diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation_20251016_062341.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation_20251016_062341.log
new file mode 100644
index 0000000000000000000000000000000000000000..ead5e952807d1af8ee141ae0f62a5252a4e2d3eb
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation_20251016_062341.log
@@ -0,0 +1,2312 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation_20251016_062341.log
+Timestamp: 2025-10-16 06:23:41
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 06:23:43,903] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:47,163] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 06:23:47,164] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 1.0 --temperature_attn_text 1.7 --temperature_mlp_text 1.7 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 1.0 --temperature_attn_vision 1.7 --temperature_mlp_vision 1.7 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 1.0 --temperature_connector 1.7 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 06:23:49,737] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:50,790] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 06:23:50,790] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 06:23:50,791] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 06:23:50,791] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 06:23:50,791] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 06:23:50,791] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 06:23:50,791] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 06:23:50,793] [INFO] [launch.py:253:main] process 2450662 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,795] [INFO] [launch.py:253:main] process 2450663 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,797] [INFO] [launch.py:253:main] process 2450664 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,798] [INFO] [launch.py:253:main] process 2450665 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,800] [INFO] [launch.py:253:main] process 2450666 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,802] [INFO] [launch.py:253:main] process 2450667 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,804] [INFO] [launch.py:253:main] process 2450668 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:23:50,806] [INFO] [launch.py:253:main] process 2450669 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '1.7', '--temperature_mlp_text', '1.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '1.7', '--temperature_mlp_vision', '1.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '1.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 06:23:57,648] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,701] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,701] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,703] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,712] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,721] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,721] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:57,721] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,187] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 06:23:58,187] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:23:58,188] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 1.7, 'temperature_mlp': 1.7, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 1.7, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 1.7,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 1.7,
+    "temperature_mlp": 1.7,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2450662:2450662 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450662:2450662 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450662:2450662 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450662:2450662 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450662:2450662 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2450662:2450662 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2450666:2450666 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450666:2450666 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450666:2450666 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450664:2450664 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450664:2450664 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450664:2450664 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450666:2450666 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450666:2450666 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450666:2450666 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2450664:2450664 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450664:2450664 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450664:2450664 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+ywang29-vrdb-test2-worker-0:2450663:2450663 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450663:2450663 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450663:2450663 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450663:2450663 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450663:2450663 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450663:2450663 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2450665:2450665 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450665:2450665 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450665:2450665 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450665:2450665 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450665:2450665 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450665:2450665 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2450667:2450667 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450667:2450667 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450667:2450667 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450667:2450667 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450667:2450667 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450667:2450667 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450668:2450668 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450668:2450668 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450668:2450668 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450668:2450668 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450668:2450668 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450668:2450668 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2450669:2450669 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2450669:2450669 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450669:2450669 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450669:2450669 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2450669:2450669 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2450669:2450669 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO ncclCommInitRank comm 0x558465d3c060 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO ncclCommInitRank comm 0x564c71db62e0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO ncclCommInitRank comm 0x564603007740 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO ncclCommInitRank comm 0x563fee077fd0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO ncclCommInitRank comm 0x55648cb4dc10 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO ncclCommInitRank comm 0x558f9f143250 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO ncclCommInitRank comm 0x558fa0d09e70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO ncclCommInitRank comm 0x56108f513290 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x97b5935be2469642 - Init START
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO comm 0x563fee077fd0 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO comm 0x558f9f143250 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO comm 0x558fa0d09e70 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO comm 0x56108f513290 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO comm 0x55648cb4dc10 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO comm 0x564603007740 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO comm 0x564c71db62e0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO comm 0x558465d3c060 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450664:2452307 [2] NCCL INFO ncclCommInitRank comm 0x55648cb4dc10 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450665:2452309 [3] NCCL INFO ncclCommInitRank comm 0x56108f513290 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450668:2452311 [6] NCCL INFO ncclCommInitRank comm 0x563fee077fd0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450669:2452312 [7] NCCL INFO ncclCommInitRank comm 0x558465d3c060 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450666:2452306 [4] NCCL INFO ncclCommInitRank comm 0x558f9f143250 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450662:2452305 [0] NCCL INFO ncclCommInitRank comm 0x564c71db62e0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450667:2452310 [5] NCCL INFO ncclCommInitRank comm 0x558fa0d09e70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x97b5935be2469642 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2450663:2452308 [1] NCCL INFO ncclCommInitRank comm 0x564603007740 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x97b5935be2469642 - Init COMPLETE
+[2025-10-16 06:24:40,751] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 06:24:44,390] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=1.000000
+Pre-training init connector._connector.0.scores: Mean=1.000005
+Pre-training init connector._connector.2.scores: Mean=0.999970
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 06:25:04,390 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 06:25:04,395 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO ncclCommInitRank comm 0x7f54b006acc0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO ncclCommInitRank comm 0x7f296006b080 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO ncclCommInitRank comm 0x7fa47c06aa40 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO ncclCommInitRank comm 0x7efad006b400 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO ncclCommInitRank comm 0x7f981806afb0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO ncclCommInitRank comm 0x7f7b9c06a7b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO ncclCommInitRank comm 0x7f19b006b020 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO ncclCommInitRank comm 0x7f884006aa90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xdc24789395ca56c2 - Init START
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO comm 0x7f19b006b020 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO comm 0x7efad006b400 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO comm 0x7fa47c06aa40 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO comm 0x7f884006aa90 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO comm 0x7f981806afb0 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO comm 0x7f54b006acc0 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO comm 0x7f296006b080 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO comm 0x7f7b9c06a7b0 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2450662:2457358 [0] NCCL INFO ncclCommInitRank comm 0x7efad006b400 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450664:2457359 [2] NCCL INFO ncclCommInitRank comm 0x7f19b006b020 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450668:2457363 [6] NCCL INFO ncclCommInitRank comm 0x7f981806afb0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450666:2457361 [4] NCCL INFO ncclCommInitRank comm 0x7f296006b080 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450667:2457365 [5] NCCL INFO ncclCommInitRank comm 0x7f54b006acc0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450669:2457362 [7] NCCL INFO ncclCommInitRank comm 0x7fa47c06aa40 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450663:2457364 [1] NCCL INFO ncclCommInitRank comm 0x7f884006aa90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xdc24789395ca56c2 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2450665:2457360 [3] NCCL INFO ncclCommInitRank comm 0x7f7b9c06a7b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xdc24789395ca56c2 - Init COMPLETE
+
  0%|          | 1/520 [00:14<2:08:17, 14.83s/it]
                                                 
{'loss': 7.8159, 'grad_norm': 0.43970953987327077, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:14<2:08:17, 14.83s/it]
  0%|          | 2/520 [00:18<1:12:08,  8.36s/it]
                                                 
{'loss': 7.0813, 'grad_norm': 0.4565463999971547, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:18<1:12:08,  8.36s/it]
  1%|          | 3/520 [00:22<54:10,  6.29s/it]  
                                               
{'loss': 6.5681, 'grad_norm': 0.26554326194639916, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:22<54:10,  6.29s/it]
  1%|          | 4/520 [00:26<45:47,  5.32s/it]
                                               
{'loss': 4.9785, 'grad_norm': 0.15355197406313056, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:26<45:47,  5.32s/it]
  1%|          | 5/520 [00:30<41:01,  4.78s/it]
                                               
{'loss': 4.5983, 'grad_norm': 0.33942625192761505, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:30<41:01,  4.78s/it]
  1%|          | 6/520 [00:33<38:11,  4.46s/it]
                                               
{'loss': 4.9659, 'grad_norm': 0.12992565021204927, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:33<38:11,  4.46s/it]
  1%|▏         | 7/520 [00:37<36:20,  4.25s/it]
                                               
{'loss': 3.4083, 'grad_norm': 0.07752778713180918, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:37<36:20,  4.25s/it]
  2%|▏         | 8/520 [00:42<36:28,  4.27s/it]
                                               
{'loss': 3.1945, 'grad_norm': 0.1981610334323709, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:42<36:28,  4.27s/it]
  2%|▏         | 9/520 [00:46<36:21,  4.27s/it]
                                               
{'loss': 2.8166, 'grad_norm': 0.03456615351080223, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:46<36:21,  4.27s/it]
  2%|▏         | 10/520 [00:50<34:41,  4.08s/it]
                                                
{'loss': 2.3844, 'grad_norm': 0.029432168415101857, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:50<34:41,  4.08s/it]
  2%|▏         | 11/520 [00:53<33:52,  3.99s/it]
                                                
{'loss': 2.3582, 'grad_norm': 0.02115123524081498, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [00:53<33:52,  3.99s/it]
  2%|▏         | 12/520 [00:57<32:56,  3.89s/it]
                                                
{'loss': 2.6345, 'grad_norm': 0.027842620589165517, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [00:57<32:56,  3.89s/it][2025-10-16 06:26:11,395] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:01<33:58,  4.02s/it]
                                                
{'loss': 2.1064, 'grad_norm': 0.01692852387214569, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:01<33:58,  4.02s/it]
  3%|▎         | 14/520 [01:05<33:02,  3.92s/it]
                                                
{'loss': 2.1278, 'grad_norm': 0.01814333817191686, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:05<33:02,  3.92s/it]
  3%|▎         | 15/520 [01:09<32:27,  3.86s/it]
                                                
{'loss': 2.3348, 'grad_norm': 0.019435395485399867, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:09<32:27,  3.86s/it]
  3%|▎         | 16/520 [01:12<31:51,  3.79s/it]
                                                
{'loss': 2.2867, 'grad_norm': 0.03434413977346103, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:12<31:51,  3.79s/it]
  3%|▎         | 17/520 [01:16<31:27,  3.75s/it]
                                                
{'loss': 2.5347, 'grad_norm': 0.10764037257874869, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:16<31:27,  3.75s/it]
  3%|▎         | 18/520 [01:20<31:12,  3.73s/it]
                                                
{'loss': 2.184, 'grad_norm': 0.05415076003880489, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:20<31:12,  3.73s/it]
  4%|▎         | 19/520 [01:23<31:06,  3.72s/it]
                                                
{'loss': 3.4027, 'grad_norm': 0.16129733625298584, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:23<31:06,  3.72s/it]
  4%|▍         | 20/520 [01:27<30:50,  3.70s/it]
                                                
{'loss': 2.7879, 'grad_norm': 0.0722917434760068, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:27<30:50,  3.70s/it]
  4%|▍         | 21/520 [01:31<30:43,  3.69s/it]
                                                
{'loss': 2.8347, 'grad_norm': 0.037312376055898266, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:31<30:43,  3.69s/it]
  4%|▍         | 22/520 [01:34<30:35,  3.69s/it]
                                                
{'loss': 2.4355, 'grad_norm': 0.03351777764224293, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:34<30:35,  3.69s/it]
  4%|▍         | 23/520 [01:38<30:24,  3.67s/it]
                                                
{'loss': 2.2557, 'grad_norm': 0.03215776173571134, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:38<30:24,  3.67s/it]
  5%|▍         | 24/520 [01:42<30:10,  3.65s/it]
                                                
{'loss': 2.4713, 'grad_norm': 0.022778390643126877, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [01:42<30:10,  3.65s/it]
  5%|▍         | 25/520 [01:45<30:15,  3.67s/it]
                                                
{'loss': 2.2893, 'grad_norm': 0.034748622889463864, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [01:45<30:15,  3.67s/it]
  5%|▌         | 26/520 [01:49<30:37,  3.72s/it]
                                                
{'loss': 2.1296, 'grad_norm': 0.016306385924809836, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [01:49<30:37,  3.72s/it]
  5%|▌         | 27/520 [01:53<30:46,  3.74s/it]
                                                
{'loss': 1.942, 'grad_norm': 0.01929347987494001, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [01:53<30:46,  3.74s/it]
  5%|▌         | 28/520 [01:57<30:55,  3.77s/it]
                                                
{'loss': 1.9557, 'grad_norm': 0.024881482606772816, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [01:57<30:55,  3.77s/it]
  6%|▌         | 29/520 [02:01<31:01,  3.79s/it]
                                                
{'loss': 1.887, 'grad_norm': 0.012163545654289995, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [02:01<31:01,  3.79s/it]
  6%|▌         | 30/520 [02:04<31:04,  3.81s/it]
                                                
{'loss': 2.352, 'grad_norm': 0.015232679170892357, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:04<31:04,  3.81s/it]
  6%|▌         | 31/520 [02:08<31:01,  3.81s/it]
                                                
{'loss': 1.8844, 'grad_norm': 0.01638573851170903, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:08<31:01,  3.81s/it]
  6%|▌         | 32/520 [02:12<30:59,  3.81s/it]
                                                
{'loss': 2.454, 'grad_norm': 0.022298797153434358, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:12<30:59,  3.81s/it]
  6%|▋         | 33/520 [02:16<31:00,  3.82s/it]
                                                
{'loss': 1.8368, 'grad_norm': 0.008518197084876807, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:16<31:00,  3.82s/it]
  7%|▋         | 34/520 [02:20<31:00,  3.83s/it]
                                                
{'loss': 1.809, 'grad_norm': 0.009924276871729495, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:20<31:00,  3.83s/it]
  7%|▋         | 35/520 [02:24<31:07,  3.85s/it]
                                                
{'loss': 1.8056, 'grad_norm': 0.007910624481857524, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:24<31:07,  3.85s/it]
  7%|▋         | 36/520 [02:28<31:01,  3.85s/it]
                                                
{'loss': 1.9224, 'grad_norm': 0.006603211385854794, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:28<31:01,  3.85s/it]
  7%|▋         | 37/520 [02:31<30:51,  3.83s/it]
                                                
{'loss': 2.2619, 'grad_norm': 0.03018451631680643, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:31<30:51,  3.83s/it]
  7%|▋         | 38/520 [02:35<30:48,  3.83s/it]
                                                
{'loss': 1.9827, 'grad_norm': 0.00739185887325306, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:35<30:48,  3.83s/it]
  8%|▊         | 39/520 [02:39<30:47,  3.84s/it]
                                                
{'loss': 1.8244, 'grad_norm': 0.007177206916465484, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [02:39<30:47,  3.84s/it]
  8%|▊         | 40/520 [02:43<30:38,  3.83s/it]
                                                
{'loss': 1.7889, 'grad_norm': 0.006831943636226794, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [02:43<30:38,  3.83s/it]
  8%|▊         | 41/520 [02:47<30:34,  3.83s/it]
                                                
{'loss': 1.7665, 'grad_norm': 0.005566807392140723, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [02:47<30:34,  3.83s/it]
  8%|▊         | 42/520 [02:51<30:30,  3.83s/it]
                                                
{'loss': 1.8344, 'grad_norm': 0.00801283544470498, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [02:51<30:30,  3.83s/it]
  8%|▊         | 43/520 [02:54<30:27,  3.83s/it]
                                                
{'loss': 2.0074, 'grad_norm': 0.010645771682366412, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [02:54<30:27,  3.83s/it]
  8%|▊         | 44/520 [02:58<30:24,  3.83s/it]
                                                
{'loss': 2.1258, 'grad_norm': 0.0077825659095028315, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [02:58<30:24,  3.83s/it]
  9%|▊         | 45/520 [03:02<30:23,  3.84s/it]
                                                
{'loss': 1.7464, 'grad_norm': 0.006648735036504383, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [03:02<30:23,  3.84s/it]
  9%|▉         | 46/520 [03:06<30:17,  3.83s/it]
                                                
{'loss': 2.1307, 'grad_norm': 0.011019869471928024, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:06<30:17,  3.83s/it]
  9%|▉         | 47/520 [03:10<30:08,  3.82s/it]
                                                
{'loss': 1.7311, 'grad_norm': 0.007187182065186074, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:10<30:08,  3.82s/it]
  9%|▉         | 48/520 [03:13<29:58,  3.81s/it]
                                                
{'loss': 1.7234, 'grad_norm': 0.00922680167175871, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:13<29:58,  3.81s/it]
  9%|▉         | 49/520 [03:17<29:50,  3.80s/it]
                                                
{'loss': 1.7135, 'grad_norm': 0.005969029660907946, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:17<29:50,  3.80s/it]
 10%|▉         | 50/520 [03:21<29:24,  3.75s/it]
                                                
{'loss': 1.7008, 'grad_norm': 0.006532514655790746, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:21<29:24,  3.75s/it]
 10%|▉         | 51/520 [03:24<29:01,  3.71s/it]
                                                
{'loss': 1.6154, 'grad_norm': 0.007140613471135835, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:24<29:01,  3.71s/it]
 10%|█         | 52/520 [03:28<28:45,  3.69s/it]
                                                
{'loss': 1.7641, 'grad_norm': 0.006796523016825306, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:28<28:45,  3.69s/it]
 10%|█         | 53/520 [03:32<28:37,  3.68s/it]
                                                
{'loss': 1.759, 'grad_norm': 0.0071262441884677764, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:32<28:37,  3.68s/it]
 10%|█         | 54/520 [03:35<28:26,  3.66s/it]
                                                
{'loss': 1.615, 'grad_norm': 0.0075376296512798075, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:35<28:26,  3.66s/it]
 11%|█         | 55/520 [03:39<28:18,  3.65s/it]
                                                
{'loss': 1.6189, 'grad_norm': 0.006529565450626214, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [03:39<28:18,  3.65s/it]
 11%|█         | 56/520 [03:43<28:06,  3.63s/it]
                                                
{'loss': 1.7589, 'grad_norm': 0.00738071949694295, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [03:43<28:06,  3.63s/it]
 11%|█         | 57/520 [03:46<28:03,  3.64s/it]
                                                
{'loss': 1.6101, 'grad_norm': 0.009805666065557406, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [03:46<28:03,  3.64s/it]
 11%|█         | 58/520 [03:50<28:01,  3.64s/it]
                                                
{'loss': 1.7539, 'grad_norm': 0.004515435127339485, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [03:50<28:01,  3.64s/it]
 11%|█▏        | 59/520 [03:54<27:55,  3.63s/it]
                                                
{'loss': 1.8022, 'grad_norm': 0.009740094690582323, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [03:54<27:55,  3.63s/it]
 12%|█▏        | 60/520 [03:57<27:50,  3.63s/it]
                                                
{'loss': 1.6841, 'grad_norm': 0.008279578682029195, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [03:57<27:50,  3.63s/it]
 12%|█▏        | 61/520 [04:01<27:47,  3.63s/it]
                                                
{'loss': 1.8958, 'grad_norm': 0.007767573521053134, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [04:01<27:47,  3.63s/it]
 12%|█▏        | 62/520 [04:04<27:40,  3.63s/it]
                                                
{'loss': 1.6353, 'grad_norm': 0.006821789872066126, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [04:04<27:40,  3.63s/it]
 12%|█▏        | 63/520 [04:08<27:43,  3.64s/it]
                                                
{'loss': 1.63, 'grad_norm': 0.006570199836945317, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:08<27:43,  3.64s/it]
 12%|█▏        | 64/520 [04:12<27:42,  3.65s/it]
                                                
{'loss': 1.6548, 'grad_norm': 0.005670119265347389, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:12<27:42,  3.65s/it]
 12%|█▎        | 65/520 [04:16<28:05,  3.70s/it]
                                                
{'loss': 1.6733, 'grad_norm': 0.006205206318452912, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:16<28:05,  3.70s/it]
 13%|█▎        | 66/520 [04:19<28:18,  3.74s/it]
                                                
{'loss': 1.6318, 'grad_norm': 0.007883408553624998, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:19<28:18,  3.74s/it]
 13%|█▎        | 67/520 [04:23<28:27,  3.77s/it]
                                                
{'loss': 1.4722, 'grad_norm': 0.0066647493930190906, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:23<28:27,  3.77s/it]
 13%|█▎        | 68/520 [04:27<28:29,  3.78s/it]
                                                
{'loss': 1.5211, 'grad_norm': 0.004853550944070299, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:27<28:29,  3.78s/it]
 13%|█▎        | 69/520 [04:31<28:29,  3.79s/it]
                                                
{'loss': 1.512, 'grad_norm': 0.00864984009911031, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:31<28:29,  3.79s/it]
 13%|█▎        | 70/520 [04:35<28:31,  3.80s/it]
                                                
{'loss': 1.5689, 'grad_norm': 0.006904568256010773, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [04:35<28:31,  3.80s/it]
 14%|█▎        | 71/520 [04:39<28:32,  3.81s/it]
                                                
{'loss': 1.472, 'grad_norm': 0.005392195001443912, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [04:39<28:32,  3.81s/it]
 14%|█▍        | 72/520 [04:42<28:27,  3.81s/it]
                                                
{'loss': 1.6343, 'grad_norm': 0.006067765425008376, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [04:42<28:27,  3.81s/it]
 14%|█▍        | 73/520 [04:46<28:31,  3.83s/it]
                                                
{'loss': 1.4525, 'grad_norm': 0.005320227735194359, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [04:46<28:31,  3.83s/it]
 14%|█▍        | 74/520 [04:50<28:28,  3.83s/it]
                                                
{'loss': 1.5871, 'grad_norm': 0.006421544977184317, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [04:50<28:28,  3.83s/it]
 14%|█▍        | 75/520 [04:54<28:26,  3.84s/it]
                                                
{'loss': 1.4612, 'grad_norm': 0.005605938985592228, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [04:54<28:26,  3.84s/it]
 15%|█▍        | 76/520 [04:58<28:21,  3.83s/it]
                                                
{'loss': 1.8659, 'grad_norm': 0.008045082061128812, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [04:58<28:21,  3.83s/it]
 15%|█▍        | 77/520 [05:02<28:22,  3.84s/it]
                                                
{'loss': 1.3801, 'grad_norm': 0.005985150498642125, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [05:02<28:22,  3.84s/it]
 15%|█▌        | 78/520 [05:05<28:22,  3.85s/it]
                                                
{'loss': 1.5286, 'grad_norm': 0.0054435338660941265, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:05<28:22,  3.85s/it]
 15%|█▌        | 79/520 [05:09<28:24,  3.86s/it]
                                                
{'loss': 1.5023, 'grad_norm': 0.005332738808959049, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:09<28:24,  3.86s/it]
 15%|█▌        | 80/520 [05:13<28:21,  3.87s/it]
                                                
{'loss': 1.8615, 'grad_norm': 0.007631621774757579, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:13<28:21,  3.87s/it]
 16%|█▌        | 81/520 [05:17<28:14,  3.86s/it]
                                                
{'loss': 1.6808, 'grad_norm': 0.006568243467414207, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:17<28:14,  3.86s/it]
 16%|█▌        | 82/520 [05:21<28:11,  3.86s/it]
                                                
{'loss': 1.5681, 'grad_norm': 0.005284179726756832, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:21<28:11,  3.86s/it]
 16%|█▌        | 83/520 [05:25<28:11,  3.87s/it]
                                                
{'loss': 1.624, 'grad_norm': 0.00548344303711468, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:25<28:11,  3.87s/it]
 16%|█▌        | 84/520 [05:29<28:09,  3.88s/it]
                                                
{'loss': 1.6068, 'grad_norm': 0.006394137204205169, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:29<28:09,  3.88s/it]
 16%|█▋        | 85/520 [05:33<28:02,  3.87s/it]
                                                
{'loss': 1.6012, 'grad_norm': 0.005008265087780791, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:33<28:02,  3.87s/it]
 17%|█▋        | 86/520 [05:36<28:00,  3.87s/it]
                                                
{'loss': 1.6549, 'grad_norm': 0.00579822351654609, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:36<28:00,  3.87s/it]
 17%|█▋        | 87/520 [05:40<27:40,  3.83s/it]
                                                
{'loss': 1.8131, 'grad_norm': 0.009110171553119793, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [05:40<27:40,  3.83s/it]
 17%|█▋        | 88/520 [05:44<27:20,  3.80s/it]
                                                
{'loss': 1.8427, 'grad_norm': 0.012905993549176013, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [05:44<27:20,  3.80s/it]
 17%|█▋        | 89/520 [05:48<27:03,  3.77s/it]
                                                
{'loss': 1.5647, 'grad_norm': 0.005773572553990653, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [05:48<27:03,  3.77s/it]
 17%|█▋        | 90/520 [05:51<26:51,  3.75s/it]
                                                
{'loss': 1.4958, 'grad_norm': 0.006127439710368184, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [05:51<26:51,  3.75s/it]
 18%|█▊        | 91/520 [05:55<26:40,  3.73s/it]
                                                
{'loss': 1.5822, 'grad_norm': 0.0049926896451864735, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [05:55<26:40,  3.73s/it]
 18%|█▊        | 92/520 [05:59<26:35,  3.73s/it]
                                                
{'loss': 1.5112, 'grad_norm': 0.006190719354974129, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [05:59<26:35,  3.73s/it]
 18%|█▊        | 93/520 [06:02<26:35,  3.74s/it]
                                                
{'loss': 1.5127, 'grad_norm': 0.00589429962972921, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [06:02<26:35,  3.74s/it]
 18%|█▊        | 94/520 [06:06<26:47,  3.77s/it]
                                                
{'loss': 1.6393, 'grad_norm': 0.006020558793696292, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [06:06<26:47,  3.77s/it]
 18%|█▊        | 95/520 [06:10<26:52,  3.79s/it]
                                                
{'loss': 1.5067, 'grad_norm': 0.005944279142634599, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:10<26:52,  3.79s/it]
 18%|█▊        | 96/520 [06:14<26:53,  3.81s/it]
                                                
{'loss': 1.4993, 'grad_norm': 0.0052737648622621395, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:14<26:53,  3.81s/it]
 19%|█▊        | 97/520 [06:18<26:50,  3.81s/it]
                                                
{'loss': 1.4667, 'grad_norm': 0.006230242367643348, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:18<26:50,  3.81s/it]
 19%|█▉        | 98/520 [06:22<26:52,  3.82s/it]
                                                
{'loss': 1.465, 'grad_norm': 0.0042739982626709624, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:22<26:52,  3.82s/it]
 19%|█▉        | 99/520 [06:25<26:50,  3.83s/it]
                                                
{'loss': 1.5035, 'grad_norm': 0.005435407935354457, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:25<26:50,  3.83s/it]
 19%|█▉        | 100/520 [06:29<26:49,  3.83s/it]
                                                 
{'loss': 1.678, 'grad_norm': 0.00818537526173001, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:29<26:49,  3.83s/it]
 19%|█▉        | 101/520 [06:33<26:41,  3.82s/it]
                                                 
{'loss': 1.479, 'grad_norm': 0.0049449369723615225, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:33<26:41,  3.82s/it]
 20%|█▉        | 102/520 [06:37<26:36,  3.82s/it]
                                                 
{'loss': 1.4963, 'grad_norm': 0.005550053419713821, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [06:37<26:36,  3.82s/it]
 20%|█▉        | 103/520 [06:41<26:29,  3.81s/it]
                                                 
{'loss': 1.4179, 'grad_norm': 0.004491806295554794, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [06:41<26:29,  3.81s/it]
 20%|██        | 104/520 [06:45<26:23,  3.81s/it]
                                                 
{'loss': 1.5085, 'grad_norm': 0.005334543734991399, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [06:45<26:23,  3.81s/it]
 20%|██        | 105/520 [06:48<26:22,  3.81s/it]
                                                 
{'loss': 1.4986, 'grad_norm': 0.004918646787641118, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [06:48<26:22,  3.81s/it]
 20%|██        | 106/520 [06:52<26:25,  3.83s/it]
                                                 
{'loss': 1.6294, 'grad_norm': 0.005436330719660856, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [06:52<26:25,  3.83s/it]
 21%|██        | 107/520 [06:56<26:24,  3.84s/it]
                                                 
{'loss': 1.6184, 'grad_norm': 0.005488251260338863, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [06:56<26:24,  3.84s/it]
 21%|██        | 108/520 [07:00<26:24,  3.85s/it]
                                                 
{'loss': 1.4591, 'grad_norm': 0.006102643329779599, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [07:00<26:24,  3.85s/it]
 21%|██        | 109/520 [07:04<26:19,  3.84s/it]
                                                 
{'loss': 1.6041, 'grad_norm': 0.004895725222725142, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [07:04<26:19,  3.84s/it]
 21%|██        | 110/520 [07:08<26:15,  3.84s/it]
                                                 
{'loss': 1.6426, 'grad_norm': 0.005431354762361593, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [07:08<26:15,  3.84s/it]
 21%|██▏       | 111/520 [07:11<25:46,  3.78s/it]
                                                 
{'loss': 1.6632, 'grad_norm': 0.006304414423865752, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:11<25:46,  3.78s/it]
 22%|██▏       | 112/520 [07:15<25:31,  3.75s/it]
                                                 
{'loss': 1.5202, 'grad_norm': 0.005168317049221944, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:15<25:31,  3.75s/it]
 22%|██▏       | 113/520 [07:19<25:14,  3.72s/it]
                                                 
{'loss': 1.3782, 'grad_norm': 0.0042387289635097474, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:19<25:14,  3.72s/it]
 22%|██▏       | 114/520 [07:22<25:01,  3.70s/it]
                                                 
{'loss': 1.4985, 'grad_norm': 0.005259652214637634, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:22<25:01,  3.70s/it]
 22%|██▏       | 115/520 [07:26<24:57,  3.70s/it]
                                                 
{'loss': 1.6311, 'grad_norm': 0.0060590694668464515, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:26<24:57,  3.70s/it]
 22%|██▏       | 116/520 [07:30<24:51,  3.69s/it]
                                                 
{'loss': 1.608, 'grad_norm': 0.004580191438664996, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:30<24:51,  3.69s/it]
 22%|██▎       | 117/520 [07:33<24:44,  3.68s/it]
                                                 
{'loss': 1.602, 'grad_norm': 0.00575561717124184, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:33<24:44,  3.68s/it]
 23%|██▎       | 118/520 [07:37<24:41,  3.69s/it]
                                                 
{'loss': 1.4684, 'grad_norm': 0.004506981882010861, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [07:37<24:41,  3.69s/it]
 23%|██▎       | 119/520 [07:41<24:36,  3.68s/it]
                                                 
{'loss': 1.4089, 'grad_norm': 0.004688833777032494, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [07:41<24:36,  3.68s/it]
 23%|██▎       | 120/520 [07:44<24:37,  3.69s/it]
                                                 
{'loss': 1.4366, 'grad_norm': 0.005141963808799315, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [07:44<24:37,  3.69s/it]
 23%|██▎       | 121/520 [07:48<24:54,  3.75s/it]
                                                 
{'loss': 1.505, 'grad_norm': 0.006607782740604067, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [07:48<24:54,  3.75s/it]
 23%|██▎       | 122/520 [07:52<24:39,  3.72s/it]
                                                 
{'loss': 1.3812, 'grad_norm': 0.004460952190988248, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [07:52<24:39,  3.72s/it]
 24%|██▎       | 123/520 [07:56<24:29,  3.70s/it]
                                                 
{'loss': 1.6783, 'grad_norm': 0.006403481215801771, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [07:56<24:29,  3.70s/it]
 24%|██▍       | 124/520 [07:59<24:21,  3.69s/it]
                                                 
{'loss': 1.4913, 'grad_norm': 0.006579073896082786, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [07:59<24:21,  3.69s/it]
 24%|██▍       | 125/520 [08:03<24:12,  3.68s/it]
                                                 
{'loss': 1.4681, 'grad_norm': 0.0050792122618557355, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [08:03<24:12,  3.68s/it]
 24%|██▍       | 126/520 [08:07<25:25,  3.87s/it]
                                                 
{'loss': 1.5582, 'grad_norm': 0.00492835374504549, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [08:07<25:25,  3.87s/it]
 24%|██▍       | 127/520 [08:11<24:57,  3.81s/it]
                                                 
{'loss': 1.4382, 'grad_norm': 0.006074014330338032, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:11<24:57,  3.81s/it]
 25%|██▍       | 128/520 [08:15<24:35,  3.76s/it]
                                                 
{'loss': 1.4985, 'grad_norm': 0.005444498120432727, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:15<24:35,  3.76s/it]
 25%|██▍       | 129/520 [08:18<24:17,  3.73s/it]
                                                 
{'loss': 1.4092, 'grad_norm': 0.00412208500014528, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:18<24:17,  3.73s/it]
 25%|██▌       | 130/520 [08:22<24:05,  3.71s/it]
                                                 
{'loss': 1.4803, 'grad_norm': 0.005250877594685346, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:22<24:05,  3.71s/it]
 25%|██▌       | 131/520 [08:25<23:57,  3.70s/it]
                                                 
{'loss': 1.5416, 'grad_norm': 0.006332721056017293, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:26<23:57,  3.70s/it]
 25%|██▌       | 132/520 [08:29<23:47,  3.68s/it]
                                                 
{'loss': 1.5209, 'grad_norm': 0.004759281756986831, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:29<23:47,  3.68s/it]
 26%|██▌       | 133/520 [08:33<23:36,  3.66s/it]
                                                 
{'loss': 1.4339, 'grad_norm': 0.005341487469341472, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:33<23:36,  3.66s/it]
 26%|██▌       | 134/520 [08:36<23:32,  3.66s/it]
                                                 
{'loss': 1.51, 'grad_norm': 0.0053912602296668985, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [08:36<23:32,  3.66s/it]
 26%|██▌       | 135/520 [08:40<23:23,  3.65s/it]
                                                 
{'loss': 1.5925, 'grad_norm': 0.004695042204581576, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [08:40<23:23,  3.65s/it]
 26%|██▌       | 136/520 [08:44<23:18,  3.64s/it]
                                                 
{'loss': 1.5073, 'grad_norm': 0.004755613699144567, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [08:44<23:18,  3.64s/it]
 26%|██▋       | 137/520 [08:47<23:15,  3.64s/it]
                                                 
{'loss': 1.4233, 'grad_norm': 0.005136276329335617, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [08:47<23:15,  3.64s/it]
 27%|██▋       | 138/520 [08:51<23:10,  3.64s/it]
                                                 
{'loss': 1.4273, 'grad_norm': 0.004235133349765746, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [08:51<23:10,  3.64s/it]
 27%|██▋       | 139/520 [08:55<23:11,  3.65s/it]
                                                 
{'loss': 1.4249, 'grad_norm': 0.005204505024612289, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [08:55<23:11,  3.65s/it]
 27%|██▋       | 140/520 [08:58<23:08,  3.65s/it]
                                                 
{'loss': 1.5914, 'grad_norm': 0.004722129789260828, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [08:58<23:08,  3.65s/it]
 27%|██▋       | 141/520 [09:02<23:06,  3.66s/it]
                                                 
{'loss': 1.5613, 'grad_norm': 0.004450587447996984, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [09:02<23:06,  3.66s/it]
 27%|██▋       | 142/520 [09:06<23:02,  3.66s/it]
                                                 
{'loss': 1.6217, 'grad_norm': 0.004516689288247079, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [09:06<23:02,  3.66s/it]
 28%|██▊       | 143/520 [09:09<22:58,  3.66s/it]
                                                 
{'loss': 1.4656, 'grad_norm': 0.005684099892504029, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:09<22:58,  3.66s/it]
 28%|██▊       | 144/520 [09:13<22:52,  3.65s/it]
                                                 
{'loss': 1.4083, 'grad_norm': 0.0049652753237298935, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:13<22:52,  3.65s/it]
 28%|██▊       | 145/520 [09:17<23:09,  3.70s/it]
                                                 
{'loss': 1.3396, 'grad_norm': 0.004206186356274259, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:17<23:09,  3.70s/it]
 28%|██▊       | 146/520 [09:21<23:20,  3.75s/it]
                                                 
{'loss': 1.6658, 'grad_norm': 0.005347704117044293, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:21<23:20,  3.75s/it]
 28%|██▊       | 147/520 [09:24<23:27,  3.77s/it]
                                                 
{'loss': 1.3913, 'grad_norm': 0.004653059597245748, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:24<23:27,  3.77s/it]
 28%|██▊       | 148/520 [09:28<23:29,  3.79s/it]
                                                 
{'loss': 1.4271, 'grad_norm': 0.004642827157580894, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:28<23:29,  3.79s/it]
 29%|██▊       | 149/520 [09:32<23:32,  3.81s/it]
                                                 
{'loss': 1.3834, 'grad_norm': 0.004925385488361721, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:32<23:32,  3.81s/it]
 29%|██▉       | 150/520 [09:36<23:34,  3.82s/it]
                                                 
{'loss': 1.6105, 'grad_norm': 0.005075466803546029, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [09:36<23:34,  3.82s/it]
 29%|██▉       | 151/520 [09:40<23:32,  3.83s/it]
                                                 
{'loss': 1.4091, 'grad_norm': 0.004300054302336075, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [09:40<23:32,  3.83s/it]
 29%|██▉       | 152/520 [09:44<23:27,  3.83s/it]
                                                 
{'loss': 1.3728, 'grad_norm': 0.00468200250940736, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [09:44<23:27,  3.83s/it]
 29%|██▉       | 153/520 [09:47<23:24,  3.83s/it]
                                                 
{'loss': 1.4264, 'grad_norm': 0.005233170575725289, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [09:47<23:24,  3.83s/it]
 30%|██▉       | 154/520 [09:51<23:20,  3.83s/it]
                                                 
{'loss': 1.5062, 'grad_norm': 0.0042933308838393296, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [09:51<23:20,  3.83s/it]
 30%|██▉       | 155/520 [09:55<23:18,  3.83s/it]
                                                 
{'loss': 1.4081, 'grad_norm': 0.0045384536249819234, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [09:55<23:18,  3.83s/it]
 30%|███       | 156/520 [09:59<23:13,  3.83s/it]
                                                 
{'loss': 1.4443, 'grad_norm': 0.0049104602040791175, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [09:59<23:13,  3.83s/it]
 30%|███       | 157/520 [10:03<22:54,  3.79s/it]
                                                 
{'loss': 1.6681, 'grad_norm': 0.005734771878033296, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [10:03<22:54,  3.79s/it]
 30%|███       | 158/520 [10:06<22:40,  3.76s/it]
                                                 
{'loss': 1.4082, 'grad_norm': 0.005057005756517819, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [10:06<22:40,  3.76s/it]
 31%|███       | 159/520 [10:10<22:28,  3.73s/it]
                                                 
{'loss': 1.442, 'grad_norm': 0.004121913062189089, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [10:10<22:28,  3.73s/it]
 31%|███       | 160/520 [10:14<22:15,  3.71s/it]
                                                 
{'loss': 1.4834, 'grad_norm': 0.004400583545271872, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:14<22:15,  3.71s/it]
 31%|███       | 161/520 [10:17<22:23,  3.74s/it]
                                                 
{'loss': 1.4678, 'grad_norm': 0.004781624737307886, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:17<22:23,  3.74s/it]
 31%|███       | 162/520 [10:21<22:31,  3.78s/it]
                                                 
{'loss': 1.5704, 'grad_norm': 0.00489734483835994, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:21<22:31,  3.78s/it]
 31%|███▏      | 163/520 [10:25<22:35,  3.80s/it]
                                                 
{'loss': 1.3294, 'grad_norm': 0.0056402273392643035, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:25<22:35,  3.80s/it]
 32%|███▏      | 164/520 [10:29<22:38,  3.82s/it]
                                                 
{'loss': 1.2919, 'grad_norm': 0.004491757066501604, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:29<22:38,  3.82s/it]
 32%|███▏      | 165/520 [10:33<22:39,  3.83s/it]
                                                 
{'loss': 1.437, 'grad_norm': 0.004334347506778019, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:33<22:39,  3.83s/it]
 32%|███▏      | 166/520 [10:37<22:39,  3.84s/it]
                                                 
{'loss': 1.4319, 'grad_norm': 0.004981211430084119, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:37<22:39,  3.84s/it]
 32%|███▏      | 167/520 [10:41<22:32,  3.83s/it]
                                                 
{'loss': 1.4275, 'grad_norm': 0.004685978748984113, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [10:41<22:32,  3.83s/it]
 32%|███▏      | 168/520 [10:44<22:09,  3.78s/it]
                                                 
{'loss': 1.3658, 'grad_norm': 0.004472834632371248, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [10:44<22:09,  3.78s/it]
 32%|███▎      | 169/520 [10:48<21:55,  3.75s/it]
                                                 
{'loss': 1.4396, 'grad_norm': 0.00421975263356354, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [10:48<21:55,  3.75s/it]
 33%|███▎      | 170/520 [10:52<21:43,  3.72s/it]
                                                 
{'loss': 1.4967, 'grad_norm': 0.004789292715254809, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [10:52<21:43,  3.72s/it]
 33%|███▎      | 171/520 [10:55<21:31,  3.70s/it]
                                                 
{'loss': 1.3713, 'grad_norm': 0.00499208774700857, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [10:55<21:31,  3.70s/it]
 33%|███▎      | 172/520 [10:59<21:21,  3.68s/it]
                                                 
{'loss': 1.4376, 'grad_norm': 0.004955506265056071, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [10:59<21:21,  3.68s/it]
 33%|███▎      | 173/520 [11:02<21:15,  3.68s/it]
                                                 
{'loss': 1.3633, 'grad_norm': 0.004178017113002678, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [11:02<21:15,  3.68s/it]
 33%|███▎      | 174/520 [11:06<21:13,  3.68s/it]
                                                 
{'loss': 1.4523, 'grad_norm': 0.005278940461203836, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [11:06<21:13,  3.68s/it]
 34%|███▎      | 175/520 [11:10<21:07,  3.67s/it]
                                                 
{'loss': 1.3445, 'grad_norm': 0.004224286465397514, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [11:10<21:07,  3.67s/it]
 34%|███▍      | 176/520 [11:13<21:03,  3.67s/it]
                                                 
{'loss': 1.5856, 'grad_norm': 0.004797176804468026, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:14<21:03,  3.67s/it]
 34%|███▍      | 177/520 [11:17<20:58,  3.67s/it]
                                                 
{'loss': 1.4448, 'grad_norm': 0.005431558436738939, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:17<20:58,  3.67s/it]
 34%|███▍      | 178/520 [11:21<20:56,  3.67s/it]
                                                 
{'loss': 1.4196, 'grad_norm': 0.004812913577153232, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:21<20:56,  3.67s/it]
 34%|███▍      | 179/520 [11:25<20:57,  3.69s/it]
                                                 
{'loss': 1.5036, 'grad_norm': 0.004168674207610028, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:25<20:57,  3.69s/it]
 35%|███▍      | 180/520 [11:28<20:54,  3.69s/it]
                                                 
{'loss': 1.415, 'grad_norm': 0.005147157280098958, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:28<20:54,  3.69s/it]
 35%|███▍      | 181/520 [11:32<20:49,  3.69s/it]
                                                 
{'loss': 1.3829, 'grad_norm': 0.004320916971841666, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:32<20:49,  3.69s/it]
 35%|███▌      | 182/520 [11:36<20:48,  3.69s/it]
                                                 
{'loss': 1.4072, 'grad_norm': 0.004349509498039605, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:36<20:48,  3.69s/it]
 35%|███▌      | 183/520 [11:39<20:45,  3.70s/it]
                                                 
{'loss': 1.4274, 'grad_norm': 0.005049564725733441, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [11:39<20:45,  3.70s/it]
 35%|███▌      | 184/520 [11:43<20:41,  3.70s/it]
                                                 
{'loss': 1.3379, 'grad_norm': 0.004603998236240537, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [11:43<20:41,  3.70s/it]
 36%|███▌      | 185/520 [11:47<20:36,  3.69s/it]
                                                 
{'loss': 1.5244, 'grad_norm': 0.004449378741584875, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [11:47<20:36,  3.69s/it]
 36%|███▌      | 186/520 [11:50<20:29,  3.68s/it]
                                                 
{'loss': 1.3531, 'grad_norm': 0.004280264036884546, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [11:50<20:29,  3.68s/it]
 36%|███▌      | 187/520 [11:54<20:22,  3.67s/it]
                                                 
{'loss': 1.3691, 'grad_norm': 0.005191242305760531, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [11:54<20:22,  3.67s/it]
 36%|███▌      | 188/520 [11:58<20:18,  3.67s/it]
                                                 
{'loss': 1.4435, 'grad_norm': 0.004708598729854467, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [11:58<20:18,  3.67s/it]
 36%|███▋      | 189/520 [12:01<20:23,  3.70s/it]
                                                 
{'loss': 1.4623, 'grad_norm': 0.004073562963008043, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [12:01<20:23,  3.70s/it]
 37%|███▋      | 190/520 [12:05<20:32,  3.74s/it]
                                                 
{'loss': 1.3724, 'grad_norm': 0.004649755775523317, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [12:05<20:32,  3.74s/it]
 37%|███▋      | 191/520 [12:09<20:38,  3.76s/it]
                                                 
{'loss': 1.3248, 'grad_norm': 0.0041525172490652455, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [12:09<20:38,  3.76s/it]
 37%|███▋      | 192/520 [12:13<20:35,  3.77s/it]
                                                 
{'loss': 1.4268, 'grad_norm': 0.004556107753714613, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:13<20:35,  3.77s/it]
 37%|███▋      | 193/520 [12:17<20:25,  3.75s/it]
                                                 
{'loss': 1.5092, 'grad_norm': 0.005130463724985093, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:17<20:25,  3.75s/it]
 37%|███▋      | 194/520 [12:20<20:17,  3.73s/it]
                                                 
{'loss': 1.3776, 'grad_norm': 0.0058884463249506075, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:20<20:17,  3.73s/it]
 38%|███▊      | 195/520 [12:24<20:05,  3.71s/it]
                                                 
{'loss': 1.4359, 'grad_norm': 0.004197648053720576, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:24<20:05,  3.71s/it]
 38%|███▊      | 196/520 [12:28<19:57,  3.70s/it]
                                                 
{'loss': 1.3973, 'grad_norm': 0.004940748288150987, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:28<19:57,  3.70s/it]
 38%|███▊      | 197/520 [12:31<19:51,  3.69s/it]
                                                 
{'loss': 1.3654, 'grad_norm': 0.004607237611439345, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:31<19:51,  3.69s/it]
 38%|███▊      | 198/520 [12:35<19:46,  3.68s/it]
                                                 
{'loss': 1.45, 'grad_norm': 0.0047456433590902045, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:35<19:46,  3.68s/it]
 38%|███▊      | 199/520 [12:39<19:40,  3.68s/it]
                                                 
{'loss': 1.3601, 'grad_norm': 0.004449620458339757, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [12:39<19:40,  3.68s/it]
 38%|███▊      | 200/520 [12:42<19:39,  3.69s/it]
                                                 
{'loss': 1.4426, 'grad_norm': 0.005020257790146785, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [12:42<19:39,  3.69s/it]
 39%|███▊      | 201/520 [12:46<19:37,  3.69s/it]
                                                 
{'loss': 1.4596, 'grad_norm': 0.0041478024927893925, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [12:46<19:37,  3.69s/it]
 39%|███▉      | 202/520 [12:50<19:30,  3.68s/it]
                                                 
{'loss': 1.3463, 'grad_norm': 0.004466236135486162, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [12:50<19:30,  3.68s/it]
 39%|███▉      | 203/520 [12:53<19:22,  3.67s/it]
                                                 
{'loss': 1.3988, 'grad_norm': 0.004407225871965885, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [12:53<19:22,  3.67s/it]
 39%|███▉      | 204/520 [12:57<19:15,  3.66s/it]
                                                 
{'loss': 1.4413, 'grad_norm': 0.004374582947084023, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [12:57<19:15,  3.66s/it]
 39%|███▉      | 205/520 [13:01<19:13,  3.66s/it]
                                                 
{'loss': 1.4728, 'grad_norm': 0.005404123052086189, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [13:01<19:13,  3.66s/it]
 40%|███▉      | 206/520 [13:04<19:07,  3.65s/it]
                                                 
{'loss': 1.4737, 'grad_norm': 0.004280181445647264, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [13:04<19:07,  3.65s/it]
 40%|███▉      | 207/520 [13:08<19:04,  3.66s/it]
                                                 
{'loss': 1.4464, 'grad_norm': 0.004429276911566513, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [13:08<19:04,  3.66s/it]
 40%|████      | 208/520 [13:12<18:59,  3.65s/it]
                                                 
{'loss': 1.444, 'grad_norm': 0.004800120644715747, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [13:12<18:59,  3.65s/it]
 40%|████      | 209/520 [13:15<18:55,  3.65s/it]
                                                 
{'loss': 1.3677, 'grad_norm': 0.0043582032255337555, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:15<18:55,  3.65s/it]
 40%|████      | 210/520 [13:19<18:48,  3.64s/it]
                                                 
{'loss': 1.4396, 'grad_norm': 0.004805545256793099, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:19<18:48,  3.64s/it]
 41%|████      | 211/520 [13:22<18:44,  3.64s/it]
                                                 
{'loss': 1.4456, 'grad_norm': 0.00400299224166497, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:22<18:44,  3.64s/it]
 41%|████      | 212/520 [13:26<18:39,  3.63s/it]
                                                 
{'loss': 1.4165, 'grad_norm': 0.004137104783219967, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:26<18:39,  3.63s/it]
 41%|████      | 213/520 [13:30<18:36,  3.64s/it]
                                                 
{'loss': 1.3869, 'grad_norm': 0.004883093417641553, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:30<18:36,  3.64s/it]
 41%|████      | 214/520 [13:33<18:31,  3.63s/it]
                                                 
{'loss': 1.3731, 'grad_norm': 0.004434087916701376, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:33<18:31,  3.63s/it]
 41%|████▏     | 215/520 [13:37<18:34,  3.65s/it]
                                                 
{'loss': 1.3743, 'grad_norm': 0.004575880538605672, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [13:37<18:34,  3.65s/it]
 42%|████▏     | 216/520 [13:41<18:28,  3.65s/it]
                                                 
{'loss': 1.2841, 'grad_norm': 0.0042295391815006435, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [13:41<18:28,  3.65s/it]
 42%|████▏     | 217/520 [13:44<18:25,  3.65s/it]
                                                 
{'loss': 1.4143, 'grad_norm': 0.004328455201247795, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [13:44<18:25,  3.65s/it]
 42%|████▏     | 218/520 [13:48<18:24,  3.66s/it]
                                                 
{'loss': 1.4051, 'grad_norm': 0.004454790570768995, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [13:48<18:24,  3.66s/it]
 42%|████▏     | 219/520 [13:52<18:20,  3.66s/it]
                                                 
{'loss': 1.3688, 'grad_norm': 0.003796282908924954, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [13:52<18:20,  3.66s/it]
 42%|████▏     | 220/520 [13:55<18:15,  3.65s/it]
                                                 
{'loss': 1.4318, 'grad_norm': 0.004938477728663886, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [13:55<18:15,  3.65s/it]
 42%|████▎     | 221/520 [13:59<18:13,  3.66s/it]
                                                 
{'loss': 1.4122, 'grad_norm': 0.004273396064311793, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [13:59<18:13,  3.66s/it]
 43%|████▎     | 222/520 [14:03<18:08,  3.65s/it]
                                                 
{'loss': 1.3177, 'grad_norm': 0.004060149461454274, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [14:03<18:08,  3.65s/it]
 43%|████▎     | 223/520 [14:06<18:06,  3.66s/it]
                                                 
{'loss': 1.3004, 'grad_norm': 0.004196527395243219, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [14:06<18:06,  3.66s/it]
 43%|████▎     | 224/520 [14:10<18:04,  3.66s/it]
                                                 
{'loss': 1.6403, 'grad_norm': 0.0075597478643579285, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [14:10<18:04,  3.66s/it]
 43%|████▎     | 225/520 [14:14<17:59,  3.66s/it]
                                                 
{'loss': 1.3332, 'grad_norm': 0.0044397676832995305, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:14<17:59,  3.66s/it]
 43%|████▎     | 226/520 [14:17<17:56,  3.66s/it]
                                                 
{'loss': 1.4344, 'grad_norm': 0.004201713867972385, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:17<17:56,  3.66s/it]
 44%|████▎     | 227/520 [14:21<17:54,  3.67s/it]
                                                 
{'loss': 1.4136, 'grad_norm': 0.004042588438572285, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:21<17:54,  3.67s/it]
 44%|████▍     | 228/520 [14:25<17:49,  3.66s/it]
                                                 
{'loss': 1.5842, 'grad_norm': 0.004705616511010748, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:25<17:49,  3.66s/it]
 44%|████▍     | 229/520 [14:28<17:47,  3.67s/it]
                                                 
{'loss': 1.3994, 'grad_norm': 0.004115866698191579, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:28<17:47,  3.67s/it]
 44%|████▍     | 230/520 [14:32<17:44,  3.67s/it]
                                                 
{'loss': 1.2713, 'grad_norm': 0.004480028898857866, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:32<17:44,  3.67s/it]
 44%|████▍     | 231/520 [14:36<17:39,  3.67s/it]
                                                 
{'loss': 1.3308, 'grad_norm': 0.0038864939519623167, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:36<17:39,  3.67s/it]
 45%|████▍     | 232/520 [14:39<17:34,  3.66s/it]
                                                 
{'loss': 1.6094, 'grad_norm': 0.004593441359797257, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [14:39<17:34,  3.66s/it]
 45%|████▍     | 233/520 [14:43<17:30,  3.66s/it]
                                                 
{'loss': 1.4785, 'grad_norm': 0.0051138719592116, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [14:43<17:30,  3.66s/it]
 45%|████▌     | 234/520 [14:47<17:38,  3.70s/it]
                                                 
{'loss': 1.278, 'grad_norm': 0.004184365243695636, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [14:47<17:38,  3.70s/it]
 45%|████▌     | 235/520 [14:51<17:42,  3.73s/it]
                                                 
{'loss': 1.3282, 'grad_norm': 0.004817097605823089, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [14:51<17:42,  3.73s/it]
 45%|████▌     | 236/520 [14:54<17:31,  3.70s/it]
                                                 
{'loss': 1.4569, 'grad_norm': 0.0038550205157249716, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [14:54<17:31,  3.70s/it]
 46%|████▌     | 237/520 [14:58<17:23,  3.69s/it]
                                                 
{'loss': 1.4173, 'grad_norm': 0.0041194432678913605, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [14:58<17:23,  3.69s/it]
 46%|████▌     | 238/520 [15:01<17:14,  3.67s/it]
                                                 
{'loss': 1.3522, 'grad_norm': 0.004440433333590445, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [15:01<17:14,  3.67s/it]
 46%|████▌     | 239/520 [15:05<17:08,  3.66s/it]
                                                 
{'loss': 1.4631, 'grad_norm': 0.004489529877156505, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [15:05<17:08,  3.66s/it]
 46%|████▌     | 240/520 [15:09<17:03,  3.66s/it]
                                                 
{'loss': 1.2066, 'grad_norm': 0.004806405580017891, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:09<17:03,  3.66s/it]
 46%|████▋     | 241/520 [15:12<16:59,  3.65s/it]
                                                 
{'loss': 1.3027, 'grad_norm': 0.004133751653376319, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:12<16:59,  3.65s/it]
 47%|████▋     | 242/520 [15:16<16:57,  3.66s/it]
                                                 
{'loss': 1.3224, 'grad_norm': 0.003763887966995485, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:16<16:57,  3.66s/it]
 47%|████▋     | 243/520 [15:20<16:53,  3.66s/it]
                                                 
{'loss': 1.317, 'grad_norm': 0.004145135134474499, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:20<16:53,  3.66s/it]
 47%|████▋     | 244/520 [15:23<16:52,  3.67s/it]
                                                 
{'loss': 1.4653, 'grad_norm': 0.004289253528936598, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:23<16:52,  3.67s/it]
 47%|████▋     | 245/520 [15:27<16:50,  3.67s/it]
                                                 
{'loss': 1.3103, 'grad_norm': 0.004028200588514878, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:27<16:50,  3.67s/it]
 47%|████▋     | 246/520 [15:31<16:46,  3.67s/it]
                                                 
{'loss': 1.5744, 'grad_norm': 0.004362097760632629, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:31<16:46,  3.67s/it]
 48%|████▊     | 247/520 [15:34<16:41,  3.67s/it]
                                                 
{'loss': 1.4775, 'grad_norm': 0.004058922151854034, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:34<16:41,  3.67s/it]
 48%|████▊     | 248/520 [15:38<16:40,  3.68s/it]
                                                 
{'loss': 1.296, 'grad_norm': 0.004157401172321326, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [15:38<16:40,  3.68s/it]
 48%|████▊     | 249/520 [15:42<16:40,  3.69s/it]
                                                 
{'loss': 1.4186, 'grad_norm': 0.004377571264908201, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [15:42<16:40,  3.69s/it]
 48%|████▊     | 250/520 [15:46<16:35,  3.69s/it]
                                                 
{'loss': 1.3593, 'grad_norm': 0.004855364226602438, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [15:46<16:35,  3.69s/it]
 48%|████▊     | 251/520 [15:49<16:31,  3.69s/it]
                                                 
{'loss': 1.4126, 'grad_norm': 0.0037922820802587243, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [15:49<16:31,  3.69s/it]
 48%|████▊     | 252/520 [15:53<16:25,  3.68s/it]
                                                 
{'loss': 1.4472, 'grad_norm': 0.004132688634192368, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [15:53<16:25,  3.68s/it]
 49%|████▊     | 253/520 [15:57<16:23,  3.68s/it]
                                                 
{'loss': 1.4129, 'grad_norm': 0.004716884546417268, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [15:57<16:23,  3.68s/it]
 49%|████▉     | 254/520 [16:00<16:15,  3.67s/it]
                                                 
{'loss': 1.3197, 'grad_norm': 0.003999666696694521, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [16:00<16:15,  3.67s/it]
 49%|████▉     | 255/520 [16:04<16:09,  3.66s/it]
                                                 
{'loss': 1.3318, 'grad_norm': 0.004434908929244165, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [16:04<16:09,  3.66s/it]
 49%|████▉     | 256/520 [16:07<16:06,  3.66s/it]
                                                 
{'loss': 1.3781, 'grad_norm': 0.004541577550730151, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:07<16:06,  3.66s/it]
 49%|████▉     | 257/520 [16:11<16:01,  3.66s/it]
                                                 
{'loss': 1.3763, 'grad_norm': 0.004331758306873519, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:11<16:01,  3.66s/it]
 50%|████▉     | 258/520 [16:15<15:57,  3.66s/it]
                                                 
{'loss': 1.3937, 'grad_norm': 0.0036091178101941663, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:15<15:57,  3.66s/it]
 50%|████▉     | 259/520 [16:18<15:52,  3.65s/it]
                                                 
{'loss': 1.4573, 'grad_norm': 0.004703438939471208, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:18<15:52,  3.65s/it]
 50%|█████     | 260/520 [16:22<15:48,  3.65s/it]
                                                 
{'loss': 1.5246, 'grad_norm': 0.003967955222138265, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:22<15:48,  3.65s/it]
 50%|█████     | 261/520 [16:26<15:46,  3.66s/it]
                                                 
{'loss': 1.4561, 'grad_norm': 0.0041601351473908406, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:26<15:46,  3.66s/it]
 50%|█████     | 262/520 [16:29<15:43,  3.66s/it]
                                                 
{'loss': 1.2956, 'grad_norm': 0.004469965006462439, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:29<15:43,  3.66s/it]
 51%|█████     | 263/520 [16:33<15:50,  3.70s/it]
                                                 
{'loss': 1.4601, 'grad_norm': 0.004407465204833223, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:33<15:50,  3.70s/it]
 51%|█████     | 264/520 [16:37<15:59,  3.75s/it]
                                                 
{'loss': 1.4175, 'grad_norm': 0.004034811793370448, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [16:37<15:59,  3.75s/it]
 51%|█████     | 265/520 [16:41<16:05,  3.79s/it]
                                                 
{'loss': 1.316, 'grad_norm': 0.005161436061894714, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [16:41<16:05,  3.79s/it]
 51%|█████     | 266/520 [16:45<16:05,  3.80s/it]
                                                 
{'loss': 1.1609, 'grad_norm': 0.0035948784179094936, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [16:45<16:05,  3.80s/it]
 51%|█████▏    | 267/520 [16:49<16:08,  3.83s/it]
                                                 
{'loss': 1.3067, 'grad_norm': 0.003933326350398873, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [16:49<16:08,  3.83s/it]
 52%|█████▏    | 268/520 [16:53<16:06,  3.84s/it]
                                                 
{'loss': 1.5919, 'grad_norm': 0.005497439422813538, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [16:53<16:06,  3.84s/it]
 52%|█████▏    | 269/520 [16:56<16:06,  3.85s/it]
                                                 
{'loss': 1.4324, 'grad_norm': 0.0043459530099447955, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [16:56<16:06,  3.85s/it]
 52%|█████▏    | 270/520 [17:00<15:48,  3.79s/it]
                                                 
{'loss': 1.3751, 'grad_norm': 0.003985771250686427, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [17:00<15:48,  3.79s/it]
 52%|█████▏    | 271/520 [17:04<15:35,  3.76s/it]
                                                 
{'loss': 1.4113, 'grad_norm': 0.004086400711578378, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [17:04<15:35,  3.76s/it]
 52%|█████▏    | 272/520 [17:07<15:27,  3.74s/it]
                                                 
{'loss': 1.386, 'grad_norm': 0.004795760414216506, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:07<15:27,  3.74s/it]
 52%|█████▎    | 273/520 [17:11<15:17,  3.71s/it]
                                                 
{'loss': 1.5279, 'grad_norm': 0.004759462216118722, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:11<15:17,  3.71s/it]
 53%|█████▎    | 274/520 [17:15<15:08,  3.69s/it]
                                                 
{'loss': 1.3677, 'grad_norm': 0.004469993852463788, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:15<15:08,  3.69s/it]
 53%|█████▎    | 275/520 [17:18<15:03,  3.69s/it]
                                                 
{'loss': 1.3032, 'grad_norm': 0.00460301742482514, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:18<15:03,  3.69s/it]
 53%|█████▎    | 276/520 [17:22<14:59,  3.69s/it]
                                                 
{'loss': 1.3935, 'grad_norm': 0.004375178223752344, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:22<14:59,  3.69s/it]
 53%|█████▎    | 277/520 [17:26<14:57,  3.69s/it]
                                                 
{'loss': 1.5114, 'grad_norm': 0.005691170418719667, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:26<14:57,  3.69s/it]
 53%|█████▎    | 278/520 [17:30<14:54,  3.70s/it]
                                                 
{'loss': 1.2586, 'grad_norm': 0.003974250687016974, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:30<14:54,  3.70s/it]
 54%|█████▎    | 279/520 [17:33<14:48,  3.69s/it]
                                                 
{'loss': 1.4102, 'grad_norm': 0.00636443891782005, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:33<14:48,  3.69s/it]
 54%|█████▍    | 280/520 [17:37<14:47,  3.70s/it]
                                                 
{'loss': 1.3071, 'grad_norm': 0.004510065573721173, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [17:37<14:47,  3.70s/it]
 54%|█████▍    | 281/520 [17:41<14:40,  3.68s/it]
                                                 
{'loss': 1.4266, 'grad_norm': 0.004424755207342497, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [17:41<14:40,  3.68s/it]
 54%|█████▍    | 282/520 [17:44<14:37,  3.69s/it]
                                                 
{'loss': 1.2644, 'grad_norm': 0.004262179835849369, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [17:44<14:37,  3.69s/it]
 54%|█████▍    | 283/520 [17:48<14:36,  3.70s/it]
                                                 
{'loss': 1.4519, 'grad_norm': 0.004749679723122533, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [17:48<14:36,  3.70s/it]
 55%|█████▍    | 284/520 [17:52<14:32,  3.70s/it]
                                                 
{'loss': 1.3747, 'grad_norm': 0.004643944999040611, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [17:52<14:32,  3.70s/it]
 55%|█████▍    | 285/520 [17:55<14:27,  3.69s/it]
                                                 
{'loss': 1.298, 'grad_norm': 0.004260840224894914, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [17:55<14:27,  3.69s/it]
 55%|█████▌    | 286/520 [17:59<14:26,  3.70s/it]
                                                 
{'loss': 1.158, 'grad_norm': 0.004401926309780154, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [17:59<14:26,  3.70s/it]
 55%|█████▌    | 287/520 [18:03<14:21,  3.70s/it]
                                                 
{'loss': 1.4172, 'grad_norm': 0.004511241765839045, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [18:03<14:21,  3.70s/it]
 55%|█████▌    | 288/520 [18:06<14:19,  3.71s/it]
                                                 
{'loss': 1.4708, 'grad_norm': 0.004487727231419834, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:06<14:19,  3.71s/it]
 56%|█████▌    | 289/520 [18:10<14:17,  3.71s/it]
                                                 
{'loss': 1.3081, 'grad_norm': 0.003978210310502783, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:10<14:17,  3.71s/it]
 56%|█████▌    | 290/520 [18:14<14:11,  3.70s/it]
                                                 
{'loss': 1.2225, 'grad_norm': 0.0038557911346834904, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:14<14:11,  3.70s/it]
 56%|█████▌    | 291/520 [18:18<14:06,  3.70s/it]
                                                 
{'loss': 1.2859, 'grad_norm': 0.00487151203878196, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:18<14:06,  3.70s/it]
 56%|█████▌    | 292/520 [18:21<14:02,  3.69s/it]
                                                 
{'loss': 1.336, 'grad_norm': 0.003930491143232352, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:21<14:02,  3.69s/it]
 56%|█████▋    | 293/520 [18:25<13:57,  3.69s/it]
                                                 
{'loss': 1.2731, 'grad_norm': 0.004700323461452995, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:25<13:57,  3.69s/it]
 57%|█████▋    | 294/520 [18:29<13:54,  3.69s/it]
                                                 
{'loss': 1.3106, 'grad_norm': 0.004380411953238329, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:29<13:54,  3.69s/it]
 57%|█████▋    | 295/520 [18:32<13:49,  3.69s/it]
                                                 
{'loss': 1.4519, 'grad_norm': 0.005615635122717599, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:32<13:49,  3.69s/it]
 57%|█████▋    | 296/520 [18:36<13:44,  3.68s/it]
                                                 
{'loss': 1.2409, 'grad_norm': 0.004414005560246888, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [18:36<13:44,  3.68s/it]
 57%|█████▋    | 297/520 [18:40<13:39,  3.68s/it]
                                                 
{'loss': 1.3901, 'grad_norm': 0.004548506029577162, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [18:40<13:39,  3.68s/it]
 57%|█████▋    | 298/520 [18:43<13:35,  3.67s/it]
                                                 
{'loss': 1.3472, 'grad_norm': 0.00377369593838872, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [18:43<13:35,  3.67s/it]
 57%|█████▊    | 299/520 [18:47<13:32,  3.68s/it]
                                                 
{'loss': 1.4547, 'grad_norm': 0.004248809730848529, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [18:47<13:32,  3.68s/it]
 58%|█████▊    | 300/520 [18:51<13:27,  3.67s/it]
                                                 
{'loss': 1.4107, 'grad_norm': 0.004458847536967402, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [18:51<13:27,  3.67s/it]
 58%|█████▊    | 301/520 [18:55<13:42,  3.76s/it]
                                                 
{'loss': 1.3732, 'grad_norm': 0.0041094014346831345, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [18:55<13:42,  3.76s/it]
 58%|█████▊    | 302/520 [18:59<13:52,  3.82s/it]
                                                 
{'loss': 1.4777, 'grad_norm': 0.004277495045633765, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [18:59<13:52,  3.82s/it]
 58%|█████▊    | 303/520 [19:03<13:55,  3.85s/it]
                                                 
{'loss': 1.3056, 'grad_norm': 0.004336677395022478, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [19:03<13:55,  3.85s/it]
 58%|█████▊    | 304/520 [19:07<14:01,  3.90s/it]
                                                 
{'loss': 1.3674, 'grad_norm': 0.004403378290326746, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:07<14:01,  3.90s/it]
 59%|█████▊    | 305/520 [19:10<13:53,  3.88s/it]
                                                 
{'loss': 1.4348, 'grad_norm': 0.00490548741840277, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:10<13:53,  3.88s/it]
 59%|█████▉    | 306/520 [19:14<13:39,  3.83s/it]
                                                 
{'loss': 1.3555, 'grad_norm': 0.0040838871783450675, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:14<13:39,  3.83s/it]
 59%|█████▉    | 307/520 [19:18<13:49,  3.90s/it]
                                                 
{'loss': 1.2919, 'grad_norm': 0.003961878562728322, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:18<13:49,  3.90s/it]
 59%|█████▉    | 308/520 [19:22<13:30,  3.82s/it]
                                                 
{'loss': 1.4171, 'grad_norm': 0.004192209223866029, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:22<13:30,  3.82s/it]
 59%|█████▉    | 309/520 [19:25<13:14,  3.76s/it]
                                                 
{'loss': 1.283, 'grad_norm': 0.0039114224837664855, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:25<13:14,  3.76s/it]
 60%|█████▉    | 310/520 [19:29<13:04,  3.74s/it]
                                                 
{'loss': 1.2675, 'grad_norm': 0.004047126698825762, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:29<13:04,  3.74s/it]
 60%|█████▉    | 311/520 [19:33<12:56,  3.72s/it]
                                                 
{'loss': 1.2314, 'grad_norm': 0.004005294629048275, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [19:33<12:56,  3.72s/it]
 60%|██████    | 312/520 [19:36<12:48,  3.69s/it]
                                                 
{'loss': 1.224, 'grad_norm': 0.004817439552502824, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [19:36<12:48,  3.69s/it]
 60%|██████    | 313/520 [19:40<12:41,  3.68s/it]
                                                 
{'loss': 1.2129, 'grad_norm': 0.0036841170735105754, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [19:40<12:41,  3.68s/it]
 60%|██████    | 314/520 [19:44<13:00,  3.79s/it]
                                                 
{'loss': 1.2557, 'grad_norm': 0.0040767391392939, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [19:44<13:00,  3.79s/it]
 61%|██████    | 315/520 [19:48<12:49,  3.75s/it]
                                                 
{'loss': 1.4459, 'grad_norm': 0.005557294254899473, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [19:48<12:49,  3.75s/it]
 61%|██████    | 316/520 [19:52<13:04,  3.85s/it]
                                                 
{'loss': 1.2284, 'grad_norm': 0.0048706406574788265, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [19:52<13:04,  3.85s/it]
 61%|██████    | 317/520 [19:55<12:49,  3.79s/it]
                                                 
{'loss': 1.2463, 'grad_norm': 0.003776099273651555, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [19:55<12:49,  3.79s/it]
 61%|██████    | 318/520 [19:59<12:41,  3.77s/it]
                                                 
{'loss': 1.3777, 'grad_norm': 0.004229222887268496, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [19:59<12:41,  3.77s/it]
 61%|██████▏   | 319/520 [20:03<12:58,  3.87s/it]
                                                 
{'loss': 1.2385, 'grad_norm': 0.00407032811923965, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [20:03<12:58,  3.87s/it]
 62%|██████▏   | 320/520 [20:07<12:41,  3.81s/it]
                                                 
{'loss': 1.1778, 'grad_norm': 0.004196643863798075, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:07<12:41,  3.81s/it]
 62%|██████▏   | 321/520 [20:11<12:28,  3.76s/it]
                                                 
{'loss': 1.391, 'grad_norm': 0.0042350124198092245, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:11<12:28,  3.76s/it]
 62%|██████▏   | 322/520 [20:14<12:18,  3.73s/it]
                                                 
{'loss': 1.288, 'grad_norm': 0.004422999673282651, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:14<12:18,  3.73s/it]
 62%|██████▏   | 323/520 [20:18<12:09,  3.70s/it]
                                                 
{'loss': 1.3741, 'grad_norm': 0.005137490707398243, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:18<12:09,  3.70s/it]
 62%|██████▏   | 324/520 [20:22<12:00,  3.68s/it]
                                                 
{'loss': 1.3137, 'grad_norm': 0.004348896044696881, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:22<12:00,  3.68s/it]
 62%|██████▎   | 325/520 [20:25<11:55,  3.67s/it]
                                                 
{'loss': 1.3273, 'grad_norm': 0.0044135014209218055, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:25<11:55,  3.67s/it]
 63%|██████▎   | 326/520 [20:29<11:49,  3.66s/it]
                                                 
{'loss': 1.3114, 'grad_norm': 0.0041817699546673895, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:29<11:49,  3.66s/it]
 63%|██████▎   | 327/520 [20:32<11:44,  3.65s/it]
                                                 
{'loss': 1.4506, 'grad_norm': 0.005647818060604807, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [20:32<11:44,  3.65s/it]
 63%|██████▎   | 328/520 [20:36<11:38,  3.64s/it]
                                                 
{'loss': 1.3838, 'grad_norm': 0.004224868504875673, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [20:36<11:38,  3.64s/it]
 63%|██████▎   | 329/520 [20:40<11:37,  3.65s/it]
                                                 
{'loss': 1.2223, 'grad_norm': 0.00367121573655905, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [20:40<11:37,  3.65s/it]
 63%|██████▎   | 330/520 [20:43<11:34,  3.65s/it]
                                                 
{'loss': 1.2926, 'grad_norm': 0.0037327068004085525, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [20:43<11:34,  3.65s/it]
 64%|██████▎   | 331/520 [20:47<11:31,  3.66s/it]
                                                 
{'loss': 1.265, 'grad_norm': 0.003869252221910313, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [20:47<11:31,  3.66s/it]
 64%|██████▍   | 332/520 [20:51<11:28,  3.66s/it]
                                                 
{'loss': 1.4444, 'grad_norm': 0.004099074356840049, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [20:51<11:28,  3.66s/it]
 64%|██████▍   | 333/520 [20:54<11:25,  3.67s/it]
                                                 
{'loss': 1.4411, 'grad_norm': 0.004350600657517941, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [20:54<11:25,  3.67s/it]
 64%|██████▍   | 334/520 [20:58<11:21,  3.67s/it]
                                                 
{'loss': 1.3187, 'grad_norm': 0.004733525537497669, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [20:58<11:21,  3.67s/it]
 64%|██████▍   | 335/520 [21:02<11:19,  3.67s/it]
                                                 
{'loss': 1.3062, 'grad_norm': 0.003966021566931298, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [21:02<11:19,  3.67s/it]
 65%|██████▍   | 336/520 [21:05<11:17,  3.68s/it]
                                                 
{'loss': 1.1896, 'grad_norm': 0.0046557618005457916, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:05<11:17,  3.68s/it]
 65%|██████▍   | 337/520 [21:09<11:13,  3.68s/it]
                                                 
{'loss': 1.1944, 'grad_norm': 0.004248250580321976, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:09<11:13,  3.68s/it]
 65%|██████▌   | 338/520 [21:13<11:09,  3.68s/it]
                                                 
{'loss': 1.3291, 'grad_norm': 0.004283975446197311, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:13<11:09,  3.68s/it]
 65%|██████▌   | 339/520 [21:16<11:03,  3.67s/it]
                                                 
{'loss': 1.2615, 'grad_norm': 0.004197514979797665, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:16<11:03,  3.67s/it]
 65%|██████▌   | 340/520 [21:20<10:58,  3.66s/it]
                                                 
{'loss': 1.2494, 'grad_norm': 0.003995351098660928, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:20<10:58,  3.66s/it]
 66%|██████▌   | 341/520 [21:24<10:54,  3.66s/it]
                                                 
{'loss': 1.2754, 'grad_norm': 0.004238005795245443, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:24<10:54,  3.66s/it]
 66%|██████▌   | 342/520 [21:27<10:50,  3.65s/it]
                                                 
{'loss': 1.436, 'grad_norm': 0.006902032649706279, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:27<10:50,  3.65s/it]
 66%|██████▌   | 343/520 [21:31<10:45,  3.65s/it]
                                                 
{'loss': 1.3823, 'grad_norm': 0.004399201836303718, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [21:31<10:45,  3.65s/it]
 66%|██████▌   | 344/520 [21:35<10:41,  3.64s/it]
                                                 
{'loss': 1.2188, 'grad_norm': 0.003984553275321056, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [21:35<10:41,  3.64s/it]
 66%|██████▋   | 345/520 [21:38<10:37,  3.64s/it]
                                                 
{'loss': 1.3392, 'grad_norm': 0.004426573793946866, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [21:38<10:37,  3.64s/it]
 67%|██████▋   | 346/520 [21:42<10:32,  3.64s/it]
                                                 
{'loss': 1.3683, 'grad_norm': 0.004375397756524048, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [21:42<10:32,  3.64s/it]
 67%|██████▋   | 347/520 [21:46<10:29,  3.64s/it]
                                                 
{'loss': 1.2413, 'grad_norm': 0.003795779296716695, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [21:46<10:29,  3.64s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [21:49<10:26,  3.64s/it]
                                                 
{'loss': 1.1976, 'grad_norm': 0.004817143918513408, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [21:49<10:26,  3.64s/it]
 67%|██████▋   | 349/520 [21:53<10:23,  3.65s/it]
                                                 
{'loss': 1.2462, 'grad_norm': 0.004135456843120722, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [21:53<10:23,  3.65s/it]
 67%|██████▋   | 350/520 [21:56<10:19,  3.64s/it]
                                                 
{'loss': 1.2748, 'grad_norm': 0.004736797883612023, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [21:56<10:19,  3.64s/it]
 68%|██████▊   | 351/520 [22:00<10:16,  3.65s/it]
                                                 
{'loss': 1.182, 'grad_norm': 0.004020955078247993, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [22:00<10:16,  3.65s/it]
 68%|██████▊   | 352/520 [22:04<10:12,  3.65s/it]
                                                 
{'loss': 1.3094, 'grad_norm': 0.00403982954078973, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [22:04<10:12,  3.65s/it]
 68%|██████▊   | 353/520 [22:08<10:12,  3.67s/it]
                                                 
{'loss': 1.3177, 'grad_norm': 0.004059218528477298, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:08<10:12,  3.67s/it]
 68%|██████▊   | 354/520 [22:11<10:06,  3.66s/it]
                                                 
{'loss': 1.4555, 'grad_norm': 0.004293713506356266, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:11<10:06,  3.66s/it]
 68%|██████▊   | 355/520 [22:15<10:03,  3.65s/it]
                                                 
{'loss': 1.2464, 'grad_norm': 0.0039525103632659705, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:15<10:03,  3.65s/it]
 68%|██████▊   | 356/520 [22:18<10:00,  3.66s/it]
                                                 
{'loss': 1.2461, 'grad_norm': 0.00393831856113755, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:18<10:00,  3.66s/it]
 69%|██████▊   | 357/520 [22:22<09:55,  3.65s/it]
                                                 
{'loss': 1.2661, 'grad_norm': 0.0035796781628484966, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:22<09:55,  3.65s/it]
 69%|██████▉   | 358/520 [22:26<09:51,  3.65s/it]
                                                 
{'loss': 1.1993, 'grad_norm': 0.004120622261578449, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:26<09:51,  3.65s/it]
 69%|██████▉   | 359/520 [22:29<09:47,  3.65s/it]
                                                 
{'loss': 1.3771, 'grad_norm': 0.004438032213020061, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [22:29<09:47,  3.65s/it]
 69%|██████▉   | 360/520 [22:33<09:42,  3.64s/it]
                                                 
{'loss': 1.4167, 'grad_norm': 0.005286564364372078, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [22:33<09:42,  3.64s/it]
 69%|██████▉   | 361/520 [22:37<09:38,  3.64s/it]
                                                 
{'loss': 1.3836, 'grad_norm': 0.00409563123652336, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [22:37<09:38,  3.64s/it]
 70%|██████▉   | 362/520 [22:40<09:34,  3.63s/it]
                                                 
{'loss': 1.2613, 'grad_norm': 0.004341007753916833, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [22:40<09:34,  3.63s/it]
 70%|██████▉   | 363/520 [22:44<09:31,  3.64s/it]
                                                 
{'loss': 1.289, 'grad_norm': 0.0037784112378936863, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [22:44<09:31,  3.64s/it]
 70%|███████   | 364/520 [22:48<09:30,  3.65s/it]
                                                 
{'loss': 1.4034, 'grad_norm': 0.004266948178042626, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [22:48<09:30,  3.65s/it]
 70%|███████   | 365/520 [22:51<09:25,  3.65s/it]
                                                 
{'loss': 1.357, 'grad_norm': 0.004200658076108288, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [22:51<09:25,  3.65s/it]
 70%|███████   | 366/520 [22:55<09:21,  3.65s/it]
                                                 
{'loss': 1.3081, 'grad_norm': 0.0038877096997321663, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [22:55<09:21,  3.65s/it]
 71%|███████   | 367/520 [22:59<09:19,  3.66s/it]
                                                 
{'loss': 1.3025, 'grad_norm': 0.0038779845198781315, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [22:59<09:19,  3.66s/it]
 71%|███████   | 368/520 [23:02<09:15,  3.65s/it]
                                                 
{'loss': 1.1611, 'grad_norm': 0.004095089806231945, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [23:02<09:15,  3.65s/it]
 71%|███████   | 369/520 [23:06<09:11,  3.65s/it]
                                                 
{'loss': 1.3504, 'grad_norm': 0.003851609546869183, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:06<09:11,  3.65s/it]
 71%|███████   | 370/520 [23:10<09:07,  3.65s/it]
                                                 
{'loss': 1.2143, 'grad_norm': 0.003684634639735272, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:10<09:07,  3.65s/it]
 71%|███████▏  | 371/520 [23:13<09:06,  3.67s/it]
                                                 
{'loss': 1.2106, 'grad_norm': 0.00408831090242096, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:13<09:06,  3.67s/it]
 72%|███████▏  | 372/520 [23:17<09:02,  3.67s/it]
                                                 
{'loss': 1.4546, 'grad_norm': 0.0039046988191327138, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:17<09:02,  3.67s/it]
 72%|███████▏  | 373/520 [23:21<09:00,  3.67s/it]
                                                 
{'loss': 1.3213, 'grad_norm': 0.004362779812270388, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:21<09:00,  3.67s/it]
 72%|███████▏  | 374/520 [23:24<08:57,  3.68s/it]
                                                 
{'loss': 1.2969, 'grad_norm': 0.0040298644300388485, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:24<08:57,  3.68s/it]
 72%|███████▏  | 375/520 [23:28<08:55,  3.69s/it]
                                                 
{'loss': 1.2013, 'grad_norm': 0.0038947822229878034, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [23:28<08:55,  3.69s/it]
 72%|███████▏  | 376/520 [23:32<08:59,  3.74s/it]
                                                 
{'loss': 1.3236, 'grad_norm': 0.0036908616849408837, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [23:32<08:59,  3.74s/it]
 72%|███████▎  | 377/520 [23:36<09:00,  3.78s/it]
                                                 
{'loss': 1.2652, 'grad_norm': 0.004290324140155706, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [23:36<09:00,  3.78s/it]
 73%|███████▎  | 378/520 [23:40<08:59,  3.80s/it]
                                                 
{'loss': 1.3187, 'grad_norm': 0.0038708381128385832, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [23:40<08:59,  3.80s/it]
 73%|███████▎  | 379/520 [23:43<08:57,  3.81s/it]
                                                 
{'loss': 1.3035, 'grad_norm': 0.003805962125651013, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [23:43<08:57,  3.81s/it]
 73%|███████▎  | 380/520 [23:47<08:55,  3.82s/it]
                                                 
{'loss': 1.4323, 'grad_norm': 0.0047046858869299, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [23:47<08:55,  3.82s/it]
 73%|███████▎  | 381/520 [23:51<08:52,  3.83s/it]
                                                 
{'loss': 1.2906, 'grad_norm': 0.003990392392827848, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [23:51<08:52,  3.83s/it]
 73%|███████▎  | 382/520 [23:55<08:50,  3.84s/it]
                                                 
{'loss': 1.3655, 'grad_norm': 0.0037398569183077295, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [23:55<08:50,  3.84s/it]
 74%|███████▎  | 383/520 [23:59<08:46,  3.84s/it]
                                                 
{'loss': 1.1319, 'grad_norm': 0.0043349250570015145, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [23:59<08:46,  3.84s/it]
 74%|███████▍  | 384/520 [24:03<08:42,  3.84s/it]
                                                 
{'loss': 1.4878, 'grad_norm': 0.00425632313160007, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [24:03<08:42,  3.84s/it]
 74%|███████▍  | 385/520 [24:06<08:37,  3.83s/it]
                                                 
{'loss': 1.2732, 'grad_norm': 0.0037713428858894486, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:06<08:37,  3.83s/it]
 74%|███████▍  | 386/520 [24:10<08:32,  3.83s/it]
                                                 
{'loss': 1.2226, 'grad_norm': 0.003627224253941338, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:10<08:32,  3.83s/it]
 74%|███████▍  | 387/520 [24:14<08:29,  3.83s/it]
                                                 
{'loss': 1.4522, 'grad_norm': 0.0040969821457329525, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:14<08:29,  3.83s/it]
 75%|███████▍  | 388/520 [24:18<08:26,  3.84s/it]
                                                 
{'loss': 1.1657, 'grad_norm': 0.0038222347383357804, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:18<08:26,  3.84s/it]
 75%|███████▍  | 389/520 [24:22<08:21,  3.83s/it]
                                                 
{'loss': 1.2346, 'grad_norm': 0.004386205634294334, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}
+
 75%|███████▍  | 389/520 [24:22<08:21,  3.83s/it]
 75%|███████▌  | 390/520 [24:26<08:19,  3.84s/it]
                                                 
{'loss': 1.2909, 'grad_norm': 0.0037648825832791266, 'learning_rate': 0.031074166151605298, 'epoch': 0.75}
+
 75%|███████▌  | 390/520 [24:26<08:19,  3.84s/it]
 75%|███████▌  | 391/520 [24:29<08:15,  3.84s/it]
                                                 
{'loss': 1.3807, 'grad_norm': 0.004153134520310542, 'learning_rate': 0.030623894397145836, 'epoch': 0.75}
+
 75%|███████▌  | 391/520 [24:30<08:15,  3.84s/it]
 75%|███████▌  | 392/520 [24:33<08:12,  3.85s/it]
                                                 
{'loss': 1.1758, 'grad_norm': 0.003806944621862005, 'learning_rate': 0.03017631819139273, 'epoch': 0.75}
+
 75%|███████▌  | 392/520 [24:33<08:12,  3.85s/it]
 76%|███████▌  | 393/520 [24:37<08:07,  3.84s/it]
                                                 
{'loss': 1.2419, 'grad_norm': 0.003578574902783716, 'learning_rate': 0.029731454924533086, 'epoch': 0.76}
+
 76%|███████▌  | 393/520 [24:37<08:07,  3.84s/it]
 76%|███████▌  | 394/520 [24:41<08:03,  3.84s/it]
                                                 
{'loss': 1.2471, 'grad_norm': 0.004389934036229887, 'learning_rate': 0.029289321881345254, 'epoch': 0.76}
+
 76%|███████▌  | 394/520 [24:41<08:03,  3.84s/it]
 76%|███████▌  | 395/520 [24:45<08:00,  3.84s/it]
                                                 
{'loss': 1.2047, 'grad_norm': 0.004323706443244483, 'learning_rate': 0.028849936240527008, 'epoch': 0.76}
+
 76%|███████▌  | 395/520 [24:45<08:00,  3.84s/it]
 76%|███████▌  | 396/520 [24:49<07:52,  3.81s/it]
                                                 
{'loss': 1.2982, 'grad_norm': 0.004067672727314966, 'learning_rate': 0.028413315074028157, 'epoch': 0.76}
+
 76%|███████▌  | 396/520 [24:49<07:52,  3.81s/it]
 76%|███████▋  | 397/520 [24:52<07:44,  3.78s/it]
                                                 
{'loss': 1.2772, 'grad_norm': 0.003823026052428202, 'learning_rate': 0.027979475346387363, 'epoch': 0.76}
+
 76%|███████▋  | 397/520 [24:52<07:44,  3.78s/it]
 77%|███████▋  | 398/520 [24:56<07:36,  3.74s/it]
                                                 
{'loss': 1.2717, 'grad_norm': 0.004053324785262316, 'learning_rate': 0.027548433914072735, 'epoch': 0.77}
+
 77%|███████▋  | 398/520 [24:56<07:36,  3.74s/it]
 77%|███████▋  | 399/520 [25:00<07:30,  3.72s/it]
                                                 
{'loss': 1.3005, 'grad_norm': 0.004133180481549631, 'learning_rate': 0.027120207524827168, 'epoch': 0.77}
+
 77%|███████▋  | 399/520 [25:00<07:30,  3.72s/it]
 77%|███████▋  | 400/520 [25:03<07:24,  3.71s/it]
                                                 
{'loss': 1.3579, 'grad_norm': 0.004039480380428322, 'learning_rate': 0.02669481281701739, 'epoch': 0.77}
+
 77%|███████▋  | 400/520 [25:03<07:24,  3.71s/it]
 77%|███████▋  | 401/520 [25:07<07:22,  3.72s/it]
                                                 
{'loss': 1.0901, 'grad_norm': 0.004191613096682287, 'learning_rate': 0.026272266318987603, 'epoch': 0.77}
+
 77%|███████▋  | 401/520 [25:07<07:22,  3.72s/it]
 77%|███████▋  | 402/520 [25:11<07:23,  3.76s/it]
                                                 
{'loss': 1.213, 'grad_norm': 0.004032707006126415, 'learning_rate': 0.02585258444841733, 'epoch': 0.77}
+
 77%|███████▋  | 402/520 [25:11<07:23,  3.76s/it]
 78%|███████▊  | 403/520 [25:15<07:23,  3.79s/it]
                                                 
{'loss': 1.2467, 'grad_norm': 0.00428633090459074, 'learning_rate': 0.025435783511683442, 'epoch': 0.78}
+
 78%|███████▊  | 403/520 [25:15<07:23,  3.79s/it]
 78%|███████▊  | 404/520 [25:19<07:21,  3.81s/it]
                                                 
{'loss': 1.1515, 'grad_norm': 0.00506760025380778, 'learning_rate': 0.02502187970322657, 'epoch': 0.78}
+
 78%|███████▊  | 404/520 [25:19<07:21,  3.81s/it]
 78%|███████▊  | 405/520 [25:22<07:19,  3.82s/it]
                                                 
{'loss': 1.2952, 'grad_norm': 0.0040620303566581365, 'learning_rate': 0.02461088910492202, 'epoch': 0.78}
+
 78%|███████▊  | 405/520 [25:22<07:19,  3.82s/it]
 78%|███████▊  | 406/520 [25:26<07:15,  3.82s/it]
                                                 
{'loss': 1.2379, 'grad_norm': 0.004557878135467499, 'learning_rate': 0.02420282768545469, 'epoch': 0.78}
+
 78%|███████▊  | 406/520 [25:26<07:15,  3.82s/it]
 78%|███████▊  | 407/520 [25:30<07:12,  3.83s/it]
                                                 
{'loss': 1.3472, 'grad_norm': 0.004278289709382089, 'learning_rate': 0.02379771129969892, 'epoch': 0.78}
+
 78%|███████▊  | 407/520 [25:30<07:12,  3.83s/it]
 78%|███████▊  | 408/520 [25:34<07:06,  3.81s/it]
                                                 
{'loss': 1.2377, 'grad_norm': 0.0042540717581734935, 'learning_rate': 0.023395555688102213, 'epoch': 0.78}
+
 78%|███████▊  | 408/520 [25:34<07:06,  3.81s/it]
 79%|███████▊  | 409/520 [25:38<06:58,  3.77s/it]
                                                 
{'loss': 1.366, 'grad_norm': 0.004476344581926865, 'learning_rate': 0.02299637647607372, 'epoch': 0.79}
+
 79%|███████▊  | 409/520 [25:38<06:58,  3.77s/it]
 79%|███████▉  | 410/520 [25:41<06:51,  3.74s/it]
                                                 
{'loss': 1.0656, 'grad_norm': 0.004287967451037019, 'learning_rate': 0.022600189173377264, 'epoch': 0.79}
+
 79%|███████▉  | 410/520 [25:41<06:51,  3.74s/it]
 79%|███████▉  | 411/520 [25:45<06:45,  3.72s/it]
                                                 
{'loss': 1.3357, 'grad_norm': 0.004223817129296374, 'learning_rate': 0.022207009173528525, 'epoch': 0.79}
+
 79%|███████▉  | 411/520 [25:45<06:45,  3.72s/it]
 79%|███████▉  | 412/520 [25:49<06:40,  3.70s/it]
                                                 
{'loss': 1.2467, 'grad_norm': 0.00396001157285675, 'learning_rate': 0.02181685175319702, 'epoch': 0.79}
+
 79%|███████▉  | 412/520 [25:49<06:40,  3.70s/it]
 79%|███████▉  | 413/520 [25:52<06:34,  3.69s/it]
                                                 
{'loss': 1.3364, 'grad_norm': 0.003940531039145668, 'learning_rate': 0.021429732071612653, 'epoch': 0.79}
+
 79%|███████▉  | 413/520 [25:52<06:34,  3.69s/it]
 80%|███████▉  | 414/520 [25:56<06:31,  3.69s/it]
                                                 
{'loss': 1.1152, 'grad_norm': 0.0036465800613090624, 'learning_rate': 0.02104566516997647, 'epoch': 0.8}
+
 80%|███████▉  | 414/520 [25:56<06:31,  3.69s/it]
 80%|███████▉  | 415/520 [26:00<06:26,  3.68s/it]
                                                 
{'loss': 1.2233, 'grad_norm': 0.0038730329613674316, 'learning_rate': 0.020664665970876496, 'epoch': 0.8}
+
 80%|███████▉  | 415/520 [26:00<06:26,  3.68s/it]
 80%|████████  | 416/520 [26:03<06:23,  3.69s/it]
                                                 
{'loss': 1.1324, 'grad_norm': 0.004567385002553746, 'learning_rate': 0.020286749277707784, 'epoch': 0.8}
+
 80%|████████  | 416/520 [26:03<06:23,  3.69s/it]
 80%|████████  | 417/520 [26:07<06:18,  3.68s/it]
                                                 
{'loss': 1.3117, 'grad_norm': 0.004616838741130314, 'learning_rate': 0.019911929774097215, 'epoch': 0.8}
+
 80%|████████  | 417/520 [26:07<06:18,  3.68s/it]
 80%|████████  | 418/520 [26:11<06:14,  3.68s/it]
                                                 
{'loss': 1.2907, 'grad_norm': 0.0037450832170694715, 'learning_rate': 0.019540222023333165, 'epoch': 0.8}
+
 80%|████████  | 418/520 [26:11<06:14,  3.68s/it]
 81%|████████  | 419/520 [26:14<06:11,  3.67s/it]
                                                 
{'loss': 1.2795, 'grad_norm': 0.004304990225116143, 'learning_rate': 0.01917164046779948, 'epoch': 0.81}
+
 81%|████████  | 419/520 [26:14<06:11,  3.67s/it]
 81%|████████  | 420/520 [26:18<06:08,  3.68s/it]
                                                 
{'loss': 1.1556, 'grad_norm': 0.004156821883082135, 'learning_rate': 0.018806199428414352, 'epoch': 0.81}
+
 81%|████████  | 420/520 [26:18<06:08,  3.68s/it]
 81%|████████  | 421/520 [26:22<06:03,  3.68s/it]
                                                 
{'loss': 1.085, 'grad_norm': 0.004111388539909275, 'learning_rate': 0.018443913104073985, 'epoch': 0.81}
+
 81%|████████  | 421/520 [26:22<06:03,  3.68s/it]
 81%|████████  | 422/520 [26:25<06:00,  3.68s/it]
                                                 
{'loss': 1.2155, 'grad_norm': 0.004313174026675046, 'learning_rate': 0.01808479557110081, 'epoch': 0.81}
+
 81%|████████  | 422/520 [26:25<06:00,  3.68s/it]
 81%|████████▏ | 423/520 [26:29<05:56,  3.68s/it]
                                                 
{'loss': 1.2137, 'grad_norm': 0.004772252796929866, 'learning_rate': 0.017728860782696667, 'epoch': 0.81}
+
 81%|████████▏ | 423/520 [26:29<05:56,  3.68s/it]
 82%|████████▏ | 424/520 [26:33<05:52,  3.67s/it]
                                                 
{'loss': 1.415, 'grad_norm': 0.0044204571068566295, 'learning_rate': 0.017376122568400532, 'epoch': 0.82}
+
 82%|████████▏ | 424/520 [26:33<05:52,  3.67s/it]
 82%|████████▏ | 425/520 [26:36<05:48,  3.67s/it]
                                                 
{'loss': 1.2156, 'grad_norm': 0.0038410826559133896, 'learning_rate': 0.017026594633551252, 'epoch': 0.82}
+
 82%|████████▏ | 425/520 [26:36<05:48,  3.67s/it]
 82%|████████▏ | 426/520 [26:40<05:44,  3.67s/it]
                                                 
{'loss': 1.2382, 'grad_norm': 0.0055208414841711574, 'learning_rate': 0.01668029055875512, 'epoch': 0.82}
+
 82%|████████▏ | 426/520 [26:40<05:44,  3.67s/it]
 82%|████████▏ | 427/520 [26:44<05:40,  3.66s/it]
                                                 
{'loss': 1.1523, 'grad_norm': 0.004238733846153601, 'learning_rate': 0.016337223799358026, 'epoch': 0.82}
+
 82%|████████▏ | 427/520 [26:44<05:40,  3.66s/it]
 82%|████████▏ | 428/520 [26:47<05:36,  3.66s/it]
                                                 
{'loss': 1.1165, 'grad_norm': 0.004007997775099366, 'learning_rate': 0.01599740768492286, 'epoch': 0.82}
+
 82%|████████▏ | 428/520 [26:47<05:36,  3.66s/it]
 82%|████████▎ | 429/520 [26:51<05:33,  3.66s/it]
                                                 
{'loss': 1.2241, 'grad_norm': 0.003801295882790296, 'learning_rate': 0.015660855418711452, 'epoch': 0.82}
+
 82%|████████▎ | 429/520 [26:51<05:33,  3.66s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2076 > 2048). Running this sequence through the model will result in indexing errors
+
 83%|████████▎ | 430/520 [26:55<05:29,  3.66s/it]
                                                 
{'loss': 1.2174, 'grad_norm': 0.0036256169694414693, 'learning_rate': 0.015327580077171589, 'epoch': 0.83}
+
 83%|████████▎ | 430/520 [26:55<05:29,  3.66s/it]
 83%|████████▎ | 431/520 [26:58<05:31,  3.72s/it]
                                                 
{'loss': 1.2902, 'grad_norm': 0.004329489669106471, 'learning_rate': 0.014997594609429088, 'epoch': 0.83}
+
 83%|████████▎ | 431/520 [26:58<05:31,  3.72s/it]
 83%|████████▎ | 432/520 [27:02<05:30,  3.75s/it]
                                                 
{'loss': 1.1274, 'grad_norm': 0.004590112073026319, 'learning_rate': 0.01467091183678444, 'epoch': 0.83}
+
 83%|████████▎ | 432/520 [27:02<05:30,  3.75s/it]
 83%|████████▎ | 433/520 [27:06<05:24,  3.73s/it]
                                                 
{'loss': 1.2678, 'grad_norm': 0.004103173484125776, 'learning_rate': 0.014347544452214867, 'epoch': 0.83}
+
 83%|████████▎ | 433/520 [27:06<05:24,  3.73s/it]
 83%|████████▎ | 434/520 [27:10<05:19,  3.71s/it]
                                                 
{'loss': 1.0015, 'grad_norm': 0.003813458712727371, 'learning_rate': 0.014027505019880971, 'epoch': 0.83}
+
 83%|████████▎ | 434/520 [27:10<05:19,  3.71s/it]
 84%|████████▎ | 435/520 [27:13<05:13,  3.69s/it]
                                                 
{'loss': 1.3024, 'grad_norm': 0.0045580842468441775, 'learning_rate': 0.013710805974638696, 'epoch': 0.84}
+
 84%|████████▎ | 435/520 [27:13<05:13,  3.69s/it]
 84%|████████▍ | 436/520 [27:17<05:08,  3.68s/it]
                                                 
{'loss': 1.0912, 'grad_norm': 0.003994821507076144, 'learning_rate': 0.01339745962155613, 'epoch': 0.84}
+
 84%|████████▍ | 436/520 [27:17<05:08,  3.68s/it]
 84%|████████▍ | 437/520 [27:21<05:04,  3.67s/it]
                                                 
{'loss': 1.3327, 'grad_norm': 0.0039371168302820125, 'learning_rate': 0.01308747813543536, 'epoch': 0.84}
+
 84%|████████▍ | 437/520 [27:21<05:04,  3.67s/it]
 84%|████████▍ | 438/520 [27:24<05:01,  3.67s/it]
                                                 
{'loss': 1.123, 'grad_norm': 0.003784659087073013, 'learning_rate': 0.012780873560339467, 'epoch': 0.84}
+
 84%|████████▍ | 438/520 [27:24<05:01,  3.67s/it]
 84%|████████▍ | 439/520 [27:28<04:57,  3.67s/it]
                                                 
{'loss': 1.2583, 'grad_norm': 0.003501580660529986, 'learning_rate': 0.012477657809124632, 'epoch': 0.84}
+
 84%|████████▍ | 439/520 [27:28<04:57,  3.67s/it]
 85%|████████▍ | 440/520 [27:32<04:53,  3.67s/it]
                                                 
{'loss': 1.184, 'grad_norm': 0.003953537041671082, 'learning_rate': 0.012177842662977134, 'epoch': 0.85}
+
 85%|████████▍ | 440/520 [27:32<04:53,  3.67s/it]
 85%|████████▍ | 441/520 [27:35<04:50,  3.67s/it]
                                                 
{'loss': 1.291, 'grad_norm': 0.004022798178452184, 'learning_rate': 0.01188143977095576, 'epoch': 0.85}
+
 85%|████████▍ | 441/520 [27:35<04:50,  3.67s/it]
 85%|████████▌ | 442/520 [27:39<04:46,  3.67s/it]
                                                 
{'loss': 1.2401, 'grad_norm': 0.00462015188528124, 'learning_rate': 0.011588460649539035, 'epoch': 0.85}
+
 85%|████████▌ | 442/520 [27:39<04:46,  3.67s/it]
 85%|████████▌ | 443/520 [27:43<04:42,  3.67s/it]
                                                 
{'loss': 1.2619, 'grad_norm': 0.0042525789481886655, 'learning_rate': 0.011298916682177829, 'epoch': 0.85}
+
 85%|████████▌ | 443/520 [27:43<04:42,  3.67s/it]
 85%|████████▌ | 444/520 [27:46<04:39,  3.68s/it]
                                                 
{'loss': 1.2241, 'grad_norm': 0.0035970687177349324, 'learning_rate': 0.011012819118853146, 'epoch': 0.85}
+
 85%|████████▌ | 444/520 [27:46<04:39,  3.68s/it]
 86%|████████▌ | 445/520 [27:50<04:35,  3.68s/it]
                                                 
{'loss': 1.1434, 'grad_norm': 0.003956041364666467, 'learning_rate': 0.01073017907563887, 'epoch': 0.86}
+
 86%|████████▌ | 445/520 [27:50<04:35,  3.68s/it]
 86%|████████▌ | 446/520 [27:54<04:32,  3.68s/it]
                                                 
{'loss': 1.3713, 'grad_norm': 0.00390232369647536, 'learning_rate': 0.010451007534269908, 'epoch': 0.86}
+
 86%|████████▌ | 446/520 [27:54<04:32,  3.68s/it]
 86%|████████▌ | 447/520 [27:57<04:28,  3.68s/it]
                                                 
{'loss': 1.2414, 'grad_norm': 0.004087448886119014, 'learning_rate': 0.010175315341715598, 'epoch': 0.86}
+
 86%|████████▌ | 447/520 [27:57<04:28,  3.68s/it]
 86%|████████▌ | 448/520 [28:01<04:24,  3.68s/it]
                                                 
{'loss': 1.2113, 'grad_norm': 0.0045372109331261136, 'learning_rate': 0.009903113209758098, 'epoch': 0.86}
+
 86%|████████▌ | 448/520 [28:01<04:24,  3.68s/it]
 86%|████████▋ | 449/520 [28:05<04:21,  3.69s/it]
                                                 
{'loss': 1.3227, 'grad_norm': 0.004085140224445111, 'learning_rate': 0.009634411714576352, 'epoch': 0.86}
+
 86%|████████▋ | 449/520 [28:05<04:21,  3.69s/it]
 87%|████████▋ | 450/520 [28:08<04:17,  3.67s/it]
                                                 
{'loss': 1.2593, 'grad_norm': 0.003947994248801975, 'learning_rate': 0.009369221296335007, 'epoch': 0.87}
+
 87%|████████▋ | 450/520 [28:08<04:17,  3.67s/it]
 87%|████████▋ | 451/520 [28:12<04:13,  3.67s/it]
                                                 
{'loss': 1.2466, 'grad_norm': 0.004082120200746635, 'learning_rate': 0.009107552258778906, 'epoch': 0.87}
+
 87%|████████▋ | 451/520 [28:12<04:13,  3.67s/it]
 87%|████████▋ | 452/520 [28:16<04:12,  3.72s/it]
                                                 
{'loss': 1.3618, 'grad_norm': 0.004154579266883381, 'learning_rate': 0.008849414768832687, 'epoch': 0.87}
+
 87%|████████▋ | 452/520 [28:16<04:12,  3.72s/it]
 87%|████████▋ | 453/520 [28:20<04:11,  3.75s/it]
                                                 
{'loss': 1.3402, 'grad_norm': 0.004126077151894212, 'learning_rate': 0.008594818856205699, 'epoch': 0.87}
+
 87%|████████▋ | 453/520 [28:20<04:11,  3.75s/it]
 87%|████████▋ | 454/520 [28:24<04:09,  3.78s/it]
                                                 
{'loss': 1.1622, 'grad_norm': 0.004284135927069758, 'learning_rate': 0.00834377441300238, 'epoch': 0.87}
+
 87%|████████▋ | 454/520 [28:24<04:09,  3.78s/it]
 88%|████████▊ | 455/520 [28:27<04:07,  3.82s/it]
                                                 
{'loss': 1.2929, 'grad_norm': 0.0039006796572652, 'learning_rate': 0.008096291193337934, 'epoch': 0.88}
+
 88%|████████▊ | 455/520 [28:27<04:07,  3.82s/it]
 88%|████████▊ | 456/520 [28:31<04:04,  3.82s/it]
                                                 
{'loss': 1.2113, 'grad_norm': 0.004745071254409722, 'learning_rate': 0.007852378812959226, 'epoch': 0.88}
+
 88%|████████▊ | 456/520 [28:31<04:04,  3.82s/it]
 88%|████████▊ | 457/520 [28:35<04:01,  3.83s/it]
                                                 
{'loss': 1.3145, 'grad_norm': 0.004220289857596339, 'learning_rate': 0.007612046748871327, 'epoch': 0.88}
+
 88%|████████▊ | 457/520 [28:35<04:01,  3.83s/it]
 88%|████████▊ | 458/520 [28:39<03:57,  3.83s/it]
                                                 
{'loss': 1.3566, 'grad_norm': 0.004050085585657556, 'learning_rate': 0.007375304338969136, 'epoch': 0.88}
+
 88%|████████▊ | 458/520 [28:39<03:57,  3.83s/it]
 88%|████████▊ | 459/520 [28:43<03:54,  3.84s/it]
                                                 
{'loss': 1.2866, 'grad_norm': 0.003788297200803749, 'learning_rate': 0.007142160781674645, 'epoch': 0.88}
+
 88%|████████▊ | 459/520 [28:43<03:54,  3.84s/it]
 88%|████████▊ | 460/520 [28:47<03:50,  3.84s/it]
                                                 
{'loss': 1.1543, 'grad_norm': 0.003766725751261823, 'learning_rate': 0.006912625135579587, 'epoch': 0.88}
+
 88%|████████▊ | 460/520 [28:47<03:50,  3.84s/it]
 89%|████████▊ | 461/520 [28:51<03:46,  3.84s/it]
                                                 
{'loss': 1.4005, 'grad_norm': 0.0038602557933787736, 'learning_rate': 0.0066867063190933496, 'epoch': 0.89}
+
 89%|████████▊ | 461/520 [28:51<03:46,  3.84s/it]
 89%|████████▉ | 462/520 [28:54<03:42,  3.83s/it]
                                                 
{'loss': 1.4109, 'grad_norm': 0.003938442538007118, 'learning_rate': 0.006464413110096601, 'epoch': 0.89}
+
 89%|████████▉ | 462/520 [28:54<03:42,  3.83s/it]
 89%|████████▉ | 463/520 [28:58<03:38,  3.83s/it]
                                                 
{'loss': 1.1213, 'grad_norm': 0.004213435158222209, 'learning_rate': 0.006245754145600091, 'epoch': 0.89}
+
 89%|████████▉ | 463/520 [28:58<03:38,  3.83s/it]
 89%|████████▉ | 464/520 [29:02<03:33,  3.81s/it]
                                                 
{'loss': 1.2695, 'grad_norm': 0.00421723625380015, 'learning_rate': 0.006030737921409169, 'epoch': 0.89}
+
 89%|████████▉ | 464/520 [29:02<03:33,  3.81s/it]
 89%|████████▉ | 465/520 [29:06<03:26,  3.76s/it]
                                                 
{'loss': 1.3911, 'grad_norm': 0.0045011584942024774, 'learning_rate': 0.005819372791793654, 'epoch': 0.89}
+
 89%|████████▉ | 465/520 [29:06<03:26,  3.76s/it]
 90%|████████▉ | 466/520 [29:09<03:21,  3.73s/it]
                                                 
{'loss': 1.252, 'grad_norm': 0.003476247221234463, 'learning_rate': 0.005611666969163243, 'epoch': 0.9}
+
 90%|████████▉ | 466/520 [29:09<03:21,  3.73s/it]
 90%|████████▉ | 467/520 [29:13<03:16,  3.71s/it]
                                                 
{'loss': 1.2919, 'grad_norm': 0.003953986957132499, 'learning_rate': 0.005407628523748398, 'epoch': 0.9}
+
 90%|████████▉ | 467/520 [29:13<03:16,  3.71s/it]
 90%|█████████ | 468/520 [29:17<03:12,  3.71s/it]
                                                 
{'loss': 1.2368, 'grad_norm': 0.00451015755800606, 'learning_rate': 0.00520726538328683, 'epoch': 0.9}
+
 90%|█████████ | 468/520 [29:17<03:12,  3.71s/it]
 90%|█████████ | 469/520 [29:20<03:08,  3.69s/it]
                                                 
{'loss': 1.2851, 'grad_norm': 0.004223030401918961, 'learning_rate': 0.005010585332715401, 'epoch': 0.9}
+
 90%|█████████ | 469/520 [29:20<03:08,  3.69s/it]
 90%|█████████ | 470/520 [29:24<03:04,  3.68s/it]
                                                 
{'loss': 1.1618, 'grad_norm': 0.003510929411902853, 'learning_rate': 0.004817596013867765, 'epoch': 0.9}
+
 90%|█████████ | 470/520 [29:24<03:04,  3.68s/it]
 91%|█████████ | 471/520 [29:28<03:00,  3.69s/it]
                                                 
{'loss': 1.178, 'grad_norm': 0.004019523010712186, 'learning_rate': 0.004628304925177318, 'epoch': 0.91}
+
 91%|█████████ | 471/520 [29:28<03:00,  3.69s/it]
 91%|█████████ | 472/520 [29:31<02:59,  3.73s/it]
                                                 
{'loss': 1.1543, 'grad_norm': 0.003737121194398204, 'learning_rate': 0.004442719421385921, 'epoch': 0.91}
+
 91%|█████████ | 472/520 [29:31<02:59,  3.73s/it]
 91%|█████████ | 473/520 [29:35<02:55,  3.73s/it]
                                                 
{'loss': 1.2087, 'grad_norm': 0.004129736421310799, 'learning_rate': 0.004260846713258193, 'epoch': 0.91}
+
 91%|█████████ | 473/520 [29:35<02:55,  3.73s/it]
 91%|█████████ | 474/520 [29:39<02:52,  3.75s/it]
                                                 
{'loss': 1.3206, 'grad_norm': 0.003871099889301877, 'learning_rate': 0.004082693867301224, 'epoch': 0.91}
+
 91%|█████████ | 474/520 [29:39<02:52,  3.75s/it]
 91%|█████████▏| 475/520 [29:43<02:48,  3.75s/it]
                                                 
{'loss': 1.2352, 'grad_norm': 0.0037256533749625747, 'learning_rate': 0.003908267805490051, 'epoch': 0.91}
+
 91%|█████████▏| 475/520 [29:43<02:48,  3.75s/it]
 92%|█████████▏| 476/520 [29:47<02:46,  3.79s/it]
                                                 
{'loss': 1.21, 'grad_norm': 0.004339179095912478, 'learning_rate': 0.003737575304998797, 'epoch': 0.92}
+
 92%|█████████▏| 476/520 [29:47<02:46,  3.79s/it]
 92%|█████████▏| 477/520 [29:50<02:43,  3.80s/it]
                                                 
{'loss': 1.196, 'grad_norm': 0.004595362386606664, 'learning_rate': 0.003570622997937234, 'epoch': 0.92}
+
 92%|█████████▏| 477/520 [29:50<02:43,  3.80s/it]
 92%|█████████▏| 478/520 [29:54<02:39,  3.79s/it]
                                                 
{'loss': 1.1507, 'grad_norm': 0.0038924472558954297, 'learning_rate': 0.00340741737109318, 'epoch': 0.92}
+
 92%|█████████▏| 478/520 [29:54<02:39,  3.79s/it]
 92%|█████████▏| 479/520 [29:58<02:34,  3.76s/it]
                                                 
{'loss': 1.2953, 'grad_norm': 0.00422665216072056, 'learning_rate': 0.003247964765680389, 'epoch': 0.92}
+
 92%|█████████▏| 479/520 [29:58<02:34,  3.76s/it]
 92%|█████████▏| 480/520 [30:02<02:29,  3.73s/it]
                                                 
{'loss': 1.3135, 'grad_norm': 0.0041076997710668235, 'learning_rate': 0.0030922713770922153, 'epoch': 0.92}
+
 92%|█████████▏| 480/520 [30:02<02:29,  3.73s/it]
 92%|█████████▎| 481/520 [30:05<02:25,  3.74s/it]
                                                 
{'loss': 1.3283, 'grad_norm': 0.0036620300858139477, 'learning_rate': 0.0029403432546609046, 'epoch': 0.93}
+
 92%|█████████▎| 481/520 [30:05<02:25,  3.74s/it]
 93%|█████████▎| 482/520 [30:09<02:22,  3.76s/it]
                                                 
{'loss': 1.3332, 'grad_norm': 0.004328119902283388, 'learning_rate': 0.0027921863014225504, 'epoch': 0.93}
+
 93%|█████████▎| 482/520 [30:09<02:22,  3.76s/it]
 93%|█████████▎| 483/520 [30:13<02:18,  3.74s/it]
                                                 
{'loss': 1.2308, 'grad_norm': 0.004386083581336805, 'learning_rate': 0.002647806273887665, 'epoch': 0.93}
+
 93%|█████████▎| 483/520 [30:13<02:18,  3.74s/it]
 93%|█████████▎| 484/520 [30:16<02:13,  3.72s/it]
                                                 
{'loss': 1.2318, 'grad_norm': 0.003988879975815538, 'learning_rate': 0.0025072087818176383, 'epoch': 0.93}
+
 93%|█████████▎| 484/520 [30:16<02:13,  3.72s/it]
 93%|█████████▎| 485/520 [30:20<02:10,  3.72s/it]
                                                 
{'loss': 1.1732, 'grad_norm': 0.0037581091020006964, 'learning_rate': 0.002370399288006664, 'epoch': 0.93}
+
 93%|█████████▎| 485/520 [30:20<02:10,  3.72s/it]
 93%|█████████▎| 486/520 [30:24<02:06,  3.73s/it]
                                                 
{'loss': 1.3062, 'grad_norm': 0.004373795679535535, 'learning_rate': 0.0022373831080695463, 'epoch': 0.93}
+
 93%|█████████▎| 486/520 [30:24<02:06,  3.73s/it]
 94%|█████████▎| 487/520 [30:28<02:03,  3.74s/it]
                                                 
{'loss': 1.1503, 'grad_norm': 0.003699413036525545, 'learning_rate': 0.0021081654102351635, 'epoch': 0.94}
+
 94%|█████████▎| 487/520 [30:28<02:03,  3.74s/it]
 94%|█████████▍| 488/520 [30:32<02:00,  3.78s/it]
                                                 
{'loss': 1.0928, 'grad_norm': 0.004017437505730763, 'learning_rate': 0.0019827512151456175, 'epoch': 0.94}
+
 94%|█████████▍| 488/520 [30:32<02:00,  3.78s/it]
 94%|█████████▍| 489/520 [30:35<01:56,  3.76s/it]
                                                 
{'loss': 1.3181, 'grad_norm': 0.003434977234504906, 'learning_rate': 0.0018611453956612345, 'epoch': 0.94}
+
 94%|█████████▍| 489/520 [30:35<01:56,  3.76s/it]
 94%|█████████▍| 490/520 [30:39<01:51,  3.73s/it]
                                                 
{'loss': 1.2175, 'grad_norm': 0.003999426278148538, 'learning_rate': 0.0017433526766711727, 'epoch': 0.94}
+
 94%|█████████▍| 490/520 [30:39<01:51,  3.73s/it]
 94%|█████████▍| 491/520 [30:43<01:48,  3.73s/it]
                                                 
{'loss': 1.1775, 'grad_norm': 0.004061447679007399, 'learning_rate': 0.0016293776349098677, 'epoch': 0.94}
+
 94%|█████████▍| 491/520 [30:43<01:48,  3.73s/it]
 95%|█████████▍| 492/520 [30:47<01:45,  3.76s/it]
                                                 
{'loss': 1.301, 'grad_norm': 0.004022787749846593, 'learning_rate': 0.0015192246987791981, 'epoch': 0.95}
+
 95%|█████████▍| 492/520 [30:47<01:45,  3.76s/it]
 95%|█████████▍| 493/520 [30:50<01:40,  3.73s/it]
                                                 
{'loss': 1.3692, 'grad_norm': 0.004431011824841104, 'learning_rate': 0.0014128981481764114, 'epoch': 0.95}
+
 95%|█████████▍| 493/520 [30:50<01:40,  3.73s/it]
 95%|█████████▌| 494/520 [30:54<01:36,  3.71s/it]
                                                 
{'loss': 1.241, 'grad_norm': 0.0036064818632825226, 'learning_rate': 0.0013104021143278911, 'epoch': 0.95}
+
 95%|█████████▌| 494/520 [30:54<01:36,  3.71s/it]
 95%|█████████▌| 495/520 [30:58<01:32,  3.71s/it]
                                                 
{'loss': 1.1872, 'grad_norm': 0.004106760859137228, 'learning_rate': 0.0012117405796285285, 'epoch': 0.95}
+
 95%|█████████▌| 495/520 [30:58<01:32,  3.71s/it]
 95%|█████████▌| 496/520 [31:01<01:29,  3.75s/it]
                                                 
{'loss': 1.1122, 'grad_norm': 0.003924433528982079, 'learning_rate': 0.0011169173774871477, 'epoch': 0.95}
+
 95%|█████████▌| 496/520 [31:01<01:29,  3.75s/it]
 96%|█████████▌| 497/520 [31:05<01:26,  3.76s/it]
                                                 
{'loss': 1.2488, 'grad_norm': 0.003466845089839699, 'learning_rate': 0.0010259361921774012, 'epoch': 0.96}
+
 96%|█████████▌| 497/520 [31:05<01:26,  3.76s/it]
 96%|█████████▌| 498/520 [31:09<01:22,  3.77s/it]
                                                 
{'loss': 1.1996, 'grad_norm': 0.004275813713710447, 'learning_rate': 0.000938800558694719, 'epoch': 0.96}
+
 96%|█████████▌| 498/520 [31:09<01:22,  3.77s/it]
 96%|█████████▌| 499/520 [31:13<01:18,  3.75s/it]
                                                 
{'loss': 1.3905, 'grad_norm': 0.003924739279973648, 'learning_rate': 0.0008555138626189618, 'epoch': 0.96}
+
 96%|█████████▌| 499/520 [31:13<01:18,  3.75s/it]
 96%|█████████▌| 500/520 [31:17<01:15,  3.78s/it]
                                                 
{'loss': 1.3161, 'grad_norm': 0.0046802058275385435, 'learning_rate': 0.0007760793399827937, 'epoch': 0.96}
+
 96%|█████████▌| 500/520 [31:17<01:15,  3.78s/it]
 96%|█████████▋| 501/520 [31:20<01:11,  3.77s/it]
                                                 
{'loss': 1.3502, 'grad_norm': 0.006263181910580194, 'learning_rate': 0.000700500077146038, 'epoch': 0.96}
+
 96%|█████████▋| 501/520 [31:20<01:11,  3.77s/it]
 97%|█████████▋| 502/520 [31:24<01:07,  3.77s/it]
                                                 
{'loss': 1.2361, 'grad_norm': 0.003767799116880923, 'learning_rate': 0.0006287790106757397, 'epoch': 0.97}
+
 97%|█████████▋| 502/520 [31:24<01:07,  3.77s/it]
 97%|█████████▋| 503/520 [31:28<01:03,  3.76s/it]
                                                 
{'loss': 1.28, 'grad_norm': 0.004040248950173226, 'learning_rate': 0.0005609189272320237, 'epoch': 0.97}
+
 97%|█████████▋| 503/520 [31:28<01:03,  3.76s/it]
 97%|█████████▋| 504/520 [31:32<01:00,  3.76s/it]
                                                 
{'loss': 1.2291, 'grad_norm': 0.004391594048389926, 'learning_rate': 0.000496922463459859, 'epoch': 0.97}
+
 97%|█████████▋| 504/520 [31:32<01:00,  3.76s/it]
 97%|█████████▋| 505/520 [31:35<00:56,  3.74s/it]
                                                 
{'loss': 1.2761, 'grad_norm': 0.003839515301905443, 'learning_rate': 0.0004367921058866187, 'epoch': 0.97}
+
 97%|█████████▋| 505/520 [31:35<00:56,  3.74s/it]
 97%|█████████▋| 506/520 [31:39<00:52,  3.77s/it]
                                                 
{'loss': 1.1814, 'grad_norm': 0.0043129503154115435, 'learning_rate': 0.0003805301908254455, 'epoch': 0.97}
+
 97%|█████████▋| 506/520 [31:39<00:52,  3.77s/it]
 98%|█████████▊| 507/520 [31:43<00:49,  3.78s/it]
                                                 
{'loss': 1.4358, 'grad_norm': 0.0037810917371253363, 'learning_rate': 0.0003281389042844918, 'epoch': 0.97}
+
 98%|█████████▊| 507/520 [31:43<00:49,  3.78s/it]
 98%|█████████▊| 508/520 [31:47<00:45,  3.76s/it]
                                                 
{'loss': 1.3035, 'grad_norm': 0.003864152190954269, 'learning_rate': 0.00027962028188198705, 'epoch': 0.98}
+
 98%|█████████▊| 508/520 [31:47<00:45,  3.76s/it]
 98%|█████████▊| 509/520 [31:50<00:41,  3.73s/it]
                                                 
{'loss': 1.2722, 'grad_norm': 0.0038038582948963237, 'learning_rate': 0.00023497620876711256, 'epoch': 0.98}
+
 98%|█████████▊| 509/520 [31:50<00:41,  3.73s/it]
 98%|█████████▊| 510/520 [31:54<00:37,  3.70s/it]
                                                 
{'loss': 1.229, 'grad_norm': 0.0038023256137826636, 'learning_rate': 0.00019420841954681523, 'epoch': 0.98}
+
 98%|█████████▊| 510/520 [31:54<00:37,  3.70s/it]
 98%|█████████▊| 511/520 [31:58<00:33,  3.72s/it]
                                                 
{'loss': 1.1999, 'grad_norm': 0.003701437567140411, 'learning_rate': 0.00015731849821833956, 'epoch': 0.98}
+
 98%|█████████▊| 511/520 [31:58<00:33,  3.72s/it]
 98%|█████████▊| 512/520 [32:02<00:30,  3.76s/it]
                                                 
{'loss': 1.0776, 'grad_norm': 0.003745512566499385, 'learning_rate': 0.00012430787810776555, 'epoch': 0.98}
+
 98%|█████████▊| 512/520 [32:02<00:30,  3.76s/it]
 99%|█████████▊| 513/520 [32:05<00:26,  3.80s/it]
                                                 
{'loss': 1.2909, 'grad_norm': 0.004318781946368065, 'learning_rate': 9.517784181422018e-05, 'epoch': 0.99}
+
 99%|█████████▊| 513/520 [32:05<00:26,  3.80s/it]
 99%|█████████▉| 514/520 [32:09<00:22,  3.80s/it]
                                                 
{'loss': 1.262, 'grad_norm': 0.003542641040867754, 'learning_rate': 6.992952116013917e-05, 'epoch': 0.99}
+
 99%|█████████▉| 514/520 [32:09<00:22,  3.80s/it]
 99%|█████████▉| 515/520 [32:13<00:19,  3.82s/it]
                                                 
{'loss': 1.3154, 'grad_norm': 0.004541553708791112, 'learning_rate': 4.856389714723575e-05, 'epoch': 0.99}
+
 99%|█████████▉| 515/520 [32:13<00:19,  3.82s/it]
 99%|█████████▉| 516/520 [32:17<00:15,  3.83s/it]
                                                 
{'loss': 1.1947, 'grad_norm': 0.0038679004375289433, 'learning_rate': 3.108179991837545e-05, 'epoch': 0.99}
+
 99%|█████████▉| 516/520 [32:17<00:15,  3.83s/it]
 99%|█████████▉| 517/520 [32:21<00:11,  3.79s/it]
                                                 
{'loss': 1.3553, 'grad_norm': 0.004004921392574511, 'learning_rate': 1.7483908725357544e-05, 'epoch': 0.99}
+
 99%|█████████▉| 517/520 [32:21<00:11,  3.79s/it]
100%|█████████▉| 518/520 [32:24<00:07,  3.73s/it]
                                                 
{'loss': 1.2158, 'grad_norm': 0.004097430171333922, 'learning_rate': 7.770751902513862e-06, 'epoch': 1.0}
+
100%|█████████▉| 518/520 [32:24<00:07,  3.73s/it]
100%|█████████▉| 519/520 [32:28<00:03,  3.69s/it]
                                                 
{'loss': 1.2831, 'grad_norm': 0.0040459082939723135, 'learning_rate': 1.9427068461808086e-06, 'epoch': 1.0}
+
100%|█████████▉| 519/520 [32:28<00:03,  3.69s/it]
100%|██████████| 520/520 [32:32<00:00,  3.93s/it]
                                                 
{'loss': 1.3676, 'grad_norm': 0.004177841653423459, 'learning_rate': 0.0, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:32<00:00,  3.93s/it]
                                                 
{'train_runtime': 1952.8335, 'train_samples_per_second': 34.068, 'train_steps_per_second': 0.266, 'train_loss': 1.4982106385322718, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:32<00:00,  3.93s/it]
100%|██████████| 520/520 [32:32<00:00,  3.76s/it]
+[2025-10-16 06:57:48,007] [INFO] [launch.py:348:main] Process 2450667 exits successfully.
+[2025-10-16 06:57:49,009] [INFO] [launch.py:348:main] Process 2450666 exits successfully.
+[2025-10-16 06:57:49,010] [INFO] [launch.py:348:main] Process 2450665 exits successfully.
+[2025-10-16 06:57:49,010] [INFO] [launch.py:348:main] Process 2450664 exits successfully.
+[2025-10-16 06:57:49,011] [INFO] [launch.py:348:main] Process 2450669 exits successfully.
+[2025-10-16 06:57:50,012] [INFO] [launch.py:348:main] Process 2450668 exits successfully.
+[2025-10-16 06:57:50,013] [INFO] [launch.py:348:main] Process 2450663 exits successfully.
+[2025-10-16 06:57:53,017] [INFO] [launch.py:348:main] Process 2450662 exits successfully.
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_1.7_2e-1_connector-1.0_1.7_2e-1_ablation_20251016_062341.log
+Timestamp: 2025-10-16 06:57:55
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation_20251016_065755.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation_20251016_065755.log
new file mode 100644
index 0000000000000000000000000000000000000000..07d604c823715b6ed84f559fe34f9fd12a1db9b6
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation_20251016_065755.log
@@ -0,0 +1,2235 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation_20251016_065755.log
+Timestamp: 2025-10-16 06:57:55
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 06:57:58,324] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:01,034] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 06:58:01,035] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 1.0 --temperature_attn_text 2.3 --temperature_mlp_text 2.3 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 1.0 --temperature_attn_vision 2.3 --temperature_mlp_vision 2.3 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 1.0 --temperature_connector 2.3 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 06:58:03,612] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:04,654] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 06:58:04,654] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 06:58:04,655] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 06:58:04,655] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 06:58:04,655] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 06:58:04,655] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 06:58:04,655] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 06:58:04,657] [INFO] [launch.py:253:main] process 2472101 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,659] [INFO] [launch.py:253:main] process 2472102 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,661] [INFO] [launch.py:253:main] process 2472103 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,662] [INFO] [launch.py:253:main] process 2472104 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,664] [INFO] [launch.py:253:main] process 2472105 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,666] [INFO] [launch.py:253:main] process 2472106 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,668] [INFO] [launch.py:253:main] process 2472107 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 06:58:04,670] [INFO] [launch.py:253:main] process 2472108 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 06:58:11,358] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,473] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,687] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,690] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,716] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,719] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,760] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,761] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 06:58:11,781] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:11,891] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:12,104] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:12,120] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:12,125] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:12,125] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 06:58:12,133] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:12,162] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 06:58:12,163] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 2.3, 'temperature_mlp': 2.3, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 2.3, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 2.3,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 2.3,
+    "temperature_mlp": 2.3,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2472101:2472101 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472101:2472101 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472101:2472101 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472101:2472101 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472101:2472101 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472101:2472101 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2472104:2472104 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472104:2472104 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472104:2472104 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472104:2472104 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472104:2472104 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472104:2472104 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472102:2472102 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472102:2472102 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472102:2472102 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472102:2472102 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472102:2472102 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472102:2472102 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472103:2472103 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472103:2472103 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472103:2472103 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472105:2472105 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472103:2472103 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472103:2472103 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472103:2472103 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472105:2472105 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472105:2472105 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472105:2472105 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472105:2472105 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472105:2472105 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2472108:2472108 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472108:2472108 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472108:2472108 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472108:2472108 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472108:2472108 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472108:2472108 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2472107:2472107 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472107:2472107 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472107:2472107 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472107:2472107 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472107:2472107 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472107:2472107 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472106:2472106 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2472106:2472106 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472106:2472106 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472106:2472106 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2472106:2472106 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2472106:2472106 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO ncclCommInitRank comm 0x5652b99b3940 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO ncclCommInitRank comm 0x55b9625e5450 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO ncclCommInitRank comm 0x55c0b2143880 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO ncclCommInitRank comm 0x55e9dda8cf10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO ncclCommInitRank comm 0x5651408f8000 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO ncclCommInitRank comm 0x56184f7bdc90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO ncclCommInitRank comm 0x557f16fce170 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO ncclCommInitRank comm 0x55630a4bcf40 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x56c94708d25d1353 - Init START
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO comm 0x557f16fce170 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO comm 0x55e9dda8cf10 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO comm 0x5652b99b3940 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO comm 0x55b9625e5450 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO comm 0x56184f7bdc90 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO comm 0x5651408f8000 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO comm 0x55c0b2143880 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO comm 0x55630a4bcf40 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472103:2473593 [2] NCCL INFO ncclCommInitRank comm 0x56184f7bdc90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472102:2473592 [1] NCCL INFO ncclCommInitRank comm 0x5651408f8000 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472106:2473597 [5] NCCL INFO ncclCommInitRank comm 0x557f16fce170 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472105:2473594 [4] NCCL INFO ncclCommInitRank comm 0x5652b99b3940 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472108:2473595 [7] NCCL INFO ncclCommInitRank comm 0x55630a4bcf40 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472101:2473590 [0] NCCL INFO ncclCommInitRank comm 0x55c0b2143880 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472107:2473596 [6] NCCL INFO ncclCommInitRank comm 0x55e9dda8cf10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x56c94708d25d1353 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2472104:2473591 [3] NCCL INFO ncclCommInitRank comm 0x55b9625e5450 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x56c94708d25d1353 - Init COMPLETE
+[2025-10-16 07:55:45,859] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 07:55:49,482] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=1.000000
+Pre-training init connector._connector.0.scores: Mean=1.000005
+Pre-training init connector._connector.2.scores: Mean=0.999970
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 07:56:07,742 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 07:56:07,746 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO ncclCommInitRank comm 0x7fb2d806b260 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO ncclCommInitRank comm 0x7f64dc06b440 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO ncclCommInitRank comm 0x7f5d6406bed0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO ncclCommInitRank comm 0x7f489806b240 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO ncclCommInitRank comm 0x7fa35c06bd00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO ncclCommInitRank comm 0x7fc58406b940 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO ncclCommInitRank comm 0x7f1ec806b9d0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO ncclCommInitRank comm 0x7efb0006ab60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xdaf0bb1f969da9 - Init START
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO comm 0x7fc58406b940 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO comm 0x7fa35c06bd00 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO comm 0x7f1ec806b9d0 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO comm 0x7efb0006ab60 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO comm 0x7f5d6406bed0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO comm 0x7fb2d806b260 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO comm 0x7f489806b240 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO comm 0x7f64dc06b440 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2472104:2479520 [3] NCCL INFO ncclCommInitRank comm 0x7efb0006ab60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472108:2479516 [7] NCCL INFO ncclCommInitRank comm 0x7f489806b240 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472102:2479519 [1] NCCL INFO ncclCommInitRank comm 0x7f64dc06b440 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472106:2479515 [5] NCCL INFO ncclCommInitRank comm 0x7fc58406b940 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472101:2479513 [0] NCCL INFO ncclCommInitRank comm 0x7f5d6406bed0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472105:2479514 [4] NCCL INFO ncclCommInitRank comm 0x7f1ec806b9d0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472107:2479517 [6] NCCL INFO ncclCommInitRank comm 0x7fa35c06bd00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2472103:2479518 [2] NCCL INFO ncclCommInitRank comm 0x7fb2d806b260 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xdaf0bb1f969da9 - Init COMPLETE
+
  0%|          | 1/520 [00:14<2:03:33, 14.28s/it]
                                                 
{'loss': 8.5235, 'grad_norm': 0.3913604611317396, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:14<2:03:33, 14.28s/it]
  0%|          | 2/520 [00:18<1:10:28,  8.16s/it]
                                                 
{'loss': 7.7698, 'grad_norm': 0.4003620366409385, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:18<1:10:28,  8.16s/it]
  1%|          | 3/520 [00:22<53:27,  6.20s/it]  
                                               
{'loss': 7.1513, 'grad_norm': 0.2001050929292849, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:22<53:27,  6.20s/it]
  1%|          | 4/520 [00:25<45:21,  5.27s/it]
                                               
{'loss': 6.3053, 'grad_norm': 0.12080860920966462, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:25<45:21,  5.27s/it]
  1%|          | 5/520 [00:29<40:29,  4.72s/it]
                                               
{'loss': 5.2784, 'grad_norm': 0.10261434212881083, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:29<40:29,  4.72s/it]
  1%|          | 6/520 [00:33<37:33,  4.38s/it]
                                               
{'loss': 7.6966, 'grad_norm': 0.4612208407887934, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:33<37:33,  4.38s/it]
  1%|▏         | 7/520 [00:36<35:22,  4.14s/it]
                                               
{'loss': 4.8601, 'grad_norm': 0.06060089873830955, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:36<35:22,  4.14s/it]
  2%|▏         | 8/520 [00:41<35:57,  4.21s/it]
                                               
{'loss': 4.6654, 'grad_norm': 0.07014005285272938, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:41<35:57,  4.21s/it]
  2%|▏         | 9/520 [00:45<34:43,  4.08s/it]
                                               
{'loss': 4.1564, 'grad_norm': 0.0481083096763645, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:45<34:43,  4.08s/it]
  2%|▏         | 10/520 [00:48<33:43,  3.97s/it]
                                                
{'loss': 3.0309, 'grad_norm': 0.027126948705331343, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:48<33:43,  3.97s/it]
  2%|▏         | 11/520 [00:52<33:05,  3.90s/it]
                                                
{'loss': 3.2452, 'grad_norm': 0.0519482277310465, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [00:52<33:05,  3.90s/it]
  2%|▏         | 12/520 [00:56<32:39,  3.86s/it]
                                                
{'loss': 4.3837, 'grad_norm': 0.20239577662357044, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [00:56<32:39,  3.86s/it][2025-10-16 07:57:13,895] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:00<33:59,  4.02s/it]
                                                
{'loss': 2.7137, 'grad_norm': 0.025116909664421115, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:00<33:59,  4.02s/it]
  3%|▎         | 14/520 [01:04<32:54,  3.90s/it]
                                                
{'loss': 2.6422, 'grad_norm': 0.01762702008300904, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:04<32:54,  3.90s/it]
  3%|▎         | 15/520 [01:08<32:15,  3.83s/it]
                                                
{'loss': 2.9344, 'grad_norm': 0.03305379391126529, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:08<32:15,  3.83s/it]
  3%|▎         | 16/520 [01:11<31:45,  3.78s/it]
                                                
{'loss': 2.6592, 'grad_norm': 0.024557318039306393, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:11<31:45,  3.78s/it]
  3%|▎         | 17/520 [01:15<31:28,  3.75s/it]
                                                
{'loss': 2.44, 'grad_norm': 0.037417160888358725, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:15<31:28,  3.75s/it]
  3%|▎         | 18/520 [01:19<31:07,  3.72s/it]
                                                
{'loss': 2.3478, 'grad_norm': 0.050410425226859255, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:19<31:07,  3.72s/it]
  4%|▎         | 19/520 [01:22<30:51,  3.70s/it]
                                                
{'loss': 2.6265, 'grad_norm': 0.04530879656091422, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:22<30:51,  3.70s/it]
  4%|▍         | 20/520 [01:26<30:39,  3.68s/it]
                                                
{'loss': 2.0614, 'grad_norm': 0.02315285482877796, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:26<30:39,  3.68s/it]
  4%|▍         | 21/520 [01:30<30:34,  3.68s/it]
                                                
{'loss': 2.52, 'grad_norm': 0.021716580204494036, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:30<30:34,  3.68s/it]
  4%|▍         | 22/520 [01:33<30:23,  3.66s/it]
                                                
{'loss': 2.1042, 'grad_norm': 0.015618209657919959, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:33<30:23,  3.66s/it]
  4%|▍         | 23/520 [01:37<30:15,  3.65s/it]
                                                
{'loss': 1.9927, 'grad_norm': 0.013023033242484185, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:37<30:15,  3.65s/it]
  5%|▍         | 24/520 [01:40<30:09,  3.65s/it]
                                                
{'loss': 2.1957, 'grad_norm': 0.018915685321722938, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [01:40<30:09,  3.65s/it]
  5%|▍         | 25/520 [01:44<30:04,  3.65s/it]
                                                
{'loss': 1.9594, 'grad_norm': 0.012460797983505396, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [01:44<30:04,  3.65s/it]
  5%|▌         | 26/520 [01:48<30:06,  3.66s/it]
                                                
{'loss': 1.905, 'grad_norm': 0.010567202179627906, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [01:48<30:06,  3.66s/it]
  5%|▌         | 27/520 [01:51<29:57,  3.65s/it]
                                                
{'loss': 1.7343, 'grad_norm': 0.010157866273926097, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [01:51<29:57,  3.65s/it]
  5%|▌         | 28/520 [01:55<29:55,  3.65s/it]
                                                
{'loss': 1.712, 'grad_norm': 0.007383763231166663, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [01:55<29:55,  3.65s/it]
  6%|▌         | 29/520 [01:59<29:52,  3.65s/it]
                                                
{'loss': 1.7026, 'grad_norm': 0.009541855108650814, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [01:59<29:52,  3.65s/it]
  6%|▌         | 30/520 [02:02<29:51,  3.66s/it]
                                                
{'loss': 2.1421, 'grad_norm': 0.011095297774323219, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:02<29:51,  3.66s/it]
  6%|▌         | 31/520 [02:06<29:43,  3.65s/it]
                                                
{'loss': 1.7028, 'grad_norm': 0.00912384006358289, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:06<29:43,  3.65s/it]
  6%|▌         | 32/520 [02:10<29:41,  3.65s/it]
                                                
{'loss': 2.2718, 'grad_norm': 0.01869283314675654, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:10<29:41,  3.65s/it]
  6%|▋         | 33/520 [02:13<29:38,  3.65s/it]
                                                
{'loss': 1.6755, 'grad_norm': 0.008807650350332286, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:13<29:38,  3.65s/it]
  7%|▋         | 34/520 [02:17<29:35,  3.65s/it]
                                                
{'loss': 1.6599, 'grad_norm': 0.008250895889031119, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:17<29:35,  3.65s/it]
  7%|▋         | 35/520 [02:21<29:30,  3.65s/it]
                                                
{'loss': 1.6397, 'grad_norm': 0.007111048636772043, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:21<29:30,  3.65s/it]
  7%|▋         | 36/520 [02:24<29:26,  3.65s/it]
                                                
{'loss': 1.7603, 'grad_norm': 0.0060572921308547105, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:24<29:26,  3.65s/it]
  7%|▋         | 37/520 [02:28<29:19,  3.64s/it]
                                                
{'loss': 2.0191, 'grad_norm': 0.014276364985412522, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:28<29:19,  3.64s/it]
  7%|▋         | 38/520 [02:31<29:14,  3.64s/it]
                                                
{'loss': 1.8396, 'grad_norm': 0.01042015293734024, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:31<29:14,  3.64s/it]
  8%|▊         | 39/520 [02:35<29:07,  3.63s/it]
                                                
{'loss': 1.6612, 'grad_norm': 0.006398925448133036, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [02:35<29:07,  3.63s/it]
  8%|▊         | 40/520 [02:39<29:03,  3.63s/it]
                                                
{'loss': 1.6627, 'grad_norm': 0.008064741157293458, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [02:39<29:03,  3.63s/it]
  8%|▊         | 41/520 [02:42<29:05,  3.64s/it]
                                                
{'loss': 1.6459, 'grad_norm': 0.005968116582848455, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [02:42<29:05,  3.64s/it]
  8%|▊         | 42/520 [02:46<28:58,  3.64s/it]
                                                
{'loss': 1.6706, 'grad_norm': 0.007707694436733977, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [02:46<28:58,  3.64s/it]
  8%|▊         | 43/520 [02:50<28:53,  3.64s/it]
                                                
{'loss': 1.8048, 'grad_norm': 0.010727038003014147, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [02:50<28:53,  3.64s/it]
  8%|▊         | 44/520 [02:53<28:54,  3.64s/it]
                                                
{'loss': 1.9128, 'grad_norm': 0.006912033868365057, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [02:53<28:54,  3.64s/it]
  9%|▊         | 45/520 [02:57<28:51,  3.65s/it]
                                                
{'loss': 1.6601, 'grad_norm': 0.010955915648701128, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [02:57<28:51,  3.65s/it]
  9%|▉         | 46/520 [03:01<28:46,  3.64s/it]
                                                
{'loss': 1.968, 'grad_norm': 0.009389495105666255, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:01<28:46,  3.64s/it]
  9%|▉         | 47/520 [03:04<28:39,  3.63s/it]
                                                
{'loss': 1.6117, 'grad_norm': 0.008587744438404802, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:04<28:39,  3.63s/it]
  9%|▉         | 48/520 [03:08<28:35,  3.63s/it]
                                                
{'loss': 1.6094, 'grad_norm': 0.007133659901551127, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:08<28:35,  3.63s/it]
  9%|▉         | 49/520 [03:11<28:33,  3.64s/it]
                                                
{'loss': 1.6075, 'grad_norm': 0.005076211922649191, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:11<28:33,  3.64s/it]
 10%|▉         | 50/520 [03:15<28:30,  3.64s/it]
                                                
{'loss': 1.5869, 'grad_norm': 0.005943975637125624, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:15<28:30,  3.64s/it]
 10%|▉         | 51/520 [03:19<28:23,  3.63s/it]
                                                
{'loss': 1.5149, 'grad_norm': 0.005676324551516328, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:19<28:23,  3.63s/it]
 10%|█         | 52/520 [03:22<28:23,  3.64s/it]
                                                
{'loss': 1.66, 'grad_norm': 0.007240320944432555, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:22<28:23,  3.64s/it]
 10%|█         | 53/520 [03:26<28:19,  3.64s/it]
                                                
{'loss': 1.6602, 'grad_norm': 0.0055250371675704115, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:26<28:19,  3.64s/it]
 10%|█         | 54/520 [03:30<28:16,  3.64s/it]
                                                
{'loss': 1.5217, 'grad_norm': 0.005197845273881428, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:30<28:16,  3.64s/it]
 11%|█         | 55/520 [03:33<28:18,  3.65s/it]
                                                
{'loss': 1.5243, 'grad_norm': 0.006163374542040963, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [03:33<28:18,  3.65s/it]
 11%|█         | 56/520 [03:37<28:17,  3.66s/it]
                                                
{'loss': 1.6749, 'grad_norm': 0.005921744064374651, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [03:37<28:17,  3.66s/it]
 11%|█         | 57/520 [03:41<28:12,  3.66s/it]
                                                
{'loss': 1.5061, 'grad_norm': 0.006813505063054879, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [03:41<28:12,  3.66s/it]
 11%|█         | 58/520 [03:44<28:07,  3.65s/it]
                                                
{'loss': 1.6737, 'grad_norm': 0.00557587783332447, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [03:44<28:07,  3.65s/it]
 11%|█▏        | 59/520 [03:48<28:04,  3.66s/it]
                                                
{'loss': 1.6398, 'grad_norm': 0.006332314040253262, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [03:48<28:04,  3.66s/it]
 12%|█▏        | 60/520 [03:52<28:00,  3.65s/it]
                                                
{'loss': 1.5943, 'grad_norm': 0.006870079901720629, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [03:52<28:00,  3.65s/it]
 12%|█▏        | 61/520 [03:55<28:03,  3.67s/it]
                                                
{'loss': 1.7686, 'grad_norm': 0.006682552694844482, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [03:55<28:03,  3.67s/it]
 12%|█▏        | 62/520 [03:59<27:52,  3.65s/it]
                                                
{'loss': 1.5577, 'grad_norm': 0.005229263412692806, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [03:59<27:52,  3.65s/it]
 12%|█▏        | 63/520 [04:03<27:51,  3.66s/it]
                                                
{'loss': 1.5585, 'grad_norm': 0.005906081219436532, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:03<27:51,  3.66s/it]
 12%|█▏        | 64/520 [04:06<27:48,  3.66s/it]
                                                
{'loss': 1.5783, 'grad_norm': 0.00528270709743081, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:06<27:48,  3.66s/it]
 12%|█▎        | 65/520 [04:10<28:12,  3.72s/it]
                                                
{'loss': 1.58, 'grad_norm': 0.005901067374912828, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:10<28:12,  3.72s/it]
 13%|█▎        | 66/520 [04:14<28:26,  3.76s/it]
                                                
{'loss': 1.5465, 'grad_norm': 0.0064386355449286906, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:14<28:26,  3.76s/it]
 13%|█▎        | 67/520 [04:18<28:37,  3.79s/it]
                                                
{'loss': 1.4077, 'grad_norm': 0.006136035056006947, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:18<28:37,  3.79s/it]
 13%|█▎        | 68/520 [04:22<28:44,  3.82s/it]
                                                
{'loss': 1.4623, 'grad_norm': 0.004947024477695747, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:22<28:44,  3.82s/it]
 13%|█▎        | 69/520 [04:26<28:47,  3.83s/it]
                                                
{'loss': 1.4407, 'grad_norm': 0.006935821415960521, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:26<28:47,  3.83s/it]
 13%|█▎        | 70/520 [04:29<28:47,  3.84s/it]
                                                
{'loss': 1.5055, 'grad_norm': 0.006935547761602636, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [04:29<28:47,  3.84s/it]
 14%|█▎        | 71/520 [04:33<28:51,  3.86s/it]
                                                
{'loss': 1.4105, 'grad_norm': 0.004635464866522209, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [04:33<28:51,  3.86s/it]
 14%|█▍        | 72/520 [04:37<28:47,  3.86s/it]
                                                
{'loss': 1.5733, 'grad_norm': 0.005341108373716614, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [04:37<28:47,  3.86s/it]
 14%|█▍        | 73/520 [04:41<28:46,  3.86s/it]
                                                
{'loss': 1.4025, 'grad_norm': 0.0052534520650603135, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [04:41<28:46,  3.86s/it]
 14%|█▍        | 74/520 [04:45<28:49,  3.88s/it]
                                                
{'loss': 1.5274, 'grad_norm': 0.0060798527046051504, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [04:45<28:49,  3.88s/it]
 14%|█▍        | 75/520 [04:49<28:50,  3.89s/it]
                                                
{'loss': 1.4061, 'grad_norm': 0.004779171142658033, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [04:49<28:50,  3.89s/it]
 15%|█▍        | 76/520 [04:53<28:52,  3.90s/it]
                                                
{'loss': 1.772, 'grad_norm': 0.008443576063864047, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [04:53<28:52,  3.90s/it]
 15%|█▍        | 77/520 [04:57<28:47,  3.90s/it]
                                                
{'loss': 1.3451, 'grad_norm': 0.005766082559455853, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [04:57<28:47,  3.90s/it]
 15%|█▌        | 78/520 [05:01<28:41,  3.90s/it]
                                                
{'loss': 1.4722, 'grad_norm': 0.005059208356749269, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:01<28:41,  3.90s/it]
 15%|█▌        | 79/520 [05:05<28:37,  3.89s/it]
                                                
{'loss': 1.4505, 'grad_norm': 0.00494969160274357, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:05<28:37,  3.89s/it]
 15%|█▌        | 80/520 [05:08<28:36,  3.90s/it]
                                                
{'loss': 1.7673, 'grad_norm': 0.006191042228377533, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:08<28:36,  3.90s/it]
 16%|█▌        | 81/520 [05:12<28:32,  3.90s/it]
                                                
{'loss': 1.6158, 'grad_norm': 0.006596455367253524, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:12<28:32,  3.90s/it]
 16%|█▌        | 82/520 [05:16<28:27,  3.90s/it]
                                                
{'loss': 1.5061, 'grad_norm': 0.004896879633552718, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:16<28:27,  3.90s/it]
 16%|█▌        | 83/520 [05:20<28:23,  3.90s/it]
                                                
{'loss': 1.5657, 'grad_norm': 0.004977306229267135, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:20<28:23,  3.90s/it]
 16%|█▌        | 84/520 [05:24<28:18,  3.89s/it]
                                                
{'loss': 1.5585, 'grad_norm': 0.006513344859906679, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:24<28:18,  3.89s/it]
 16%|█▋        | 85/520 [05:28<28:12,  3.89s/it]
                                                
{'loss': 1.5647, 'grad_norm': 0.004548030707686019, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:28<28:12,  3.89s/it]
 17%|█▋        | 86/520 [05:32<27:48,  3.85s/it]
                                                
{'loss': 1.6032, 'grad_norm': 0.005152380369468009, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:32<27:48,  3.85s/it]
 17%|█▋        | 87/520 [05:35<27:25,  3.80s/it]
                                                
{'loss': 1.7249, 'grad_norm': 0.008189265375359947, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [05:35<27:25,  3.80s/it]
 17%|█▋        | 88/520 [05:39<27:06,  3.76s/it]
                                                
{'loss': 1.7409, 'grad_norm': 0.009670161415497069, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [05:39<27:06,  3.76s/it]
 17%|█▋        | 89/520 [05:43<26:55,  3.75s/it]
                                                
{'loss': 1.5329, 'grad_norm': 0.005602165098708412, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [05:43<26:55,  3.75s/it]
 17%|█▋        | 90/520 [05:47<26:55,  3.76s/it]
                                                
{'loss': 1.452, 'grad_norm': 0.004703124808500795, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [05:47<26:55,  3.76s/it]
 18%|█▊        | 91/520 [05:50<26:56,  3.77s/it]
                                                
{'loss': 1.5396, 'grad_norm': 0.004758729698482102, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [05:50<26:56,  3.77s/it]
 18%|█▊        | 92/520 [05:54<26:55,  3.77s/it]
                                                
{'loss': 1.4715, 'grad_norm': 0.004502334886638161, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [05:54<26:55,  3.77s/it]
 18%|█▊        | 93/520 [05:58<26:56,  3.79s/it]
                                                
{'loss': 1.4735, 'grad_norm': 0.005024561373811434, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [05:58<26:56,  3.79s/it]
 18%|█▊        | 94/520 [06:02<26:51,  3.78s/it]
                                                
{'loss': 1.583, 'grad_norm': 0.005898637314636734, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [06:02<26:51,  3.78s/it]
 18%|█▊        | 95/520 [06:05<26:43,  3.77s/it]
                                                
{'loss': 1.4669, 'grad_norm': 0.00621731316568937, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:05<26:43,  3.77s/it]
 18%|█▊        | 96/520 [06:09<26:40,  3.77s/it]
                                                
{'loss': 1.4575, 'grad_norm': 0.004199786653153125, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:09<26:40,  3.77s/it]
 19%|█▊        | 97/520 [06:13<26:39,  3.78s/it]
                                                
{'loss': 1.4274, 'grad_norm': 0.005530818753999383, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:13<26:39,  3.78s/it]
 19%|█▉        | 98/520 [06:17<26:25,  3.76s/it]
                                                
{'loss': 1.4287, 'grad_norm': 0.004133246983821789, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:17<26:25,  3.76s/it]
 19%|█▉        | 99/520 [06:20<26:07,  3.72s/it]
                                                
{'loss': 1.4612, 'grad_norm': 0.005118206733792272, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:20<26:07,  3.72s/it]
 19%|█▉        | 100/520 [06:24<25:55,  3.70s/it]
                                                 
{'loss': 1.6065, 'grad_norm': 0.007285111439799166, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:24<25:55,  3.70s/it]
 19%|█▉        | 101/520 [06:28<25:48,  3.70s/it]
                                                 
{'loss': 1.4318, 'grad_norm': 0.0044089199776826595, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:28<25:48,  3.70s/it]
 20%|█▉        | 102/520 [06:31<25:36,  3.68s/it]
                                                 
{'loss': 1.4548, 'grad_norm': 0.0047168127541729965, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [06:31<25:36,  3.68s/it]
 20%|█▉        | 103/520 [06:35<25:29,  3.67s/it]
                                                 
{'loss': 1.3793, 'grad_norm': 0.004018222633505492, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [06:35<25:29,  3.67s/it]
 20%|██        | 104/520 [06:39<25:23,  3.66s/it]
                                                 
{'loss': 1.4606, 'grad_norm': 0.004493043566334175, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [06:39<25:23,  3.66s/it]
 20%|██        | 105/520 [06:42<25:24,  3.67s/it]
                                                 
{'loss': 1.4598, 'grad_norm': 0.004215685193587346, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [06:42<25:24,  3.67s/it]
 20%|██        | 106/520 [06:46<25:17,  3.67s/it]
                                                 
{'loss': 1.5763, 'grad_norm': 0.004799427479535402, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [06:46<25:17,  3.67s/it]
 21%|██        | 107/520 [06:50<25:18,  3.68s/it]
                                                 
{'loss': 1.5685, 'grad_norm': 0.005033989386833178, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [06:50<25:18,  3.68s/it]
 21%|██        | 108/520 [06:53<25:21,  3.69s/it]
                                                 
{'loss': 1.4165, 'grad_norm': 0.0047167021512499645, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [06:53<25:21,  3.69s/it]
 21%|██        | 109/520 [06:57<25:13,  3.68s/it]
                                                 
{'loss': 1.5549, 'grad_norm': 0.004364796256850141, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [06:57<25:13,  3.68s/it]
 21%|██        | 110/520 [07:01<25:10,  3.68s/it]
                                                 
{'loss': 1.5905, 'grad_norm': 0.004640429118460594, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [07:01<25:10,  3.68s/it]
 21%|██▏       | 111/520 [07:04<25:09,  3.69s/it]
                                                 
{'loss': 1.6154, 'grad_norm': 0.005000162095775558, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:04<25:09,  3.69s/it]
 22%|██▏       | 112/520 [07:08<25:03,  3.69s/it]
                                                 
{'loss': 1.48, 'grad_norm': 0.004394027709113995, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:08<25:03,  3.69s/it]
 22%|██▏       | 113/520 [07:12<24:59,  3.68s/it]
                                                 
{'loss': 1.3487, 'grad_norm': 0.0042436517554487576, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:12<24:59,  3.68s/it]
 22%|██▏       | 114/520 [07:15<24:55,  3.68s/it]
                                                 
{'loss': 1.4611, 'grad_norm': 0.004305722144733402, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:15<24:55,  3.68s/it]
 22%|██▏       | 115/520 [07:19<24:51,  3.68s/it]
                                                 
{'loss': 1.5852, 'grad_norm': 0.004283255669560669, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:19<24:51,  3.68s/it]
 22%|██▏       | 116/520 [07:23<24:48,  3.68s/it]
                                                 
{'loss': 1.5655, 'grad_norm': 0.003855395307334455, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:23<24:48,  3.68s/it]
 22%|██▎       | 117/520 [07:27<24:45,  3.69s/it]
                                                 
{'loss': 1.5644, 'grad_norm': 0.00488058139384268, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:27<24:45,  3.69s/it]
 23%|██▎       | 118/520 [07:30<24:37,  3.68s/it]
                                                 
{'loss': 1.4258, 'grad_norm': 0.0038737726472269874, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [07:30<24:37,  3.68s/it]
 23%|██▎       | 119/520 [07:34<24:30,  3.67s/it]
                                                 
{'loss': 1.3824, 'grad_norm': 0.004224049031170735, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [07:34<24:30,  3.67s/it]
 23%|██▎       | 120/520 [07:37<24:23,  3.66s/it]
                                                 
{'loss': 1.4136, 'grad_norm': 0.004758478197497966, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [07:37<24:23,  3.66s/it]
 23%|██▎       | 121/520 [07:41<24:17,  3.65s/it]
                                                 
{'loss': 1.469, 'grad_norm': 0.005490518142356933, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [07:41<24:17,  3.65s/it]
 23%|██▎       | 122/520 [07:45<24:13,  3.65s/it]
                                                 
{'loss': 1.3496, 'grad_norm': 0.004160431436848574, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [07:45<24:13,  3.65s/it]
 24%|██▎       | 123/520 [07:48<24:10,  3.65s/it]
                                                 
{'loss': 1.6325, 'grad_norm': 0.006439904813560765, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [07:48<24:10,  3.65s/it]
 24%|██▍       | 124/520 [07:52<24:07,  3.65s/it]
                                                 
{'loss': 1.4416, 'grad_norm': 0.004823183577095921, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [07:52<24:07,  3.65s/it]
 24%|██▍       | 125/520 [07:56<24:05,  3.66s/it]
                                                 
{'loss': 1.4276, 'grad_norm': 0.004359216555272262, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [07:56<24:05,  3.66s/it]
 24%|██▍       | 126/520 [08:00<25:23,  3.87s/it]
                                                 
{'loss': 1.5169, 'grad_norm': 0.004128226720359357, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [08:00<25:23,  3.87s/it]
 24%|██▍       | 127/520 [08:04<24:55,  3.81s/it]
                                                 
{'loss': 1.4049, 'grad_norm': 0.005562160440311777, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:04<24:55,  3.81s/it]
 25%|██▍       | 128/520 [08:07<24:33,  3.76s/it]
                                                 
{'loss': 1.4639, 'grad_norm': 0.004594785791828756, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:07<24:33,  3.76s/it]
 25%|██▍       | 129/520 [08:11<24:13,  3.72s/it]
                                                 
{'loss': 1.3731, 'grad_norm': 0.003644123917577619, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:11<24:13,  3.72s/it]
 25%|██▌       | 130/520 [08:15<24:01,  3.69s/it]
                                                 
{'loss': 1.4429, 'grad_norm': 0.004545766922521938, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:15<24:01,  3.69s/it]
 25%|██▌       | 131/520 [08:18<23:53,  3.68s/it]
                                                 
{'loss': 1.4939, 'grad_norm': 0.0049494920663071314, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:18<23:53,  3.68s/it]
 25%|██▌       | 132/520 [08:22<23:45,  3.67s/it]
                                                 
{'loss': 1.4866, 'grad_norm': 0.0044728572262886, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:22<23:45,  3.67s/it]
 26%|██▌       | 133/520 [08:26<23:36,  3.66s/it]
                                                 
{'loss': 1.3794, 'grad_norm': 0.003997821365924465, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:26<23:36,  3.66s/it]
 26%|██▌       | 134/520 [08:29<23:30,  3.65s/it]
                                                 
{'loss': 1.4733, 'grad_norm': 0.0044542193560434575, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [08:29<23:30,  3.65s/it]
 26%|██▌       | 135/520 [08:33<23:28,  3.66s/it]
                                                 
{'loss': 1.5599, 'grad_norm': 0.004080444021566329, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [08:33<23:28,  3.66s/it]
 26%|██▌       | 136/520 [08:37<23:23,  3.66s/it]
                                                 
{'loss': 1.4718, 'grad_norm': 0.004265880570525447, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [08:37<23:23,  3.66s/it]
 26%|██▋       | 137/520 [08:40<23:17,  3.65s/it]
                                                 
{'loss': 1.3884, 'grad_norm': 0.004827728596010002, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [08:40<23:17,  3.65s/it]
 27%|██▋       | 138/520 [08:44<23:28,  3.69s/it]
                                                 
{'loss': 1.3968, 'grad_norm': 0.0039179474351121206, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [08:44<23:28,  3.69s/it]
 27%|██▋       | 139/520 [08:48<23:47,  3.75s/it]
                                                 
{'loss': 1.3922, 'grad_norm': 0.004367837150112261, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [08:48<23:47,  3.75s/it]
 27%|██▋       | 140/520 [08:52<23:58,  3.78s/it]
                                                 
{'loss': 1.5328, 'grad_norm': 0.004674528352847334, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [08:52<23:58,  3.78s/it]
 27%|██▋       | 141/520 [08:56<24:01,  3.80s/it]
                                                 
{'loss': 1.5132, 'grad_norm': 0.00403800357869065, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [08:56<24:01,  3.80s/it]
 27%|██▋       | 142/520 [08:59<24:04,  3.82s/it]
                                                 
{'loss': 1.5702, 'grad_norm': 0.004622380574441163, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [08:59<24:04,  3.82s/it]
 28%|██▊       | 143/520 [09:04<24:27,  3.89s/it]
                                                 
{'loss': 1.4315, 'grad_norm': 0.0046160484203845754, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:04<24:27,  3.89s/it]
 28%|██▊       | 144/520 [09:08<24:48,  3.96s/it]
                                                 
{'loss': 1.3827, 'grad_norm': 0.004121484684682039, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:08<24:48,  3.96s/it]
 28%|██▊       | 145/520 [09:12<25:07,  4.02s/it]
                                                 
{'loss': 1.3126, 'grad_norm': 0.004040976326138172, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:12<25:07,  4.02s/it]
 28%|██▊       | 146/520 [09:16<25:17,  4.06s/it]
                                                 
{'loss': 1.629, 'grad_norm': 0.0045697896552807146, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:16<25:17,  4.06s/it]
 28%|██▊       | 147/520 [09:20<25:24,  4.09s/it]
                                                 
{'loss': 1.3646, 'grad_norm': 0.004259186961625304, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:20<25:24,  4.09s/it]
 28%|██▊       | 148/520 [09:24<25:28,  4.11s/it]
                                                 
{'loss': 1.4065, 'grad_norm': 0.0038416695328641307, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:24<25:28,  4.11s/it]
 29%|██▊       | 149/520 [09:28<25:27,  4.12s/it]
                                                 
{'loss': 1.3494, 'grad_norm': 0.0043971566899816674, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:28<25:27,  4.12s/it]
 29%|██▉       | 150/520 [09:33<25:27,  4.13s/it]
                                                 
{'loss': 1.5767, 'grad_norm': 0.004297426141721326, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [09:33<25:27,  4.13s/it]
 29%|██▉       | 151/520 [09:37<25:23,  4.13s/it]
                                                 
{'loss': 1.3794, 'grad_norm': 0.0038999668938928126, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [09:37<25:23,  4.13s/it]
 29%|██▉       | 152/520 [09:41<25:17,  4.12s/it]
                                                 
{'loss': 1.3528, 'grad_norm': 0.00443215794556219, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [09:41<25:17,  4.12s/it]
 29%|██▉       | 153/520 [09:45<24:43,  4.04s/it]
                                                 
{'loss': 1.3939, 'grad_norm': 0.004300298161063992, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [09:45<24:43,  4.04s/it]
 30%|██▉       | 154/520 [09:49<24:20,  3.99s/it]
                                                 
{'loss': 1.4715, 'grad_norm': 0.0038155122624758202, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [09:49<24:20,  3.99s/it]
 30%|██▉       | 155/520 [09:52<24:03,  3.95s/it]
                                                 
{'loss': 1.369, 'grad_norm': 0.003948868308901894, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [09:52<24:03,  3.95s/it]
 30%|███       | 156/520 [09:56<23:51,  3.93s/it]
                                                 
{'loss': 1.4171, 'grad_norm': 0.00445905019729696, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [09:56<23:51,  3.93s/it]
 30%|███       | 157/520 [10:00<23:41,  3.92s/it]
                                                 
{'loss': 1.6177, 'grad_norm': 0.004379488371437637, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [10:00<23:41,  3.92s/it]
 30%|███       | 158/520 [10:04<23:31,  3.90s/it]
                                                 
{'loss': 1.3893, 'grad_norm': 0.005383030915305872, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [10:04<23:31,  3.90s/it]
 31%|███       | 159/520 [10:08<23:21,  3.88s/it]
                                                 
{'loss': 1.4175, 'grad_norm': 0.004238471608249297, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [10:08<23:21,  3.88s/it]
 31%|███       | 160/520 [10:12<23:13,  3.87s/it]
                                                 
{'loss': 1.4571, 'grad_norm': 0.004476280972274989, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:12<23:13,  3.87s/it]
 31%|███       | 161/520 [10:16<23:08,  3.87s/it]
                                                 
{'loss': 1.4327, 'grad_norm': 0.004121431787320304, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:16<23:08,  3.87s/it]
 31%|███       | 162/520 [10:19<23:03,  3.86s/it]
                                                 
{'loss': 1.5288, 'grad_norm': 0.004501393989905793, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:19<23:03,  3.86s/it]
 31%|███▏      | 163/520 [10:23<22:58,  3.86s/it]
                                                 
{'loss': 1.3012, 'grad_norm': 0.004818108893899766, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:23<22:58,  3.86s/it]
 32%|███▏      | 164/520 [10:27<22:54,  3.86s/it]
                                                 
{'loss': 1.2656, 'grad_norm': 0.0039002608039969155, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:27<22:54,  3.86s/it]
 32%|███▏      | 165/520 [10:31<22:50,  3.86s/it]
                                                 
{'loss': 1.416, 'grad_norm': 0.00400261102510351, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:31<22:50,  3.86s/it]
 32%|███▏      | 166/520 [10:35<22:43,  3.85s/it]
                                                 
{'loss': 1.4077, 'grad_norm': 0.004488830591876469, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:35<22:43,  3.85s/it]
 32%|███▏      | 167/520 [10:39<22:42,  3.86s/it]
                                                 
{'loss': 1.3968, 'grad_norm': 0.004620993601680692, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [10:39<22:42,  3.86s/it]
 32%|███▏      | 168/520 [10:43<22:37,  3.86s/it]
                                                 
{'loss': 1.3373, 'grad_norm': 0.00385801777315814, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [10:43<22:37,  3.86s/it]
 32%|███▎      | 169/520 [10:46<22:31,  3.85s/it]
                                                 
{'loss': 1.415, 'grad_norm': 0.003944721565041814, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [10:46<22:31,  3.85s/it]
 33%|███▎      | 170/520 [10:50<22:28,  3.85s/it]
                                                 
{'loss': 1.4527, 'grad_norm': 0.004317261537651606, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [10:50<22:28,  3.85s/it]
 33%|███▎      | 171/520 [10:54<22:27,  3.86s/it]
                                                 
{'loss': 1.3466, 'grad_norm': 0.004369497757732824, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [10:54<22:27,  3.86s/it]
 33%|███▎      | 172/520 [10:58<22:21,  3.85s/it]
                                                 
{'loss': 1.4146, 'grad_norm': 0.0042729009776936635, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [10:58<22:21,  3.85s/it]
 33%|███▎      | 173/520 [11:02<22:18,  3.86s/it]
                                                 
{'loss': 1.3411, 'grad_norm': 0.004067555268839182, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [11:02<22:18,  3.86s/it]
 33%|███▎      | 174/520 [11:06<22:19,  3.87s/it]
                                                 
{'loss': 1.4228, 'grad_norm': 0.004720295035516524, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [11:06<22:19,  3.87s/it]
 34%|███▎      | 175/520 [11:10<22:16,  3.87s/it]
                                                 
{'loss': 1.3152, 'grad_norm': 0.004027361933418872, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [11:10<22:16,  3.87s/it]
 34%|███▍      | 176/520 [11:13<22:12,  3.87s/it]
                                                 
{'loss': 1.5484, 'grad_norm': 0.004297191666366299, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:13<22:12,  3.87s/it]
 34%|███▍      | 177/520 [11:17<22:07,  3.87s/it]
                                                 
{'loss': 1.413, 'grad_norm': 0.00480017339799047, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:17<22:07,  3.87s/it]
 34%|███▍      | 178/520 [11:21<22:01,  3.86s/it]
                                                 
{'loss': 1.4014, 'grad_norm': 0.004762438451406282, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:21<22:01,  3.86s/it]
 34%|███▍      | 179/520 [11:25<22:02,  3.88s/it]
                                                 
{'loss': 1.4753, 'grad_norm': 0.003758390999258942, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:25<22:02,  3.88s/it]
 35%|███▍      | 180/520 [11:29<22:00,  3.88s/it]
                                                 
{'loss': 1.3881, 'grad_norm': 0.004151879463901869, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:29<22:00,  3.88s/it]
 35%|███▍      | 181/520 [11:33<21:56,  3.88s/it]
                                                 
{'loss': 1.3517, 'grad_norm': 0.003950515598924332, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:33<21:56,  3.88s/it]
 35%|███▌      | 182/520 [11:37<21:51,  3.88s/it]
                                                 
{'loss': 1.3737, 'grad_norm': 0.003924875855286345, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:37<21:51,  3.88s/it]
 35%|███▌      | 183/520 [11:41<21:46,  3.88s/it]
                                                 
{'loss': 1.4037, 'grad_norm': 0.004112927259111505, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [11:41<21:46,  3.88s/it]
 35%|███▌      | 184/520 [11:45<21:45,  3.88s/it]
                                                 
{'loss': 1.3063, 'grad_norm': 0.0041095361310447675, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [11:45<21:45,  3.88s/it]
 36%|███▌      | 185/520 [11:48<21:40,  3.88s/it]
                                                 
{'loss': 1.495, 'grad_norm': 0.003804163549817975, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [11:48<21:40,  3.88s/it]
 36%|███▌      | 186/520 [11:52<21:39,  3.89s/it]
                                                 
{'loss': 1.3233, 'grad_norm': 0.00438341728270825, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [11:52<21:39,  3.89s/it]
 36%|███▌      | 187/520 [11:56<21:32,  3.88s/it]
                                                 
{'loss': 1.3459, 'grad_norm': 0.004625985860941311, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [11:56<21:32,  3.88s/it]
 36%|███▌      | 188/520 [12:00<21:27,  3.88s/it]
                                                 
{'loss': 1.4198, 'grad_norm': 0.004077396682850911, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [12:00<21:27,  3.88s/it]
 36%|███▋      | 189/520 [12:04<21:26,  3.89s/it]
                                                 
{'loss': 1.4355, 'grad_norm': 0.003596951782528192, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [12:04<21:26,  3.89s/it]
 37%|███▋      | 190/520 [12:08<21:18,  3.88s/it]
                                                 
{'loss': 1.3491, 'grad_norm': 0.004070908510701648, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [12:08<21:18,  3.88s/it]
 37%|███▋      | 191/520 [12:12<21:13,  3.87s/it]
                                                 
{'loss': 1.2998, 'grad_norm': 0.0037054061563126482, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [12:12<21:13,  3.87s/it]
 37%|███▋      | 192/520 [12:16<21:10,  3.87s/it]
                                                 
{'loss': 1.4, 'grad_norm': 0.003887418802750978, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:16<21:10,  3.87s/it]
 37%|███▋      | 193/520 [12:19<21:03,  3.86s/it]
                                                 
{'loss': 1.4626, 'grad_norm': 0.004276383589723301, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:19<21:03,  3.86s/it]
 37%|███▋      | 194/520 [12:23<20:58,  3.86s/it]
                                                 
{'loss': 1.3445, 'grad_norm': 0.00485949162758663, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:23<20:58,  3.86s/it]
 38%|███▊      | 195/520 [12:27<20:51,  3.85s/it]
                                                 
{'loss': 1.4176, 'grad_norm': 0.0038381639159191607, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:27<20:51,  3.85s/it]
 38%|███▊      | 196/520 [12:31<20:43,  3.84s/it]
                                                 
{'loss': 1.3776, 'grad_norm': 0.00441275307686093, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:31<20:43,  3.84s/it]
 38%|███▊      | 197/520 [12:35<20:24,  3.79s/it]
                                                 
{'loss': 1.3366, 'grad_norm': 0.0039013118953926014, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:35<20:24,  3.79s/it]
 38%|███▊      | 198/520 [12:38<20:05,  3.74s/it]
                                                 
{'loss': 1.4145, 'grad_norm': 0.00400695461290198, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:38<20:05,  3.74s/it]
 38%|███▊      | 199/520 [12:42<19:56,  3.73s/it]
                                                 
{'loss': 1.3255, 'grad_norm': 0.003845625686431171, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [12:42<19:56,  3.73s/it]
 38%|███▊      | 200/520 [12:46<19:46,  3.71s/it]
                                                 
{'loss': 1.4083, 'grad_norm': 0.004197334594671278, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [12:46<19:46,  3.71s/it]
 39%|███▊      | 201/520 [12:49<19:38,  3.69s/it]
                                                 
{'loss': 1.419, 'grad_norm': 0.003659070615648713, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [12:49<19:38,  3.69s/it]
 39%|███▉      | 202/520 [12:53<19:31,  3.68s/it]
                                                 
{'loss': 1.32, 'grad_norm': 0.003763441713136831, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [12:53<19:31,  3.68s/it]
 39%|███▉      | 203/520 [12:56<19:21,  3.66s/it]
                                                 
{'loss': 1.3772, 'grad_norm': 0.0038283154288098692, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [12:56<19:21,  3.66s/it]
 39%|███▉      | 204/520 [13:00<19:12,  3.65s/it]
                                                 
{'loss': 1.4136, 'grad_norm': 0.00423268738096767, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [13:00<19:12,  3.65s/it]
 39%|███▉      | 205/520 [13:04<19:11,  3.66s/it]
                                                 
{'loss': 1.4256, 'grad_norm': 0.004413689247759822, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [13:04<19:11,  3.66s/it]
 40%|███▉      | 206/520 [13:07<19:03,  3.64s/it]
                                                 
{'loss': 1.4437, 'grad_norm': 0.003997632542833236, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [13:07<19:03,  3.64s/it]
 40%|███▉      | 207/520 [13:11<19:02,  3.65s/it]
                                                 
{'loss': 1.4013, 'grad_norm': 0.003980610792685143, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [13:11<19:02,  3.65s/it]
 40%|████      | 208/520 [13:15<19:01,  3.66s/it]
                                                 
{'loss': 1.4082, 'grad_norm': 0.0041902128372218695, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [13:15<19:01,  3.66s/it]
 40%|████      | 209/520 [13:18<18:57,  3.66s/it]
                                                 
{'loss': 1.3438, 'grad_norm': 0.0038667470018521966, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:18<18:57,  3.66s/it]
 40%|████      | 210/520 [13:22<18:56,  3.67s/it]
                                                 
{'loss': 1.4162, 'grad_norm': 0.004431500694855415, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:22<18:56,  3.67s/it]
 41%|████      | 211/520 [13:26<18:53,  3.67s/it]
                                                 
{'loss': 1.4175, 'grad_norm': 0.003394890921866155, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:26<18:53,  3.67s/it]
 41%|████      | 212/520 [13:29<18:46,  3.66s/it]
                                                 
{'loss': 1.389, 'grad_norm': 0.0037374862027677183, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:29<18:46,  3.66s/it]
 41%|████      | 213/520 [13:33<18:42,  3.65s/it]
                                                 
{'loss': 1.3545, 'grad_norm': 0.004715221177594393, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:33<18:42,  3.65s/it]
 41%|████      | 214/520 [13:37<18:37,  3.65s/it]
                                                 
{'loss': 1.3439, 'grad_norm': 0.004193167529291361, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:37<18:37,  3.65s/it]
 41%|████▏     | 215/520 [13:40<18:34,  3.65s/it]
                                                 
{'loss': 1.3359, 'grad_norm': 0.0038546124228375053, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [13:40<18:34,  3.65s/it]
 42%|████▏     | 216/520 [13:44<18:32,  3.66s/it]
                                                 
{'loss': 1.2516, 'grad_norm': 0.0036811288263431017, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [13:44<18:32,  3.66s/it]
 42%|████▏     | 217/520 [13:48<18:26,  3.65s/it]
                                                 
{'loss': 1.3878, 'grad_norm': 0.0037859856201215955, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [13:48<18:26,  3.65s/it]
 42%|████▏     | 218/520 [13:51<18:21,  3.65s/it]
                                                 
{'loss': 1.3783, 'grad_norm': 0.003997388264236161, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [13:51<18:21,  3.65s/it]
 42%|████▏     | 219/520 [13:55<18:19,  3.65s/it]
                                                 
{'loss': 1.3431, 'grad_norm': 0.0034589077897485923, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [13:55<18:19,  3.65s/it]
 42%|████▏     | 220/520 [13:59<18:15,  3.65s/it]
                                                 
{'loss': 1.3894, 'grad_norm': 0.004012745143574364, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [13:59<18:15,  3.65s/it]
 42%|████▎     | 221/520 [14:02<18:14,  3.66s/it]
                                                 
{'loss': 1.3893, 'grad_norm': 0.003921156932241809, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [14:02<18:14,  3.66s/it]
 43%|████▎     | 222/520 [14:06<18:09,  3.66s/it]
                                                 
{'loss': 1.2859, 'grad_norm': 0.003566713240859609, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [14:06<18:09,  3.66s/it]
 43%|████▎     | 223/520 [14:10<18:04,  3.65s/it]
                                                 
{'loss': 1.2796, 'grad_norm': 0.0038661324194353415, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [14:10<18:04,  3.65s/it]
 43%|████▎     | 224/520 [14:13<18:02,  3.66s/it]
                                                 
{'loss': 1.5766, 'grad_norm': 0.007286170966569953, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [14:13<18:02,  3.66s/it]
 43%|████▎     | 225/520 [14:17<17:57,  3.65s/it]
                                                 
{'loss': 1.3068, 'grad_norm': 0.003911890496487393, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:17<17:57,  3.65s/it]
 43%|████▎     | 226/520 [14:21<18:00,  3.67s/it]
                                                 
{'loss': 1.4063, 'grad_norm': 0.003659223298539612, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:21<18:00,  3.67s/it]
 44%|████▎     | 227/520 [14:24<17:57,  3.68s/it]
                                                 
{'loss': 1.3874, 'grad_norm': 0.0036459397566637394, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:24<17:57,  3.68s/it]
 44%|████▍     | 228/520 [14:28<17:54,  3.68s/it]
                                                 
{'loss': 1.536, 'grad_norm': 0.00437305742175094, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:28<17:54,  3.68s/it]
 44%|████▍     | 229/520 [14:32<17:50,  3.68s/it]
                                                 
{'loss': 1.3715, 'grad_norm': 0.003621027005541903, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:32<17:50,  3.68s/it]
 44%|████▍     | 230/520 [14:35<17:55,  3.71s/it]
                                                 
{'loss': 1.2368, 'grad_norm': 0.0037711892525949305, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:35<17:55,  3.71s/it]
 44%|████▍     | 231/520 [14:39<18:04,  3.75s/it]
                                                 
{'loss': 1.3055, 'grad_norm': 0.003388871372115434, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:39<18:04,  3.75s/it]
 45%|████▍     | 232/520 [14:43<18:09,  3.78s/it]
                                                 
{'loss': 1.5616, 'grad_norm': 0.003915479768765855, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [14:43<18:09,  3.78s/it]
 45%|████▍     | 233/520 [14:47<18:10,  3.80s/it]
                                                 
{'loss': 1.445, 'grad_norm': 0.004758318724793387, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [14:47<18:10,  3.80s/it]
 45%|████▌     | 234/520 [14:51<18:10,  3.81s/it]
                                                 
{'loss': 1.2565, 'grad_norm': 0.0037816147533338154, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [14:51<18:10,  3.81s/it]
 45%|████▌     | 235/520 [14:55<18:10,  3.82s/it]
                                                 
{'loss': 1.3116, 'grad_norm': 0.004147368398112213, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [14:55<18:10,  3.82s/it]
 45%|████▌     | 236/520 [14:58<18:09,  3.84s/it]
                                                 
{'loss': 1.4318, 'grad_norm': 0.0034573343751747075, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [14:58<18:09,  3.84s/it]
 46%|████▌     | 237/520 [15:02<18:08,  3.85s/it]
                                                 
{'loss': 1.3767, 'grad_norm': 0.003655305881639875, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [15:02<18:08,  3.85s/it]
 46%|████▌     | 238/520 [15:06<18:05,  3.85s/it]
                                                 
{'loss': 1.3248, 'grad_norm': 0.003788123360885646, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [15:06<18:05,  3.85s/it]
 46%|████▌     | 239/520 [15:10<18:02,  3.85s/it]
                                                 
{'loss': 1.4279, 'grad_norm': 0.003986685133851496, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [15:10<18:02,  3.85s/it]
 46%|████▌     | 240/520 [15:14<18:03,  3.87s/it]
                                                 
{'loss': 1.1932, 'grad_norm': 0.004118619176425228, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:14<18:03,  3.87s/it]
 46%|████▋     | 241/520 [15:18<17:58,  3.86s/it]
                                                 
{'loss': 1.2845, 'grad_norm': 0.0038398590537267853, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:18<17:58,  3.86s/it]
 47%|████▋     | 242/520 [15:22<17:51,  3.85s/it]
                                                 
{'loss': 1.2998, 'grad_norm': 0.0033547878304354666, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:22<17:51,  3.85s/it]
 47%|████▋     | 243/520 [15:25<17:34,  3.81s/it]
                                                 
{'loss': 1.2876, 'grad_norm': 0.003620655688887776, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:25<17:34,  3.81s/it]
 47%|████▋     | 244/520 [15:29<17:18,  3.76s/it]
                                                 
{'loss': 1.4315, 'grad_norm': 0.0038235785587633496, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:29<17:18,  3.76s/it]
 47%|████▋     | 245/520 [15:33<17:04,  3.72s/it]
                                                 
{'loss': 1.277, 'grad_norm': 0.0036535980017414825, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:33<17:04,  3.72s/it]
 47%|████▋     | 246/520 [15:36<16:54,  3.70s/it]
                                                 
{'loss': 1.5344, 'grad_norm': 0.004134197712528507, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:36<16:54,  3.70s/it]
 48%|████▊     | 247/520 [15:40<16:45,  3.68s/it]
                                                 
{'loss': 1.4565, 'grad_norm': 0.003863626377307798, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:40<16:45,  3.68s/it]
 48%|████▊     | 248/520 [15:44<16:37,  3.67s/it]
                                                 
{'loss': 1.28, 'grad_norm': 0.0036028645773778065, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [15:44<16:37,  3.67s/it]
 48%|████▊     | 249/520 [15:47<16:30,  3.66s/it]
                                                 
{'loss': 1.3922, 'grad_norm': 0.003768171592363784, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [15:47<16:30,  3.66s/it]
 48%|████▊     | 250/520 [15:51<16:27,  3.66s/it]
                                                 
{'loss': 1.3326, 'grad_norm': 0.004090622315322356, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [15:51<16:27,  3.66s/it]
 48%|████▊     | 251/520 [15:55<16:25,  3.67s/it]
                                                 
{'loss': 1.3917, 'grad_norm': 0.0035104733352699755, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [15:55<16:25,  3.67s/it]
 48%|████▊     | 252/520 [15:58<16:19,  3.65s/it]
                                                 
{'loss': 1.4138, 'grad_norm': 0.003720237247893453, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [15:58<16:19,  3.65s/it]
 49%|████▊     | 253/520 [16:02<16:17,  3.66s/it]
                                                 
{'loss': 1.3853, 'grad_norm': 0.003936941376607147, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [16:02<16:17,  3.66s/it]
 49%|████▉     | 254/520 [16:05<16:11,  3.65s/it]
                                                 
{'loss': 1.2987, 'grad_norm': 0.0033192258232531053, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [16:05<16:11,  3.65s/it]
 49%|████▉     | 255/520 [16:09<16:05,  3.64s/it]
                                                 
{'loss': 1.3159, 'grad_norm': 0.004057888948260116, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [16:09<16:05,  3.64s/it]
 49%|████▉     | 256/520 [16:13<16:00,  3.64s/it]
                                                 
{'loss': 1.3548, 'grad_norm': 0.003827637212150641, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:13<16:00,  3.64s/it]
 49%|████▉     | 257/520 [16:16<15:56,  3.64s/it]
                                                 
{'loss': 1.3483, 'grad_norm': 0.0038553892471473285, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:16<15:56,  3.64s/it]
 50%|████▉     | 258/520 [16:20<15:54,  3.64s/it]
                                                 
{'loss': 1.3678, 'grad_norm': 0.003185987257850566, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:20<15:54,  3.64s/it]
 50%|████▉     | 259/520 [16:24<15:51,  3.64s/it]
                                                 
{'loss': 1.4294, 'grad_norm': 0.004110602735363109, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:24<15:51,  3.64s/it]
 50%|█████     | 260/520 [16:27<15:50,  3.66s/it]
                                                 
{'loss': 1.4949, 'grad_norm': 0.003641636473030274, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:27<15:50,  3.66s/it]
 50%|█████     | 261/520 [16:31<15:53,  3.68s/it]
                                                 
{'loss': 1.4231, 'grad_norm': 0.003923615767088027, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:31<15:53,  3.68s/it]
 50%|█████     | 262/520 [16:35<15:52,  3.69s/it]
                                                 
{'loss': 1.2693, 'grad_norm': 0.003802976969862634, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:35<15:52,  3.69s/it]
 51%|█████     | 263/520 [16:38<15:45,  3.68s/it]
                                                 
{'loss': 1.4341, 'grad_norm': 0.0040264677817172505, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:38<15:45,  3.68s/it]
 51%|█████     | 264/520 [16:42<15:43,  3.68s/it]
                                                 
{'loss': 1.3938, 'grad_norm': 0.0035898227488808717, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [16:42<15:43,  3.68s/it]
 51%|█████     | 265/520 [16:46<15:41,  3.69s/it]
                                                 
{'loss': 1.282, 'grad_norm': 0.0042346285532251764, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [16:46<15:41,  3.69s/it]
 51%|█████     | 266/520 [16:50<15:36,  3.69s/it]
                                                 
{'loss': 1.1339, 'grad_norm': 0.0031011891084391503, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [16:50<15:36,  3.69s/it]
 51%|█████▏    | 267/520 [16:53<15:34,  3.69s/it]
                                                 
{'loss': 1.2835, 'grad_norm': 0.00355364500743181, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [16:53<15:34,  3.69s/it]
 52%|█████▏    | 268/520 [16:57<15:29,  3.69s/it]
                                                 
{'loss': 1.5589, 'grad_norm': 0.00444566763837466, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [16:57<15:29,  3.69s/it]
 52%|█████▏    | 269/520 [17:01<15:25,  3.69s/it]
                                                 
{'loss': 1.4073, 'grad_norm': 0.0038771894745384242, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [17:01<15:25,  3.69s/it]
 52%|█████▏    | 270/520 [17:04<15:21,  3.69s/it]
                                                 
{'loss': 1.34, 'grad_norm': 0.00345463518020102, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [17:04<15:21,  3.69s/it]
 52%|█████▏    | 271/520 [17:08<15:16,  3.68s/it]
                                                 
{'loss': 1.3797, 'grad_norm': 0.0036167370834349804, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [17:08<15:16,  3.68s/it]
 52%|█████▏    | 272/520 [17:12<15:11,  3.68s/it]
                                                 
{'loss': 1.3537, 'grad_norm': 0.003926833700729971, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:12<15:11,  3.68s/it]
 52%|█████▎    | 273/520 [17:15<15:06,  3.67s/it]
                                                 
{'loss': 1.4886, 'grad_norm': 0.004204122442558255, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:15<15:06,  3.67s/it]
 53%|█████▎    | 274/520 [17:19<15:00,  3.66s/it]
                                                 
{'loss': 1.3432, 'grad_norm': 0.0038609150024603483, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:19<15:00,  3.66s/it]
 53%|█████▎    | 275/520 [17:23<14:57,  3.66s/it]
                                                 
{'loss': 1.284, 'grad_norm': 0.003988181726847449, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:23<14:57,  3.66s/it]
 53%|█████▎    | 276/520 [17:26<14:53,  3.66s/it]
                                                 
{'loss': 1.3729, 'grad_norm': 0.003951580528318205, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:26<14:53,  3.66s/it]
 53%|█████▎    | 277/520 [17:30<14:50,  3.67s/it]
                                                 
{'loss': 1.4754, 'grad_norm': 0.004721757684006577, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:30<14:50,  3.67s/it]
 53%|█████▎    | 278/520 [17:34<14:46,  3.66s/it]
                                                 
{'loss': 1.236, 'grad_norm': 0.0034123515699430325, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:34<14:46,  3.66s/it]
 54%|█████▎    | 279/520 [17:37<14:42,  3.66s/it]
                                                 
{'loss': 1.3684, 'grad_norm': 0.004829137357339493, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:37<14:42,  3.66s/it]
 54%|█████▍    | 280/520 [17:41<14:40,  3.67s/it]
                                                 
{'loss': 1.2871, 'grad_norm': 0.004311580466140991, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [17:41<14:40,  3.67s/it]
 54%|█████▍    | 281/520 [17:45<14:37,  3.67s/it]
                                                 
{'loss': 1.403, 'grad_norm': 0.0039171200071355765, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [17:45<14:37,  3.67s/it]
 54%|█████▍    | 282/520 [17:48<14:33,  3.67s/it]
                                                 
{'loss': 1.2374, 'grad_norm': 0.0034898107582304783, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [17:48<14:33,  3.67s/it]
 54%|█████▍    | 283/520 [17:52<14:31,  3.68s/it]
                                                 
{'loss': 1.4209, 'grad_norm': 0.004279498806185431, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [17:52<14:31,  3.68s/it]
 55%|█████▍    | 284/520 [17:56<14:26,  3.67s/it]
                                                 
{'loss': 1.3452, 'grad_norm': 0.004149673526559456, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [17:56<14:26,  3.67s/it]
 55%|█████▍    | 285/520 [17:59<14:22,  3.67s/it]
                                                 
{'loss': 1.2792, 'grad_norm': 0.00376834928199126, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [17:59<14:22,  3.67s/it]
 55%|█████▌    | 286/520 [18:03<14:21,  3.68s/it]
                                                 
{'loss': 1.1439, 'grad_norm': 0.004338313403770008, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [18:03<14:21,  3.68s/it]
 55%|█████▌    | 287/520 [18:07<14:17,  3.68s/it]
                                                 
{'loss': 1.3938, 'grad_norm': 0.004071172303230915, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [18:07<14:17,  3.68s/it]
 55%|█████▌    | 288/520 [18:10<14:13,  3.68s/it]
                                                 
{'loss': 1.4402, 'grad_norm': 0.003864228231721783, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:10<14:13,  3.68s/it]
 56%|█████▌    | 289/520 [18:14<14:11,  3.68s/it]
                                                 
{'loss': 1.292, 'grad_norm': 0.00337191344096978, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:14<14:11,  3.68s/it]
 56%|█████▌    | 290/520 [18:18<14:05,  3.67s/it]
                                                 
{'loss': 1.2065, 'grad_norm': 0.0034054216247004577, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:18<14:05,  3.67s/it]
 56%|█████▌    | 291/520 [18:21<14:00,  3.67s/it]
                                                 
{'loss': 1.2671, 'grad_norm': 0.004112584812170378, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:21<14:00,  3.67s/it]
 56%|█████▌    | 292/520 [18:25<13:59,  3.68s/it]
                                                 
{'loss': 1.309, 'grad_norm': 0.0035592208722243345, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:25<13:59,  3.68s/it]
 56%|█████▋    | 293/520 [18:29<13:56,  3.68s/it]
                                                 
{'loss': 1.256, 'grad_norm': 0.003907175297205563, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:29<13:56,  3.68s/it]
 57%|█████▋    | 294/520 [18:32<13:53,  3.69s/it]
                                                 
{'loss': 1.2825, 'grad_norm': 0.0038362627340015596, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:32<13:53,  3.69s/it]
 57%|█████▋    | 295/520 [18:36<13:48,  3.68s/it]
                                                 
{'loss': 1.4073, 'grad_norm': 0.004448854684433617, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:36<13:48,  3.68s/it]
 57%|█████▋    | 296/520 [18:40<13:44,  3.68s/it]
                                                 
{'loss': 1.2247, 'grad_norm': 0.0038434096030084135, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [18:40<13:44,  3.68s/it]
 57%|█████▋    | 297/520 [18:43<13:40,  3.68s/it]
                                                 
{'loss': 1.3717, 'grad_norm': 0.00409998988641379, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [18:43<13:40,  3.68s/it]
 57%|█████▋    | 298/520 [18:47<13:33,  3.66s/it]
                                                 
{'loss': 1.3253, 'grad_norm': 0.0033947201979663516, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [18:47<13:33,  3.66s/it]
 57%|█████▊    | 299/520 [18:51<13:28,  3.66s/it]
                                                 
{'loss': 1.4229, 'grad_norm': 0.0035444608168033105, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [18:51<13:28,  3.66s/it]
 58%|█████▊    | 300/520 [18:54<13:23,  3.65s/it]
                                                 
{'loss': 1.3901, 'grad_norm': 0.0038345325670883475, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [18:54<13:23,  3.65s/it]
 58%|█████▊    | 301/520 [18:58<13:17,  3.64s/it]
                                                 
{'loss': 1.3526, 'grad_norm': 0.003587235333801786, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [18:58<13:17,  3.64s/it]
 58%|█████▊    | 302/520 [19:02<13:14,  3.65s/it]
                                                 
{'loss': 1.4433, 'grad_norm': 0.004027297120648425, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [19:02<13:14,  3.65s/it]
 58%|█████▊    | 303/520 [19:05<13:13,  3.66s/it]
                                                 
{'loss': 1.2921, 'grad_norm': 0.00424390791038209, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [19:05<13:13,  3.66s/it]
 58%|█████▊    | 304/520 [19:09<13:13,  3.67s/it]
                                                 
{'loss': 1.3285, 'grad_norm': 0.004255964018445352, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:09<13:13,  3.67s/it]
 59%|█████▊    | 305/520 [19:13<13:09,  3.67s/it]
                                                 
{'loss': 1.4045, 'grad_norm': 0.004138794115744359, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:13<13:09,  3.67s/it]
 59%|█████▉    | 306/520 [19:16<13:08,  3.68s/it]
                                                 
{'loss': 1.3278, 'grad_norm': 0.003492555961712439, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:16<13:08,  3.68s/it]
 59%|█████▉    | 307/520 [19:20<13:24,  3.78s/it]
                                                 
{'loss': 1.2722, 'grad_norm': 0.0036470560208065744, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:20<13:24,  3.78s/it]
 59%|█████▉    | 308/520 [19:24<13:14,  3.75s/it]
                                                 
{'loss': 1.3996, 'grad_norm': 0.0034936656165953147, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:24<13:14,  3.75s/it]
 59%|█████▉    | 309/520 [19:28<13:04,  3.72s/it]
                                                 
{'loss': 1.2666, 'grad_norm': 0.0035161308807919473, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:28<13:04,  3.72s/it]
 60%|█████▉    | 310/520 [19:31<12:57,  3.70s/it]
                                                 
{'loss': 1.2442, 'grad_norm': 0.003693226765522957, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:31<12:57,  3.70s/it]
 60%|█████▉    | 311/520 [19:35<12:52,  3.69s/it]
                                                 
{'loss': 1.2161, 'grad_norm': 0.0035463178771003964, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [19:35<12:52,  3.69s/it]
 60%|██████    | 312/520 [19:39<12:45,  3.68s/it]
                                                 
{'loss': 1.2075, 'grad_norm': 0.0040640201532730355, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [19:39<12:45,  3.68s/it]
 60%|██████    | 313/520 [19:42<12:42,  3.69s/it]
                                                 
{'loss': 1.2003, 'grad_norm': 0.0032082103074756708, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [19:42<12:42,  3.69s/it]
 60%|██████    | 314/520 [19:46<12:59,  3.78s/it]
                                                 
{'loss': 1.2358, 'grad_norm': 0.0034658492320840434, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [19:46<12:59,  3.78s/it]
 61%|██████    | 315/520 [19:50<12:45,  3.74s/it]
                                                 
{'loss': 1.4017, 'grad_norm': 0.004173753868037704, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [19:50<12:45,  3.74s/it]
 61%|██████    | 316/520 [19:54<13:04,  3.84s/it]
                                                 
{'loss': 1.2063, 'grad_norm': 0.004366227624187689, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [19:54<13:04,  3.84s/it]
 61%|██████    | 317/520 [19:58<12:48,  3.78s/it]
                                                 
{'loss': 1.2302, 'grad_norm': 0.0033191950417801, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [19:58<12:48,  3.78s/it]
 61%|██████    | 318/520 [20:01<12:37,  3.75s/it]
                                                 
{'loss': 1.3584, 'grad_norm': 0.004006411570158896, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [20:01<12:37,  3.75s/it]
 61%|██████▏   | 319/520 [20:05<12:46,  3.82s/it]
                                                 
{'loss': 1.2198, 'grad_norm': 0.0036716737600352213, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [20:05<12:46,  3.82s/it]
 62%|██████▏   | 320/520 [20:09<12:31,  3.76s/it]
                                                 
{'loss': 1.1633, 'grad_norm': 0.003938095398885777, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:09<12:31,  3.76s/it]
 62%|██████▏   | 321/520 [20:13<12:22,  3.73s/it]
                                                 
{'loss': 1.3717, 'grad_norm': 0.0039187981744018465, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:13<12:22,  3.73s/it]
 62%|██████▏   | 322/520 [20:16<12:15,  3.72s/it]
                                                 
{'loss': 1.2565, 'grad_norm': 0.0036188158506650944, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:16<12:15,  3.72s/it]
 62%|██████▏   | 323/520 [20:20<12:07,  3.69s/it]
                                                 
{'loss': 1.342, 'grad_norm': 0.004398193459761885, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:20<12:07,  3.69s/it]
 62%|██████▏   | 324/520 [20:24<12:03,  3.69s/it]
                                                 
{'loss': 1.2943, 'grad_norm': 0.004336721500745743, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:24<12:03,  3.69s/it]
 62%|██████▎   | 325/520 [20:27<11:57,  3.68s/it]
                                                 
{'loss': 1.3154, 'grad_norm': 0.003914828084307649, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:27<11:57,  3.68s/it]
 63%|██████▎   | 326/520 [20:31<11:52,  3.67s/it]
                                                 
{'loss': 1.2901, 'grad_norm': 0.003641108405813329, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:31<11:52,  3.67s/it]
 63%|██████▎   | 327/520 [20:35<11:48,  3.67s/it]
                                                 
{'loss': 1.4149, 'grad_norm': 0.004897635809473525, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [20:35<11:48,  3.67s/it]
 63%|██████▎   | 328/520 [20:38<11:42,  3.66s/it]
                                                 
{'loss': 1.3607, 'grad_norm': 0.0037753373189170656, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [20:38<11:42,  3.66s/it]
 63%|██████▎   | 329/520 [20:42<11:37,  3.65s/it]
                                                 
{'loss': 1.2049, 'grad_norm': 0.00318022934343727, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [20:42<11:37,  3.65s/it]
 63%|██████▎   | 330/520 [20:46<11:32,  3.65s/it]
                                                 
{'loss': 1.2802, 'grad_norm': 0.0032571903158216736, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [20:46<11:32,  3.65s/it]
 64%|██████▎   | 331/520 [20:49<11:27,  3.64s/it]
                                                 
{'loss': 1.2475, 'grad_norm': 0.003513542687698178, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [20:49<11:27,  3.64s/it]
 64%|██████▍   | 332/520 [20:53<11:24,  3.64s/it]
                                                 
{'loss': 1.4084, 'grad_norm': 0.0035902259393124363, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [20:53<11:24,  3.64s/it]
 64%|██████▍   | 333/520 [20:57<11:20,  3.64s/it]
                                                 
{'loss': 1.4173, 'grad_norm': 0.0038974085031834626, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [20:57<11:20,  3.64s/it]
 64%|██████▍   | 334/520 [21:00<11:18,  3.65s/it]
                                                 
{'loss': 1.299, 'grad_norm': 0.004216257347745648, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [21:00<11:18,  3.65s/it]
 64%|██████▍   | 335/520 [21:04<11:14,  3.65s/it]
                                                 
{'loss': 1.2857, 'grad_norm': 0.003387344620699763, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [21:04<11:14,  3.65s/it]
 65%|██████▍   | 336/520 [21:07<11:10,  3.64s/it]
                                                 
{'loss': 1.1751, 'grad_norm': 0.003897243024059751, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:07<11:10,  3.64s/it]
 65%|██████▍   | 337/520 [21:11<11:13,  3.68s/it]
                                                 
{'loss': 1.1793, 'grad_norm': 0.003622348173500159, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:11<11:13,  3.68s/it]
 65%|██████▌   | 338/520 [21:15<11:19,  3.74s/it]
                                                 
{'loss': 1.3077, 'grad_norm': 0.0037076896803474265, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:15<11:19,  3.74s/it]
 65%|██████▌   | 339/520 [21:19<11:20,  3.76s/it]
                                                 
{'loss': 1.2409, 'grad_norm': 0.0034937505317991454, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:19<11:20,  3.76s/it]
 65%|██████▌   | 340/520 [21:23<11:24,  3.80s/it]
                                                 
{'loss': 1.2329, 'grad_norm': 0.0035280668507783703, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:23<11:24,  3.80s/it]
 66%|██████▌   | 341/520 [21:27<11:22,  3.81s/it]
                                                 
{'loss': 1.2543, 'grad_norm': 0.003926354349798424, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:27<11:22,  3.81s/it]
 66%|██████▌   | 342/520 [21:31<11:20,  3.82s/it]
                                                 
{'loss': 1.3793, 'grad_norm': 0.004155882949204402, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:31<11:20,  3.82s/it]
 66%|██████▌   | 343/520 [21:34<11:18,  3.83s/it]
                                                 
{'loss': 1.3507, 'grad_norm': 0.0038084678155254795, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [21:34<11:18,  3.83s/it]
 66%|██████▌   | 344/520 [21:38<11:15,  3.84s/it]
                                                 
{'loss': 1.2009, 'grad_norm': 0.0038204163569435824, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [21:38<11:15,  3.84s/it]
 66%|██████▋   | 345/520 [21:42<11:14,  3.85s/it]
                                                 
{'loss': 1.3244, 'grad_norm': 0.00443898733827127, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [21:42<11:14,  3.85s/it]
 67%|██████▋   | 346/520 [21:46<11:10,  3.85s/it]
                                                 
{'loss': 1.3364, 'grad_norm': 0.003590975441889293, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [21:46<11:10,  3.85s/it]
 67%|██████▋   | 347/520 [21:50<11:05,  3.85s/it]
                                                 
{'loss': 1.2254, 'grad_norm': 0.0033486552305163975, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [21:50<11:05,  3.85s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [21:54<11:00,  3.84s/it]
                                                 
{'loss': 1.1824, 'grad_norm': 0.004403949224813527, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [21:54<11:00,  3.84s/it]
 67%|██████▋   | 349/520 [21:58<10:58,  3.85s/it]
                                                 
{'loss': 1.2309, 'grad_norm': 0.0038615472257440323, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [21:58<10:58,  3.85s/it]
 67%|██████▋   | 350/520 [22:01<10:54,  3.85s/it]
                                                 
{'loss': 1.2648, 'grad_norm': 0.003987141367131526, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [22:01<10:54,  3.85s/it]
 68%|██████▊   | 351/520 [22:05<10:51,  3.86s/it]
                                                 
{'loss': 1.1647, 'grad_norm': 0.003471806253484007, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [22:05<10:51,  3.86s/it]
 68%|██████▊   | 352/520 [22:09<10:47,  3.86s/it]
                                                 
{'loss': 1.2938, 'grad_norm': 0.0035698994561296463, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [22:09<10:47,  3.86s/it]
 68%|██████▊   | 353/520 [22:13<10:45,  3.87s/it]
                                                 
{'loss': 1.2842, 'grad_norm': 0.003240976621354375, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:13<10:45,  3.87s/it]
 68%|██████▊   | 354/520 [22:17<10:40,  3.86s/it]
                                                 
{'loss': 1.4179, 'grad_norm': 0.00361825631070284, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:17<10:40,  3.86s/it]
 68%|██████▊   | 355/520 [22:21<10:34,  3.85s/it]
                                                 
{'loss': 1.2339, 'grad_norm': 0.003571479893854371, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:21<10:34,  3.85s/it]
 68%|██████▊   | 356/520 [22:24<10:30,  3.84s/it]
                                                 
{'loss': 1.2308, 'grad_norm': 0.0037154572604216254, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:24<10:30,  3.84s/it]
 69%|██████▊   | 357/520 [22:28<10:25,  3.84s/it]
                                                 
{'loss': 1.2538, 'grad_norm': 0.0032213407651582497, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:28<10:25,  3.84s/it]
 69%|██████▉   | 358/520 [22:32<10:23,  3.85s/it]
                                                 
{'loss': 1.182, 'grad_norm': 0.0035086052631696724, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:32<10:23,  3.85s/it]
 69%|██████▉   | 359/520 [22:36<10:21,  3.86s/it]
                                                 
{'loss': 1.3512, 'grad_norm': 0.004017559347880374, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [22:36<10:21,  3.86s/it]
 69%|██████▉   | 360/520 [22:40<10:18,  3.86s/it]
                                                 
{'loss': 1.3721, 'grad_norm': 0.005327200050080632, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [22:40<10:18,  3.86s/it]
 69%|██████▉   | 361/520 [22:44<10:13,  3.86s/it]
                                                 
{'loss': 1.3531, 'grad_norm': 0.0033625180146513857, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [22:44<10:13,  3.86s/it]
 70%|██████▉   | 362/520 [22:48<10:09,  3.86s/it]
                                                 
{'loss': 1.2462, 'grad_norm': 0.0037492039926407795, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [22:48<10:09,  3.86s/it]
 70%|██████▉   | 363/520 [22:51<10:06,  3.86s/it]
                                                 
{'loss': 1.2751, 'grad_norm': 0.003452189223994529, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [22:51<10:06,  3.86s/it]
 70%|███████   | 364/520 [22:55<10:04,  3.87s/it]
                                                 
{'loss': 1.3721, 'grad_norm': 0.00362633238324422, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [22:55<10:04,  3.87s/it]
 70%|███████   | 365/520 [22:59<09:58,  3.86s/it]
                                                 
{'loss': 1.3365, 'grad_norm': 0.0037460624516991317, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [22:59<09:58,  3.86s/it]
 70%|███████   | 366/520 [23:03<09:55,  3.87s/it]
                                                 
{'loss': 1.2852, 'grad_norm': 0.003188409872448185, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [23:03<09:55,  3.87s/it]
 71%|███████   | 367/520 [23:07<09:52,  3.87s/it]
                                                 
{'loss': 1.2915, 'grad_norm': 0.0035386208081359814, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [23:07<09:52,  3.87s/it]
 71%|███████   | 368/520 [23:11<09:48,  3.87s/it]
                                                 
{'loss': 1.1365, 'grad_norm': 0.003771705428234553, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [23:11<09:48,  3.87s/it]
 71%|███████   | 369/520 [23:15<09:44,  3.87s/it]
                                                 
{'loss': 1.3234, 'grad_norm': 0.0035800750829739674, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:15<09:44,  3.87s/it]
 71%|███████   | 370/520 [23:19<09:40,  3.87s/it]
                                                 
{'loss': 1.1955, 'grad_norm': 0.003178263127956371, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:19<09:40,  3.87s/it]
 71%|███████▏  | 371/520 [23:22<09:37,  3.88s/it]
                                                 
{'loss': 1.1954, 'grad_norm': 0.003546919021729443, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:22<09:37,  3.88s/it]
 72%|███████▏  | 372/520 [23:26<09:33,  3.88s/it]
                                                 
{'loss': 1.4409, 'grad_norm': 0.0052629504939815394, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:26<09:33,  3.88s/it]
 72%|███████▏  | 373/520 [23:30<09:28,  3.87s/it]
                                                 
{'loss': 1.3163, 'grad_norm': 0.006201123862548985, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:30<09:28,  3.87s/it]
 72%|███████▏  | 374/520 [23:34<09:24,  3.87s/it]
                                                 
{'loss': 1.2806, 'grad_norm': 0.003439158888787842, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:34<09:24,  3.87s/it]
 72%|███████▏  | 375/520 [23:38<09:20,  3.87s/it]
                                                 
{'loss': 1.1868, 'grad_norm': 0.0036152559268729473, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [23:38<09:20,  3.87s/it]
 72%|███████▏  | 376/520 [23:42<09:13,  3.84s/it]
                                                 
{'loss': 1.3149, 'grad_norm': 0.0035677961395436343, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [23:42<09:13,  3.84s/it]
 72%|███████▎  | 377/520 [23:45<09:01,  3.78s/it]
                                                 
{'loss': 1.2501, 'grad_norm': 0.0036895024563328637, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [23:45<09:01,  3.78s/it]
 73%|███████▎  | 378/520 [23:49<08:50,  3.74s/it]
                                                 
{'loss': 1.2967, 'grad_norm': 0.003483233983211581, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [23:49<08:50,  3.74s/it]
 73%|███████▎  | 379/520 [23:53<08:42,  3.70s/it]
                                                 
{'loss': 1.286, 'grad_norm': 0.003338746795238868, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [23:53<08:42,  3.70s/it]
 73%|███████▎  | 380/520 [23:56<08:37,  3.70s/it]
                                                 
{'loss': 1.411, 'grad_norm': 0.004792541584851034, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [23:56<08:37,  3.70s/it]
 73%|███████▎  | 381/520 [24:00<08:32,  3.69s/it]
                                                 
{'loss': 1.2798, 'grad_norm': 0.003404658302774625, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [24:00<08:32,  3.69s/it]
 73%|███████▎  | 382/520 [24:04<08:27,  3.68s/it]
                                                 
{'loss': 1.3428, 'grad_norm': 0.003899652441919031, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [24:04<08:27,  3.68s/it]
 74%|███████▎  | 383/520 [24:07<08:22,  3.67s/it]
                                                 
{'loss': 1.116, 'grad_norm': 0.0037691992756103683, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [24:07<08:22,  3.67s/it]
 74%|███████▍  | 384/520 [24:11<08:17,  3.66s/it]
                                                 
{'loss': 1.4554, 'grad_norm': 0.00400872985816905, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [24:11<08:17,  3.66s/it]
 74%|███████▍  | 385/520 [24:15<08:14,  3.66s/it]
                                                 
{'loss': 1.2599, 'grad_norm': 0.0034090806930686017, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:15<08:14,  3.66s/it]
 74%|███████▍  | 386/520 [24:18<08:08,  3.65s/it]
                                                 
{'loss': 1.2051, 'grad_norm': 0.0031364257462190986, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:18<08:08,  3.65s/it]
 74%|███████▍  | 387/520 [24:22<08:04,  3.64s/it]
                                                 
{'loss': 1.4219, 'grad_norm': 0.003664146637136088, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:22<08:04,  3.64s/it]
 75%|███████▍  | 388/520 [24:26<08:02,  3.65s/it]
                                                 
{'loss': 1.1534, 'grad_norm': 0.0032777557785191873, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:26<08:02,  3.65s/it]
 75%|███████▍  | 389/520 [24:29<07:59,  3.66s/it]
                                                 
{'loss': 1.2266, 'grad_norm': 0.004139490775654981, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}[E ProcessGroupNCCL.cpp:474] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800025 milliseconds before timing out.
+
+[E ProcessGroupNCCL.cpp:474] [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800090 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800105 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800251 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800351 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800802 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800907 milliseconds before timing out.
+
 75%|███████▍  | 389/520 [1:00:12<07:59,  3.66s/it]ywang29-vrdb-test2-worker-0:2472108:2473624 [7] NCCL INFO [Service thread] Connection closed by localRank 7
+ywang29-vrdb-test2-worker-0:2472106:2473611 [5] NCCL INFO [Service thread] Connection closed by localRank 5
+ywang29-vrdb-test2-worker-0:2472107:2473612 [6] NCCL INFO [Service thread] Connection closed by localRank 6
+ywang29-vrdb-test2-worker-0:2472105:2473622 [4] NCCL INFO [Service thread] Connection closed by localRank 4
+ywang29-vrdb-test2-worker-0:2472104:2473618 [3] NCCL INFO [Service thread] Connection closed by localRank 3
+ywang29-vrdb-test2-worker-0:2472103:2473620 [2] NCCL INFO [Service thread] Connection closed by localRank 2
+ywang29-vrdb-test2-worker-0:2472102:2473616 [1] NCCL INFO [Service thread] Connection closed by localRank 1
+ywang29-vrdb-test2-worker-0:2472108:2473204 [7] NCCL INFO comm 0x55630a4bcf40 rank 7 nranks 8 cudaDev 7 busId a01d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 7] NCCL watchdog thread terminated with exception: [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800802 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 7] NCCL watchdog thread terminated with exception: [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800802 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2472106:2473200 [5] NCCL INFO comm 0x557f16fce170 rank 5 nranks 8 cudaDev 5 busId 901d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 5] NCCL watchdog thread terminated with exception: [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800090 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 5] NCCL watchdog thread terminated with exception: [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800090 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2472104:2473198 [3] NCCL INFO comm 0x55b9625e5450 rank 3 nranks 8 cudaDev 3 busId 201d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+ywang29-vrdb-test2-worker-0:2472102:2473188 [1] NCCL INFO comm 0x5651408f8000 rank 1 nranks 8 cudaDev 1 busId 101d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800025 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+ywang29-vrdb-test2-worker-0:2472103:2473187 [2] NCCL INFO comm 0x56184f7bdc90 rank 2 nranks 8 cudaDev 2 busId 201c0 - Abort COMPLETE
+  what():  [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800025 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 1] NCCL watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800105 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 1] NCCL watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800105 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:915] [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800351 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800351 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2472105:2473185 [4] NCCL INFO comm 0x5652b99b3940 rank 4 nranks 8 cudaDev 4 busId 901c0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 4] NCCL watchdog thread terminated with exception: [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800251 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 4] NCCL watchdog thread terminated with exception: [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800251 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2472107:2473206 [6] NCCL INFO comm 0x55e9dda8cf10 rank 6 nranks 8 cudaDev 6 busId a01c0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 6] NCCL watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800907 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 6] NCCL watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=871019, OpType=_ALLGATHER_BASE, NumelIn=189648, NumelOut=1517184, Timeout(ms)=1800000) ran for 1800907 milliseconds before timing out.
+[2025-10-16 08:57:00,095] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472101
+[2025-10-16 08:57:01,003] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472102
+[2025-10-16 08:57:01,003] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472103
+[2025-10-16 08:57:03,019] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472104
+[2025-10-16 08:57:03,022] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472105
+[2025-10-16 08:57:03,024] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472106
+[2025-10-16 08:57:03,026] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472107
+[2025-10-16 08:57:03,028] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2472108
+[2025-10-16 08:57:03,030] [ERROR] [launch.py:322:sigkill_handler] ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1'] exits with return code = -6
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.3_2e-1_connector-1.0_2.3_2e-1_ablation_20251016_065755.log
+Timestamp: 2025-10-16 08:57:03
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation_20251016_085708.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation_20251016_085708.log
new file mode 100644
index 0000000000000000000000000000000000000000..c1d54f13df753fb9b870f723d08311207d60a15c
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation_20251016_085708.log
@@ -0,0 +1,1144 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation_20251016_085708.log
+Timestamp: 2025-10-16 08:57:08
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 08:57:11,262] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:14,699] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 08:57:14,700] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 1.0 --temperature_attn_text 2.9 --temperature_mlp_text 2.9 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 1.0 --temperature_attn_vision 2.9 --temperature_mlp_vision 2.9 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 1.0 --temperature_connector 2.9 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 08:57:17,274] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:18,338] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 08:57:18,338] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 08:57:18,338] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 08:57:18,338] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 08:57:18,338] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 08:57:18,338] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 08:57:18,338] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 08:57:18,340] [INFO] [launch.py:253:main] process 2494148 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,342] [INFO] [launch.py:253:main] process 2494149 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,344] [INFO] [launch.py:253:main] process 2494150 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,346] [INFO] [launch.py:253:main] process 2494151 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,348] [INFO] [launch.py:253:main] process 2494152 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,349] [INFO] [launch.py:253:main] process 2494153 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,351] [INFO] [launch.py:253:main] process 2494154 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 08:57:18,353] [INFO] [launch.py:253:main] process 2494155 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 08:57:25,065] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,228] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,322] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,323] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,323] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,329] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,360] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,386] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 08:57:25,837] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,837] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,837] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,837] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,837] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,837] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 08:57:25,837] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,845] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 08:57:25,845] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 2.9, 'temperature_mlp': 2.9, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 2.9, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 2.9,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 2.9,
+    "temperature_mlp": 2.9,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2494148:2494148 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494148:2494148 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494148:2494148 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494148:2494148 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494148:2494148 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494148:2494148 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2494152:2494152 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494152:2494152 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494152:2494152 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494152:2494152 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494152:2494152 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494152:2494152 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494153:2494153 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494153:2494153 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494153:2494153 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494153:2494153 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494153:2494153 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494153:2494153 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494155:2494155 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494155:2494155 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494155:2494155 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494155:2494155 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494155:2494155 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494155:2494155 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494151:2494151 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494151:2494151 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494151:2494151 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494151:2494151 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494151:2494151 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494151:2494151 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494149:2494149 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494149:2494149 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494149:2494149 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494154:2494154 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494154:2494154 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494149:2494149 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494149:2494149 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494149:2494149 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494154:2494154 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494154:2494154 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494154:2494154 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494154:2494154 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494150:2494150 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2494150:2494150 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494150:2494150 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494150:2494150 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2494150:2494150 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2494150:2494150 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO ncclCommInitRank comm 0x5651567f1390 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO ncclCommInitRank comm 0x5569fdd8dd80 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO ncclCommInitRank comm 0x55e0c9c60a00 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO ncclCommInitRank comm 0x55ff178a3630 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO ncclCommInitRank comm 0x563e4d36b5b0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO ncclCommInitRank comm 0x55d6f8e18f40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO ncclCommInitRank comm 0x560fa1280710 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO ncclCommInitRank comm 0x55e9050e1f00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xaf0e7f0507422bd0 - Init START
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO comm 0x55d6f8e18f40 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO comm 0x55ff178a3630 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO comm 0x5569fdd8dd80 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO comm 0x55e0c9c60a00 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO comm 0x560fa1280710 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO comm 0x5651567f1390 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO comm 0x55e9050e1f00 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO comm 0x563e4d36b5b0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494152:2495755 [4] NCCL INFO ncclCommInitRank comm 0x55e9050e1f00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494154:2495760 [6] NCCL INFO ncclCommInitRank comm 0x5651567f1390 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494150:2495761 [2] NCCL INFO ncclCommInitRank comm 0x55e0c9c60a00 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494148:2495754 [0] NCCL INFO ncclCommInitRank comm 0x563e4d36b5b0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494155:2495757 [7] NCCL INFO ncclCommInitRank comm 0x55ff178a3630 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494151:2495758 [3] NCCL INFO ncclCommInitRank comm 0x55d6f8e18f40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494153:2495756 [5] NCCL INFO ncclCommInitRank comm 0x5569fdd8dd80 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2494149:2495759 [1] NCCL INFO ncclCommInitRank comm 0x560fa1280710 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xaf0e7f0507422bd0 - Init COMPLETE
+[2025-10-16 08:58:09,261] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 08:58:11,063] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=1.000000[E ProcessGroupNCCL.cpp:474] [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800031 milliseconds before timing out.
+
+[E ProcessGroupNCCL.cpp:474] [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800512 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800553 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800596 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800635 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800701 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800845 milliseconds before timing out.
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=1.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=1.000000
+ywang29-vrdb-test2-worker-0:2494153:2495765 [5] NCCL INFO [Service thread] Connection closed by localRank 5
+ywang29-vrdb-test2-worker-0:2494152:2495770 [4] NCCL INFO [Service thread] Connection closed by localRank 4
+ywang29-vrdb-test2-worker-0:2494150:2495764 [2] NCCL INFO [Service thread] Connection closed by localRank 2
+ywang29-vrdb-test2-worker-0:2494151:2495769 [3] NCCL INFO [Service thread] Connection closed by localRank 3
+ywang29-vrdb-test2-worker-0:2494154:2495771 [6] NCCL INFO [Service thread] Connection closed by localRank 6
+ywang29-vrdb-test2-worker-0:2494149:2495768 [1] NCCL INFO [Service thread] Connection closed by localRank 1
+ywang29-vrdb-test2-worker-0:2494155:2495762 [7] NCCL INFO [Service thread] Connection closed by localRank 7
+Pre-training init connector._connector.0.scores: Mean=1.000005
+Pre-training init connector._connector.2.scores: Mean=0.999970
+ywang29-vrdb-test2-worker-0:2494155:2495237 [7] NCCL INFO comm 0x55ff178a3630 rank 7 nranks 8 cudaDev 7 busId a01d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 7] NCCL watchdog thread terminated with exception: [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800701 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 7] NCCL watchdog thread terminated with exception: [Rank 7] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800701 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2494153:2495267 [5] NCCL INFO comm 0x5569fdd8dd80 rank 5 nranks 8 cudaDev 5 busId 901d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 5] NCCL watchdog thread terminated with exception: [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800512 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 5] NCCL watchdog thread terminated with exception: [Rank 5] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800512 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2494152:2495236 [4] NCCL INFO comm 0x55e9050e1f00 rank 4 nranks 8 cudaDev 4 busId 901c0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 4] NCCL watchdog thread terminated with exception: [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800845 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 4] NCCL watchdog thread terminated with exception: [Rank 4] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800845 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2494154:2495265 [6] NCCL INFO comm 0x5651567f1390 rank 6 nranks 8 cudaDev 6 busId a01c0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 6] NCCL watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800031 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 6] NCCL watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800031 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2494151:2495235 [3] NCCL INFO comm 0x55d6f8e18f40 rank 3 nranks 8 cudaDev 3 busId 201d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800635 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800635 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2494150:2495266 [2] NCCL INFO comm 0x55e0c9c60a00 rank 2 nranks 8 cudaDev 2 busId 201c0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800596 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800596 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2494149:2495264 [1] NCCL INFO comm 0x560fa1280710 rank 1 nranks 8 cudaDev 1 busId 101d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 1] NCCL watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800553 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 1] NCCL watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=3660, OpType=BROADCAST, NumelIn=136134656, NumelOut=136134656, Timeout(ms)=1800000) ran for 1800553 milliseconds before timing out.
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+[2025-10-16 09:28:33,324] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494148
+[2025-10-16 09:28:33,984] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494149
+[2025-10-16 09:28:34,358] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494150
+[2025-10-16 09:28:34,360] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494151
+[2025-10-16 09:28:34,360] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494152
+[2025-10-16 09:28:34,453] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494153
+[2025-10-16 09:28:34,454] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494154
+[2025-10-16 09:28:34,455] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2494155
+[2025-10-16 09:28:34,457] [ERROR] [launch.py:322:sigkill_handler] ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '1.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '1.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '1.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1'] exits with return code = -6
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-1.0_2.9_2e-1_connector-1.0_2.9_2e-1_ablation_20251016_085708.log
+Timestamp: 2025-10-16 09:28:35
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation_20251016_092835.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation_20251016_092835.log
new file mode 100644
index 0000000000000000000000000000000000000000..e9d70d72a404d4aac987fa8daef4460b0d084844
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation_20251016_092835.log
@@ -0,0 +1,2312 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation_20251016_092835.log
+Timestamp: 2025-10-16 09:28:35
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 09:28:38,398] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:41,161] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 09:28:41,162] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 3.0 --temperature_attn_text 0.5 --temperature_mlp_text 0.5 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 3.0 --temperature_attn_vision 0.5 --temperature_mlp_vision 0.5 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 3.0 --temperature_connector 0.5 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 09:28:43,783] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:44,840] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 09:28:44,840] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 09:28:44,840] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 09:28:44,840] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 09:28:44,840] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 09:28:44,840] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 09:28:44,840] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 09:28:44,843] [INFO] [launch.py:253:main] process 2496545 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,845] [INFO] [launch.py:253:main] process 2496546 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,846] [INFO] [launch.py:253:main] process 2496547 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,848] [INFO] [launch.py:253:main] process 2496548 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,850] [INFO] [launch.py:253:main] process 2496549 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,852] [INFO] [launch.py:253:main] process 2496550 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,854] [INFO] [launch.py:253:main] process 2496551 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 09:28:44,856] [INFO] [launch.py:253:main] process 2496552 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '0.5', '--temperature_mlp_text', '0.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '0.5', '--temperature_mlp_vision', '0.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '0.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 09:28:51,652] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,698] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,748] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,749] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,758] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,768] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,779] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:51,798] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 09:28:52,073] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,110] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,160] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,164] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,170] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,170] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 09:28:52,184] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,197] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 09:28:52,209] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 0.5, 'temperature_mlp': 0.5, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 0.5, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 0.5,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 0.5,
+    "temperature_mlp": 0.5,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2496545:2496545 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496545:2496545 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496545:2496545 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496545:2496545 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496545:2496545 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496545:2496545 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2496550:2496550 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496550:2496550 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496550:2496550 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496552:2496552 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496552:2496552 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496552:2496552 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496546:2496546 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496550:2496550 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496550:2496550 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496550:2496550 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496546:2496546 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496546:2496546 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496552:2496552 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496552:2496552 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496552:2496552 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496546:2496546 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496546:2496546 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496546:2496546 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496549:2496549 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496549:2496549 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496549:2496549 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496549:2496549 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496549:2496549 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496549:2496549 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2496551:2496551 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496551:2496551 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496551:2496551 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496551:2496551 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496551:2496551 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496551:2496551 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2496548:2496548 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496548:2496548 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496548:2496548 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496548:2496548 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496548:2496548 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496548:2496548 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496547:2496547 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2496547:2496547 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496547:2496547 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496547:2496547 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2496547:2496547 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2496547:2496547 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO ncclCommInitRank comm 0x558897112600 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO ncclCommInitRank comm 0x55ea9f884270 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO ncclCommInitRank comm 0x55b54750dbf0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO ncclCommInitRank comm 0x55d6687a5060 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO ncclCommInitRank comm 0x559c5e9bafb0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO ncclCommInitRank comm 0x55afb1e0c490 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO ncclCommInitRank comm 0x55f82e0b1310 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO ncclCommInitRank comm 0x56368bc5e640 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x8b3e19bd5e269d2f - Init START
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO comm 0x558897112600 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO comm 0x55d6687a5060 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO comm 0x56368bc5e640 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO comm 0x55b54750dbf0 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO comm 0x55afb1e0c490 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO comm 0x559c5e9bafb0 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO comm 0x55ea9f884270 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO comm 0x55f82e0b1310 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496549:2498159 [4] NCCL INFO ncclCommInitRank comm 0x55b54750dbf0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496552:2498157 [7] NCCL INFO ncclCommInitRank comm 0x55afb1e0c490 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496550:2498156 [5] NCCL INFO ncclCommInitRank comm 0x558897112600 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496545:2498155 [0] NCCL INFO ncclCommInitRank comm 0x55f82e0b1310 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496548:2498161 [3] NCCL INFO ncclCommInitRank comm 0x55ea9f884270 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496546:2498158 [1] NCCL INFO ncclCommInitRank comm 0x559c5e9bafb0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496551:2498160 [6] NCCL INFO ncclCommInitRank comm 0x55d6687a5060 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2496547:2498162 [2] NCCL INFO ncclCommInitRank comm 0x56368bc5e640 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x8b3e19bd5e269d2f - Init COMPLETE
+[2025-10-16 09:29:36,096] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 09:29:38,432] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=3.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=3.000000
+Pre-training init connector._connector.0.scores: Mean=3.000005
+Pre-training init connector._connector.2.scores: Mean=2.999970
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 09:37:19,646 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 09:37:19,652 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO ncclCommInitRank comm 0x7fe8cc06b730 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO ncclCommInitRank comm 0x7f713806b520 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO ncclCommInitRank comm 0x7f260006abb0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO ncclCommInitRank comm 0x7f858006af70 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO ncclCommInitRank comm 0x7ff61c06afe0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO ncclCommInitRank comm 0x7fa13406a580 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO ncclCommInitRank comm 0x7f3bcc06b550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO ncclCommInitRank comm 0x7f657406b070 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xe5f79dba64bd3dad - Init START
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO comm 0x7fe8cc06b730 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO comm 0x7f713806b520 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO comm 0x7f260006abb0 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO comm 0x7f858006af70 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO comm 0x7ff61c06afe0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO comm 0x7f657406b070 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO comm 0x7f3bcc06b550 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO comm 0x7fa13406a580 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2496551:2503342 [6] NCCL INFO ncclCommInitRank comm 0x7fe8cc06b730 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496549:2503347 [4] NCCL INFO ncclCommInitRank comm 0x7f260006abb0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496547:2503344 [2] NCCL INFO ncclCommInitRank comm 0x7ff61c06afe0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496545:2503340 [0] NCCL INFO ncclCommInitRank comm 0x7f3bcc06b550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496552:2503346 [7] NCCL INFO ncclCommInitRank comm 0x7fa13406a580 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496548:2503343 [3] NCCL INFO ncclCommInitRank comm 0x7f858006af70 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496550:2503341 [5] NCCL INFO ncclCommInitRank comm 0x7f713806b520 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2496546:2503345 [1] NCCL INFO ncclCommInitRank comm 0x7f657406b070 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xe5f79dba64bd3dad - Init COMPLETE
+
  0%|          | 1/520 [00:14<2:06:22, 14.61s/it]
                                                 
{'loss': 2.0453, 'grad_norm': 0.004835173439554429, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:14<2:06:22, 14.61s/it]
  0%|          | 2/520 [00:18<1:10:28,  8.16s/it]
                                                 
{'loss': 2.0549, 'grad_norm': 0.005249298168671081, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:18<1:10:28,  8.16s/it]
  1%|          | 3/520 [00:22<53:05,  6.16s/it]  
                                               
{'loss': 2.1899, 'grad_norm': 0.006006724812289717, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:22<53:05,  6.16s/it]
  1%|          | 4/520 [00:25<44:56,  5.22s/it]
                                               
{'loss': 2.0656, 'grad_norm': 0.004963647297850456, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:25<44:56,  5.22s/it]
  1%|          | 5/520 [00:29<40:29,  4.72s/it]
                                               
{'loss': 2.2333, 'grad_norm': 0.005481466646426385, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:29<40:29,  4.72s/it]
  1%|          | 6/520 [00:33<37:44,  4.41s/it]
                                               
{'loss': 1.4708, 'grad_norm': 0.0016545336979977329, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:33<37:44,  4.41s/it]
  1%|▏         | 7/520 [00:37<36:00,  4.21s/it]
                                               
{'loss': 1.5589, 'grad_norm': 0.0009775485749206891, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:37<36:00,  4.21s/it]
  2%|▏         | 8/520 [00:41<36:27,  4.27s/it]
                                               
{'loss': 1.5605, 'grad_norm': 0.0006639436939264217, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:41<36:27,  4.27s/it]
  2%|▏         | 9/520 [00:45<36:24,  4.28s/it]
                                               
{'loss': 1.622, 'grad_norm': 0.0005811219140305953, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:45<36:24,  4.28s/it]
  2%|▏         | 10/520 [00:49<35:07,  4.13s/it]
                                                
{'loss': 1.475, 'grad_norm': 0.0005798780749229405, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:49<35:07,  4.13s/it]
  2%|▏         | 11/520 [00:53<34:35,  4.08s/it]
                                                
{'loss': 1.4825, 'grad_norm': 0.0004969166452855392, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [00:53<34:35,  4.08s/it]
  2%|▏         | 12/520 [00:57<33:45,  3.99s/it]
                                                
{'loss': 1.355, 'grad_norm': 0.0004834569565118968, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [00:57<33:45,  3.99s/it][2025-10-16 09:38:25,963] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:01<34:41,  4.11s/it]
                                                
{'loss': 1.4377, 'grad_norm': 0.0005267034271411182, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:01<34:41,  4.11s/it]
  3%|▎         | 14/520 [01:05<33:42,  4.00s/it]
                                                
{'loss': 1.4693, 'grad_norm': 0.0005901577563437861, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:05<33:42,  4.00s/it]
  3%|▎         | 15/520 [01:09<32:59,  3.92s/it]
                                                
{'loss': 1.3818, 'grad_norm': 0.000502071222001285, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:09<32:59,  3.92s/it]
  3%|▎         | 16/520 [01:13<32:33,  3.88s/it]
                                                
{'loss': 1.3549, 'grad_norm': 0.000674181678049769, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:13<32:33,  3.88s/it]
  3%|▎         | 17/520 [01:16<31:59,  3.82s/it]
                                                
{'loss': 1.4887, 'grad_norm': 0.0008135782299271786, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:16<31:59,  3.82s/it]
  3%|▎         | 18/520 [01:20<31:32,  3.77s/it]
                                                
{'loss': 1.3437, 'grad_norm': 0.0009026311261461994, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:20<31:32,  3.77s/it]
  4%|▎         | 19/520 [01:24<31:12,  3.74s/it]
                                                
{'loss': 1.3411, 'grad_norm': 0.0008722664886884243, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:24<31:12,  3.74s/it]
  4%|▍         | 20/520 [01:27<30:53,  3.71s/it]
                                                
{'loss': 1.31, 'grad_norm': 0.0010026654542762334, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:27<30:53,  3.71s/it]
  4%|▍         | 21/520 [01:31<30:46,  3.70s/it]
                                                
{'loss': 1.3237, 'grad_norm': 0.0011511544261495082, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:31<30:46,  3.70s/it]
  4%|▍         | 22/520 [01:35<30:32,  3.68s/it]
                                                
{'loss': 1.4427, 'grad_norm': 0.0011309936329361335, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:35<30:32,  3.68s/it]
  4%|▍         | 23/520 [01:38<30:36,  3.69s/it]
                                                
{'loss': 1.3966, 'grad_norm': 0.0014873788916890233, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:38<30:36,  3.69s/it]
  5%|▍         | 24/520 [01:42<30:35,  3.70s/it]
                                                
{'loss': 1.309, 'grad_norm': 0.0012012935399482642, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [01:42<30:35,  3.70s/it]
  5%|▍         | 25/520 [01:46<30:33,  3.71s/it]
                                                
{'loss': 1.3906, 'grad_norm': 0.001567046442687115, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [01:46<30:33,  3.71s/it]
  5%|▌         | 26/520 [01:49<30:31,  3.71s/it]
                                                
{'loss': 1.3285, 'grad_norm': 0.0011100318901994676, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [01:49<30:31,  3.71s/it]
  5%|▌         | 27/520 [01:53<30:26,  3.70s/it]
                                                
{'loss': 1.27, 'grad_norm': 0.0013762236271678365, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [01:53<30:26,  3.70s/it]
  5%|▌         | 28/520 [01:57<30:20,  3.70s/it]
                                                
{'loss': 1.2915, 'grad_norm': 0.001429935609422655, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [01:57<30:20,  3.70s/it]
  6%|▌         | 29/520 [02:01<30:11,  3.69s/it]
                                                
{'loss': 1.3096, 'grad_norm': 0.0014725161848310312, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [02:01<30:11,  3.69s/it]
  6%|▌         | 30/520 [02:04<29:59,  3.67s/it]
                                                
{'loss': 1.3788, 'grad_norm': 0.0012823257274306187, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:04<29:59,  3.67s/it]
  6%|▌         | 31/520 [02:08<29:56,  3.67s/it]
                                                
{'loss': 1.2796, 'grad_norm': 0.0013408495959984163, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:08<29:56,  3.67s/it]
  6%|▌         | 32/520 [02:12<30:08,  3.71s/it]
                                                
{'loss': 1.2147, 'grad_norm': 0.0014615839028519443, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:12<30:08,  3.71s/it]
  6%|▋         | 33/520 [02:15<30:06,  3.71s/it]
                                                
{'loss': 1.2835, 'grad_norm': 0.0016780476655557368, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:15<30:06,  3.71s/it]
  7%|▋         | 34/520 [02:19<30:06,  3.72s/it]
                                                
{'loss': 1.2802, 'grad_norm': 0.0017596587330199957, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:19<30:06,  3.72s/it]
  7%|▋         | 35/520 [02:23<30:00,  3.71s/it]
                                                
{'loss': 1.2826, 'grad_norm': 0.0018685025967863213, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:23<30:00,  3.71s/it]
  7%|▋         | 36/520 [02:27<30:11,  3.74s/it]
                                                
{'loss': 1.3828, 'grad_norm': 0.0016195942096558266, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:27<30:11,  3.74s/it]
  7%|▋         | 37/520 [02:30<30:13,  3.75s/it]
                                                
{'loss': 1.3672, 'grad_norm': 0.0015729141186882193, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:30<30:13,  3.75s/it]
  7%|▋         | 38/520 [02:34<30:15,  3.77s/it]
                                                
{'loss': 1.4599, 'grad_norm': 0.0017277795293385593, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:34<30:15,  3.77s/it]
  8%|▊         | 39/520 [02:38<30:12,  3.77s/it]
                                                
{'loss': 1.3221, 'grad_norm': 0.002002165629780975, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [02:38<30:12,  3.77s/it]
  8%|▊         | 40/520 [02:42<29:59,  3.75s/it]
                                                
{'loss': 1.347, 'grad_norm': 0.0016242322307702397, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [02:42<29:59,  3.75s/it]
  8%|▊         | 41/520 [02:45<29:48,  3.73s/it]
                                                
{'loss': 1.3292, 'grad_norm': 0.0017219144238744498, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [02:45<29:48,  3.73s/it]
  8%|▊         | 42/520 [02:49<29:33,  3.71s/it]
                                                
{'loss': 1.3383, 'grad_norm': 0.0023478632275731725, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [02:49<29:33,  3.71s/it]
  8%|▊         | 43/520 [02:53<29:26,  3.70s/it]
                                                
{'loss': 1.2718, 'grad_norm': 0.0015337870635359664, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [02:53<29:26,  3.70s/it]
  8%|▊         | 44/520 [02:56<29:19,  3.70s/it]
                                                
{'loss': 1.3702, 'grad_norm': 0.001722321433491021, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [02:56<29:19,  3.70s/it]
  9%|▊         | 45/520 [03:00<29:09,  3.68s/it]
                                                
{'loss': 1.3623, 'grad_norm': 0.001983084777190461, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [03:00<29:09,  3.68s/it]
  9%|▉         | 46/520 [03:04<29:06,  3.69s/it]
                                                
{'loss': 1.4214, 'grad_norm': 0.0018800111918549419, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:04<29:06,  3.69s/it]
  9%|▉         | 47/520 [03:07<28:58,  3.67s/it]
                                                
{'loss': 1.3271, 'grad_norm': 0.0018341561031374872, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:07<28:58,  3.67s/it]
  9%|▉         | 48/520 [03:11<28:49,  3.66s/it]
                                                
{'loss': 1.3151, 'grad_norm': 0.0020767712006008066, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:11<28:49,  3.66s/it]
  9%|▉         | 49/520 [03:15<28:45,  3.66s/it]
                                                
{'loss': 1.373, 'grad_norm': 0.001997708262346234, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:15<28:45,  3.66s/it]
 10%|▉         | 50/520 [03:18<28:44,  3.67s/it]
                                                
{'loss': 1.3681, 'grad_norm': 0.0019128258329323795, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:18<28:44,  3.67s/it]
 10%|▉         | 51/520 [03:22<28:40,  3.67s/it]
                                                
{'loss': 1.2971, 'grad_norm': 0.002014642376791284, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:22<28:40,  3.67s/it]
 10%|█         | 52/520 [03:26<28:36,  3.67s/it]
                                                
{'loss': 1.4326, 'grad_norm': 0.0023055247058359893, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:26<28:36,  3.67s/it]
 10%|█         | 53/520 [03:29<28:31,  3.66s/it]
                                                
{'loss': 1.4007, 'grad_norm': 0.0018781875706317375, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:29<28:31,  3.66s/it]
 10%|█         | 54/520 [03:33<28:22,  3.65s/it]
                                                
{'loss': 1.3329, 'grad_norm': 0.001772220591563359, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:33<28:22,  3.65s/it]
 11%|█         | 55/520 [03:37<28:28,  3.67s/it]
                                                
{'loss': 1.3023, 'grad_norm': 0.0019679019555692047, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [03:37<28:28,  3.67s/it]
 11%|█         | 56/520 [03:40<28:31,  3.69s/it]
                                                
{'loss': 1.4334, 'grad_norm': 0.0020193665980659183, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [03:40<28:31,  3.69s/it]
 11%|█         | 57/520 [03:44<28:21,  3.68s/it]
                                                
{'loss': 1.2875, 'grad_norm': 0.0019857712451677944, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [03:44<28:21,  3.68s/it]
 11%|█         | 58/520 [03:48<28:10,  3.66s/it]
                                                
{'loss': 1.4505, 'grad_norm': 0.0016179664366174486, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [03:48<28:10,  3.66s/it]
 11%|█▏        | 59/520 [03:51<28:06,  3.66s/it]
                                                
{'loss': 1.2663, 'grad_norm': 0.0017183631220196562, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [03:51<28:06,  3.66s/it]
 12%|█▏        | 60/520 [03:55<28:00,  3.65s/it]
                                                
{'loss': 1.364, 'grad_norm': 0.0017093924987143857, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [03:55<28:00,  3.65s/it]
 12%|█▏        | 61/520 [03:59<27:59,  3.66s/it]
                                                
{'loss': 1.3445, 'grad_norm': 0.0018084797023183658, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [03:59<27:59,  3.66s/it]
 12%|█▏        | 62/520 [04:02<27:57,  3.66s/it]
                                                
{'loss': 1.3445, 'grad_norm': 0.0018277960074820961, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [04:02<27:57,  3.66s/it]
 12%|█▏        | 63/520 [04:06<27:53,  3.66s/it]
                                                
{'loss': 1.3266, 'grad_norm': 0.001819497815322539, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:06<27:53,  3.66s/it]
 12%|█▏        | 64/520 [04:10<28:12,  3.71s/it]
                                                
{'loss': 1.3564, 'grad_norm': 0.0018269958838543867, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:10<28:12,  3.71s/it]
 12%|█▎        | 65/520 [04:14<28:28,  3.76s/it]
                                                
{'loss': 1.3718, 'grad_norm': 0.0019948065243176206, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:14<28:28,  3.76s/it]
 13%|█▎        | 66/520 [04:17<28:37,  3.78s/it]
                                                
{'loss': 1.322, 'grad_norm': 0.0016450500285691292, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:17<28:37,  3.78s/it]
 13%|█▎        | 67/520 [04:21<28:42,  3.80s/it]
                                                
{'loss': 1.2261, 'grad_norm': 0.0017057002161580842, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:21<28:42,  3.80s/it]
 13%|█▎        | 68/520 [04:25<28:34,  3.79s/it]
                                                
{'loss': 1.2884, 'grad_norm': 0.0016373352817325073, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:25<28:34,  3.79s/it]
 13%|█▎        | 69/520 [04:29<28:10,  3.75s/it]
                                                
{'loss': 1.2657, 'grad_norm': 0.0017042329748290777, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:29<28:10,  3.75s/it]
 13%|█▎        | 70/520 [04:32<27:59,  3.73s/it]
                                                
{'loss': 1.3066, 'grad_norm': 0.001770430176637663, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [04:32<27:59,  3.73s/it]
 14%|█▎        | 71/520 [04:36<27:47,  3.71s/it]
                                                
{'loss': 1.2454, 'grad_norm': 0.0015967991445694, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [04:36<27:47,  3.71s/it]
 14%|█▍        | 72/520 [04:40<27:38,  3.70s/it]
                                                
{'loss': 1.3868, 'grad_norm': 0.001667701906717724, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [04:40<27:38,  3.70s/it]
 14%|█▍        | 73/520 [04:44<27:40,  3.72s/it]
                                                
{'loss': 1.2197, 'grad_norm': 0.0016477380561292737, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [04:44<27:40,  3.72s/it]
 14%|█▍        | 74/520 [04:47<27:43,  3.73s/it]
                                                
{'loss': 1.3274, 'grad_norm': 0.0017718560109305998, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [04:47<27:43,  3.73s/it]
 14%|█▍        | 75/520 [04:51<27:53,  3.76s/it]
                                                
{'loss': 1.2401, 'grad_norm': 0.0015108023124992978, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [04:51<27:53,  3.76s/it]
 15%|█▍        | 76/520 [04:55<27:52,  3.77s/it]
                                                
{'loss': 1.3962, 'grad_norm': 0.0014685283359830536, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [04:55<27:52,  3.77s/it]
 15%|█▍        | 77/520 [04:59<27:51,  3.77s/it]
                                                
{'loss': 1.1708, 'grad_norm': 0.0017817837330501491, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [04:59<27:51,  3.77s/it]
 15%|█▌        | 78/520 [05:03<27:52,  3.78s/it]
                                                
{'loss': 1.2822, 'grad_norm': 0.0017979791558982452, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:03<27:52,  3.78s/it]
 15%|█▌        | 79/520 [05:06<27:32,  3.75s/it]
                                                
{'loss': 1.2694, 'grad_norm': 0.0016313274063878317, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:06<27:32,  3.75s/it]
 15%|█▌        | 80/520 [05:10<27:19,  3.73s/it]
                                                
{'loss': 1.3901, 'grad_norm': 0.0017535755316351567, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:10<27:19,  3.73s/it]
 16%|█▌        | 81/520 [05:13<27:05,  3.70s/it]
                                                
{'loss': 1.4166, 'grad_norm': 0.0020327807773369974, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:13<27:05,  3.70s/it]
 16%|█▌        | 82/520 [05:17<26:58,  3.69s/it]
                                                
{'loss': 1.3358, 'grad_norm': 0.001556230553882814, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:17<26:58,  3.69s/it]
 16%|█▌        | 83/520 [05:21<26:51,  3.69s/it]
                                                
{'loss': 1.354, 'grad_norm': 0.001649927675725719, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:21<26:51,  3.69s/it]
 16%|█▌        | 84/520 [05:24<26:42,  3.68s/it]
                                                
{'loss': 1.3593, 'grad_norm': 0.001636957875335679, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:24<26:42,  3.68s/it]
 16%|█▋        | 85/520 [05:28<26:35,  3.67s/it]
                                                
{'loss': 1.3925, 'grad_norm': 0.0015230046676140425, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:28<26:35,  3.67s/it]
 17%|█▋        | 86/520 [05:32<26:42,  3.69s/it]
                                                
{'loss': 1.3899, 'grad_norm': 0.001519998684285411, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:32<26:42,  3.69s/it]
 17%|█▋        | 87/520 [05:36<26:36,  3.69s/it]
                                                
{'loss': 1.3338, 'grad_norm': 0.0015139688694013737, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [05:36<26:36,  3.69s/it]
 17%|█▋        | 88/520 [05:39<26:52,  3.73s/it]
                                                
{'loss': 1.2789, 'grad_norm': 0.00127285544749555, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [05:39<26:52,  3.73s/it]
 17%|█▋        | 89/520 [05:43<27:11,  3.79s/it]
                                                
{'loss': 1.3415, 'grad_norm': 0.001540428349709519, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [05:43<27:11,  3.79s/it]
 17%|█▋        | 90/520 [05:47<27:15,  3.80s/it]
                                                
{'loss': 1.2777, 'grad_norm': 0.0014869008123450413, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [05:47<27:15,  3.80s/it]
 18%|█▊        | 91/520 [05:51<27:18,  3.82s/it]
                                                
{'loss': 1.3509, 'grad_norm': 0.001427054831305826, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [05:51<27:18,  3.82s/it]
 18%|█▊        | 92/520 [05:55<27:20,  3.83s/it]
                                                
{'loss': 1.2863, 'grad_norm': 0.0015727609695058047, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [05:55<27:20,  3.83s/it]
 18%|█▊        | 93/520 [05:59<27:07,  3.81s/it]
                                                
{'loss': 1.3019, 'grad_norm': 0.0015402118143427797, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [05:59<27:07,  3.81s/it]
 18%|█▊        | 94/520 [06:02<26:46,  3.77s/it]
                                                
{'loss': 1.3856, 'grad_norm': 0.0014656893505991012, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [06:02<26:46,  3.77s/it]
 18%|█▊        | 95/520 [06:06<26:29,  3.74s/it]
                                                
{'loss': 1.2739, 'grad_norm': 0.0016826830685728782, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:06<26:29,  3.74s/it]
 18%|█▊        | 96/520 [06:10<26:18,  3.72s/it]
                                                
{'loss': 1.2962, 'grad_norm': 0.0013110905020461638, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:10<26:18,  3.72s/it]
 19%|█▊        | 97/520 [06:13<26:16,  3.73s/it]
                                                
{'loss': 1.258, 'grad_norm': 0.0015311828098053689, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:13<26:16,  3.73s/it]
 19%|█▉        | 98/520 [06:17<26:09,  3.72s/it]
                                                
{'loss': 1.2584, 'grad_norm': 0.001170579758388857, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:17<26:09,  3.72s/it]
 19%|█▉        | 99/520 [06:21<25:58,  3.70s/it]
                                                
{'loss': 1.2701, 'grad_norm': 0.0014805162119030217, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:21<25:58,  3.70s/it]
 19%|█▉        | 100/520 [06:25<26:02,  3.72s/it]
                                                 
{'loss': 1.2676, 'grad_norm': 0.0013297710811298268, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:25<26:02,  3.72s/it]
 19%|█▉        | 101/520 [06:28<25:52,  3.70s/it]
                                                 
{'loss': 1.2813, 'grad_norm': 0.0014112850927475362, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:28<25:52,  3.70s/it]
 20%|█▉        | 102/520 [06:32<26:01,  3.74s/it]
                                                 
{'loss': 1.2829, 'grad_norm': 0.0015499877671574162, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [06:32<26:01,  3.74s/it]
 20%|█▉        | 103/520 [06:36<26:13,  3.77s/it]
                                                 
{'loss': 1.2094, 'grad_norm': 0.0012750588721874814, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [06:36<26:13,  3.77s/it]
 20%|██        | 104/520 [06:40<26:24,  3.81s/it]
                                                 
{'loss': 1.2871, 'grad_norm': 0.001410752389241349, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [06:40<26:24,  3.81s/it]
 20%|██        | 105/520 [06:43<26:10,  3.78s/it]
                                                 
{'loss': 1.2815, 'grad_norm': 0.0013650601337415814, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [06:43<26:10,  3.78s/it]
 20%|██        | 106/520 [06:47<25:55,  3.76s/it]
                                                 
{'loss': 1.3005, 'grad_norm': 0.00130102431741545, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [06:47<25:55,  3.76s/it]
 21%|██        | 107/520 [06:51<25:44,  3.74s/it]
                                                 
{'loss': 1.262, 'grad_norm': 0.0013066648978432465, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [06:51<25:44,  3.74s/it]
 21%|██        | 108/520 [06:55<25:36,  3.73s/it]
                                                 
{'loss': 1.2306, 'grad_norm': 0.0013398262248142299, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [06:55<25:36,  3.73s/it]
 21%|██        | 109/520 [06:58<25:25,  3.71s/it]
                                                 
{'loss': 1.2447, 'grad_norm': 0.0011902552068001448, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [06:58<25:25,  3.71s/it]
 21%|██        | 110/520 [07:02<25:13,  3.69s/it]
                                                 
{'loss': 1.4222, 'grad_norm': 0.0013692998756103167, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [07:02<25:13,  3.69s/it]
 21%|██▏       | 111/520 [07:06<25:11,  3.70s/it]
                                                 
{'loss': 1.4139, 'grad_norm': 0.0014682327942140553, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:06<25:11,  3.70s/it]
 22%|██▏       | 112/520 [07:09<25:26,  3.74s/it]
                                                 
{'loss': 1.295, 'grad_norm': 0.0012676047202517884, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:09<25:26,  3.74s/it]
 22%|██▏       | 113/520 [07:13<25:33,  3.77s/it]
                                                 
{'loss': 1.1916, 'grad_norm': 0.0012355528829738611, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:13<25:33,  3.77s/it]
 22%|██▏       | 114/520 [07:17<25:35,  3.78s/it]
                                                 
{'loss': 1.2897, 'grad_norm': 0.0012467095218740864, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:17<25:35,  3.78s/it]
 22%|██▏       | 115/520 [07:21<25:34,  3.79s/it]
                                                 
{'loss': 1.3867, 'grad_norm': 0.0012341550214670237, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:21<25:34,  3.79s/it]
 22%|██▏       | 116/520 [07:25<25:33,  3.79s/it]
                                                 
{'loss': 1.3942, 'grad_norm': 0.0012126847448754915, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:25<25:33,  3.79s/it]
 22%|██▎       | 117/520 [07:29<25:32,  3.80s/it]
                                                 
{'loss': 1.3581, 'grad_norm': 0.0013354467271634173, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:29<25:32,  3.80s/it]
 23%|██▎       | 118/520 [07:32<25:30,  3.81s/it]
                                                 
{'loss': 1.2794, 'grad_norm': 0.0012246607339562697, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [07:32<25:30,  3.81s/it]
 23%|██▎       | 119/520 [07:36<25:25,  3.80s/it]
                                                 
{'loss': 1.2339, 'grad_norm': 0.001282209997532125, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [07:36<25:25,  3.80s/it]
 23%|██▎       | 120/520 [07:40<25:20,  3.80s/it]
                                                 
{'loss': 1.2445, 'grad_norm': 0.0013925032454042343, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [07:40<25:20,  3.80s/it]
 23%|██▎       | 121/520 [07:44<24:58,  3.76s/it]
                                                 
{'loss': 1.2966, 'grad_norm': 0.0012761257641190655, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [07:44<24:58,  3.76s/it]
 23%|██▎       | 122/520 [07:47<24:42,  3.72s/it]
                                                 
{'loss': 1.2024, 'grad_norm': 0.001165797712623182, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [07:47<24:42,  3.72s/it]
 24%|██▎       | 123/520 [07:51<24:29,  3.70s/it]
                                                 
{'loss': 1.3269, 'grad_norm': 0.0012232853840374394, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [07:51<24:29,  3.70s/it]
 24%|██▍       | 124/520 [07:54<24:15,  3.68s/it]
                                                 
{'loss': 1.2615, 'grad_norm': 0.0013337455005205437, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [07:55<24:15,  3.68s/it]
 24%|██▍       | 125/520 [07:58<24:11,  3.68s/it]
                                                 
{'loss': 1.2629, 'grad_norm': 0.0012345759051827637, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [07:58<24:11,  3.68s/it]
 24%|██▍       | 126/520 [08:02<25:24,  3.87s/it]
                                                 
{'loss': 1.2596, 'grad_norm': 0.0010531798888794525, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [08:02<25:24,  3.87s/it]
 24%|██▍       | 127/520 [08:06<25:18,  3.86s/it]
                                                 
{'loss': 1.2418, 'grad_norm': 0.0013430153280233517, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:06<25:18,  3.86s/it]
 25%|██▍       | 128/520 [08:10<25:15,  3.87s/it]
                                                 
{'loss': 1.2837, 'grad_norm': 0.0013100539644706108, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:10<25:15,  3.87s/it]
 25%|██▍       | 129/520 [08:14<25:28,  3.91s/it]
                                                 
{'loss': 1.2463, 'grad_norm': 0.0011299878286388434, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:14<25:28,  3.91s/it]
 25%|██▌       | 130/520 [08:18<25:20,  3.90s/it]
                                                 
{'loss': 1.2764, 'grad_norm': 0.0011406325481873884, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:18<25:20,  3.90s/it]
 25%|██▌       | 131/520 [08:22<25:19,  3.91s/it]
                                                 
{'loss': 1.2337, 'grad_norm': 0.001082294008552181, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:22<25:19,  3.91s/it]
 25%|██▌       | 132/520 [08:26<25:16,  3.91s/it]
                                                 
{'loss': 1.3168, 'grad_norm': 0.0013576664311029342, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:26<25:16,  3.91s/it]
 26%|██▌       | 133/520 [08:30<25:03,  3.89s/it]
                                                 
{'loss': 1.2482, 'grad_norm': 0.001321957762488895, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:30<25:03,  3.89s/it]
 26%|██▌       | 134/520 [08:33<24:39,  3.83s/it]
                                                 
{'loss': 1.3157, 'grad_norm': 0.0011787645183082684, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [08:33<24:39,  3.83s/it]
 26%|██▌       | 135/520 [08:37<24:18,  3.79s/it]
                                                 
{'loss': 1.3687, 'grad_norm': 0.0011894107016105061, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [08:37<24:18,  3.79s/it]
 26%|██▌       | 136/520 [08:41<24:01,  3.75s/it]
                                                 
{'loss': 1.3076, 'grad_norm': 0.0011557135596244367, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [08:41<24:01,  3.75s/it]
 26%|██▋       | 137/520 [08:45<23:53,  3.74s/it]
                                                 
{'loss': 1.224, 'grad_norm': 0.0013869364402673378, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [08:45<23:53,  3.74s/it]
 27%|██▋       | 138/520 [08:48<23:40,  3.72s/it]
                                                 
{'loss': 1.2407, 'grad_norm': 0.0010648241615413424, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [08:48<23:40,  3.72s/it]
 27%|██▋       | 139/520 [08:52<23:35,  3.72s/it]
                                                 
{'loss': 1.1352, 'grad_norm': 0.0010636778645682894, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [08:52<23:35,  3.72s/it]
 27%|██▋       | 140/520 [08:56<23:32,  3.72s/it]
                                                 
{'loss': 1.276, 'grad_norm': 0.001050759869060005, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [08:56<23:32,  3.72s/it]
 27%|██▋       | 141/520 [08:59<23:25,  3.71s/it]
                                                 
{'loss': 1.3445, 'grad_norm': 0.001054146129852929, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [08:59<23:25,  3.71s/it]
 27%|██▋       | 142/520 [09:03<23:19,  3.70s/it]
                                                 
{'loss': 1.2857, 'grad_norm': 0.0010581852227483477, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [09:03<23:19,  3.70s/it]
 28%|██▊       | 143/520 [09:07<23:15,  3.70s/it]
                                                 
{'loss': 1.2726, 'grad_norm': 0.0011936757085911009, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:07<23:15,  3.70s/it]
 28%|██▊       | 144/520 [09:10<23:10,  3.70s/it]
                                                 
{'loss': 1.2356, 'grad_norm': 0.0012273699141591918, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:10<23:10,  3.70s/it]
 28%|██▊       | 145/520 [09:14<23:11,  3.71s/it]
                                                 
{'loss': 1.1727, 'grad_norm': 0.0010682681363601944, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:14<23:11,  3.71s/it]
 28%|██▊       | 146/520 [09:18<23:02,  3.70s/it]
                                                 
{'loss': 1.3439, 'grad_norm': 0.0011367982093810568, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:18<23:02,  3.70s/it]
 28%|██▊       | 147/520 [09:22<23:00,  3.70s/it]
                                                 
{'loss': 1.2129, 'grad_norm': 0.001109729207133023, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:22<23:00,  3.70s/it]
 28%|██▊       | 148/520 [09:25<22:54,  3.69s/it]
                                                 
{'loss': 1.2356, 'grad_norm': 0.0010280990025632517, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:25<22:54,  3.69s/it]
 29%|██▊       | 149/520 [09:29<22:50,  3.69s/it]
                                                 
{'loss': 1.1809, 'grad_norm': 0.0011132715765792213, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:29<22:50,  3.69s/it]
 29%|██▉       | 150/520 [09:33<22:46,  3.69s/it]
                                                 
{'loss': 1.4046, 'grad_norm': 0.001102300337614865, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [09:33<22:46,  3.69s/it]
 29%|██▉       | 151/520 [09:36<22:41,  3.69s/it]
                                                 
{'loss': 1.2255, 'grad_norm': 0.0011370309607524995, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [09:36<22:41,  3.69s/it]
 29%|██▉       | 152/520 [09:40<22:35,  3.68s/it]
                                                 
{'loss': 1.2019, 'grad_norm': 0.0011243766027709886, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [09:40<22:35,  3.68s/it]
 29%|██▉       | 153/520 [09:44<22:22,  3.66s/it]
                                                 
{'loss': 1.2301, 'grad_norm': 0.0010518615016613935, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [09:44<22:22,  3.66s/it]
 30%|██▉       | 154/520 [09:47<22:20,  3.66s/it]
                                                 
{'loss': 1.3201, 'grad_norm': 0.001048067042258406, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [09:47<22:20,  3.66s/it]
 30%|██▉       | 155/520 [09:51<22:16,  3.66s/it]
                                                 
{'loss': 1.2328, 'grad_norm': 0.0011387739903372543, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [09:51<22:16,  3.66s/it]
 30%|███       | 156/520 [09:55<22:12,  3.66s/it]
                                                 
{'loss': 1.2594, 'grad_norm': 0.0014559380121994113, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [09:55<22:12,  3.66s/it]
 30%|███       | 157/520 [09:58<22:07,  3.66s/it]
                                                 
{'loss': 1.3331, 'grad_norm': 0.0010805274166644696, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [09:58<22:07,  3.66s/it]
 30%|███       | 158/520 [10:02<22:05,  3.66s/it]
                                                 
{'loss': 1.2299, 'grad_norm': 0.0010445218757715053, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [10:02<22:05,  3.66s/it]
 31%|███       | 159/520 [10:05<21:59,  3.66s/it]
                                                 
{'loss': 1.2736, 'grad_norm': 0.0010522069761698678, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [10:06<21:59,  3.66s/it]
 31%|███       | 160/520 [10:09<21:57,  3.66s/it]
                                                 
{'loss': 1.2847, 'grad_norm': 0.0011874123247233377, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:09<21:57,  3.66s/it]
 31%|███       | 161/520 [10:13<21:52,  3.66s/it]
                                                 
{'loss': 1.2679, 'grad_norm': 0.0010698067285364323, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:13<21:52,  3.66s/it]
 31%|███       | 162/520 [10:16<21:49,  3.66s/it]
                                                 
{'loss': 1.2749, 'grad_norm': 0.0011184349234381157, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:16<21:49,  3.66s/it]
 31%|███▏      | 163/520 [10:20<21:42,  3.65s/it]
                                                 
{'loss': 1.1656, 'grad_norm': 0.0012051242678843656, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:20<21:42,  3.65s/it]
 32%|███▏      | 164/520 [10:24<21:38,  3.65s/it]
                                                 
{'loss': 1.1276, 'grad_norm': 0.0010010899578148092, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:24<21:38,  3.65s/it]
 32%|███▏      | 165/520 [10:27<21:32,  3.64s/it]
                                                 
{'loss': 1.2692, 'grad_norm': 0.0010002488287759327, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:27<21:32,  3.64s/it]
 32%|███▏      | 166/520 [10:31<21:26,  3.63s/it]
                                                 
{'loss': 1.2392, 'grad_norm': 0.00112802450026933, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:31<21:26,  3.63s/it]
 32%|███▏      | 167/520 [10:35<21:22,  3.63s/it]
                                                 
{'loss': 1.2354, 'grad_norm': 0.001021164266416892, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [10:35<21:22,  3.63s/it]
 32%|███▏      | 168/520 [10:38<21:21,  3.64s/it]
                                                 
{'loss': 1.1812, 'grad_norm': 0.0010155628146834516, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [10:38<21:21,  3.64s/it]
 32%|███▎      | 169/520 [10:42<21:22,  3.65s/it]
                                                 
{'loss': 1.2589, 'grad_norm': 0.0009556972750906273, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [10:42<21:22,  3.65s/it]
 33%|███▎      | 170/520 [10:46<21:20,  3.66s/it]
                                                 
{'loss': 1.2221, 'grad_norm': 0.0008888292975648167, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [10:46<21:20,  3.66s/it]
 33%|███▎      | 171/520 [10:49<21:14,  3.65s/it]
                                                 
{'loss': 1.1923, 'grad_norm': 0.001057783028354286, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [10:49<21:14,  3.65s/it]
 33%|███▎      | 172/520 [10:53<21:09,  3.65s/it]
                                                 
{'loss': 1.268, 'grad_norm': 0.0010627992847538194, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [10:53<21:09,  3.65s/it]
 33%|███▎      | 173/520 [10:57<21:12,  3.67s/it]
                                                 
{'loss': 1.205, 'grad_norm': 0.0010366433503198802, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [10:57<21:12,  3.67s/it]
 33%|███▎      | 174/520 [11:00<21:08,  3.67s/it]
                                                 
{'loss': 1.2577, 'grad_norm': 0.001000254703321372, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [11:00<21:08,  3.67s/it]
 34%|███▎      | 175/520 [11:04<21:05,  3.67s/it]
                                                 
{'loss': 1.1704, 'grad_norm': 0.0010343022238107442, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [11:04<21:05,  3.67s/it]
 34%|███▍      | 176/520 [11:08<21:02,  3.67s/it]
                                                 
{'loss': 1.302, 'grad_norm': 0.0010530240158071781, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:08<21:02,  3.67s/it]
 34%|███▍      | 177/520 [11:11<21:01,  3.68s/it]
                                                 
{'loss': 1.1763, 'grad_norm': 0.0010550359002768567, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:11<21:01,  3.68s/it]
 34%|███▍      | 178/520 [11:15<20:55,  3.67s/it]
                                                 
{'loss': 1.2418, 'grad_norm': 0.0010589057605235837, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:15<20:55,  3.67s/it]
 34%|███▍      | 179/520 [11:19<20:51,  3.67s/it]
                                                 
{'loss': 1.3131, 'grad_norm': 0.0009596998271428809, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:19<20:51,  3.67s/it]
 35%|███▍      | 180/520 [11:22<20:47,  3.67s/it]
                                                 
{'loss': 1.2342, 'grad_norm': 0.0010196420466146775, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:22<20:47,  3.67s/it]
 35%|███▍      | 181/520 [11:26<20:44,  3.67s/it]
                                                 
{'loss': 1.2136, 'grad_norm': 0.0008835321520350211, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:26<20:44,  3.67s/it]
 35%|███▌      | 182/520 [11:30<20:40,  3.67s/it]
                                                 
{'loss': 1.2269, 'grad_norm': 0.0010392113267317393, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:30<20:40,  3.67s/it]
 35%|███▌      | 183/520 [11:33<20:38,  3.67s/it]
                                                 
{'loss': 1.2458, 'grad_norm': 0.0009676361394018879, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [11:33<20:38,  3.67s/it]
 35%|███▌      | 184/520 [11:37<20:36,  3.68s/it]
                                                 
{'loss': 1.1782, 'grad_norm': 0.0010387712852316159, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [11:37<20:36,  3.68s/it]
 36%|███▌      | 185/520 [11:41<20:29,  3.67s/it]
                                                 
{'loss': 1.3177, 'grad_norm': 0.0009704462831389995, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [11:41<20:29,  3.67s/it]
 36%|███▌      | 186/520 [11:44<20:26,  3.67s/it]
                                                 
{'loss': 1.1977, 'grad_norm': 0.001010574582719884, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [11:44<20:26,  3.67s/it]
 36%|███▌      | 187/520 [11:48<20:20,  3.66s/it]
                                                 
{'loss': 1.2005, 'grad_norm': 0.0010834154403055705, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [11:48<20:20,  3.66s/it]
 36%|███▌      | 188/520 [11:52<20:14,  3.66s/it]
                                                 
{'loss': 1.2924, 'grad_norm': 0.001034780875616156, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [11:52<20:14,  3.66s/it]
 36%|███▋      | 189/520 [11:55<20:12,  3.66s/it]
                                                 
{'loss': 1.2889, 'grad_norm': 0.0009114133758286846, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [11:55<20:12,  3.66s/it]
 37%|███▋      | 190/520 [11:59<20:10,  3.67s/it]
                                                 
{'loss': 1.2091, 'grad_norm': 0.0010335068892297248, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [11:59<20:10,  3.67s/it]
 37%|███▋      | 191/520 [12:03<20:09,  3.68s/it]
                                                 
{'loss': 1.1758, 'grad_norm': 0.000948974533083479, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [12:03<20:09,  3.68s/it]
 37%|███▋      | 192/520 [12:06<20:09,  3.69s/it]
                                                 
{'loss': 1.2526, 'grad_norm': 0.0009222737708943759, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:06<20:09,  3.69s/it]
 37%|███▋      | 193/520 [12:10<20:01,  3.67s/it]
                                                 
{'loss': 1.2234, 'grad_norm': 0.0011235812453095828, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:10<20:01,  3.67s/it]
 37%|███▋      | 194/520 [12:14<19:58,  3.68s/it]
                                                 
{'loss': 1.1219, 'grad_norm': 0.0008640169390099564, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:14<19:58,  3.68s/it]
 38%|███▊      | 195/520 [12:17<19:53,  3.67s/it]
                                                 
{'loss': 1.2735, 'grad_norm': 0.0009584602443280612, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:17<19:53,  3.67s/it]
 38%|███▊      | 196/520 [12:21<19:47,  3.67s/it]
                                                 
{'loss': 1.2422, 'grad_norm': 0.0010490115907846591, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:21<19:47,  3.67s/it]
 38%|███▊      | 197/520 [12:25<19:54,  3.70s/it]
                                                 
{'loss': 1.1934, 'grad_norm': 0.0009758815440015419, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:25<19:54,  3.70s/it]
 38%|███▊      | 198/520 [12:29<20:10,  3.76s/it]
                                                 
{'loss': 1.2629, 'grad_norm': 0.0010035185151051479, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:29<20:10,  3.76s/it]
 38%|███▊      | 199/520 [12:33<20:21,  3.81s/it]
                                                 
{'loss': 1.1792, 'grad_norm': 0.0009373486345579799, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [12:33<20:21,  3.81s/it]
 38%|███▊      | 200/520 [12:37<20:25,  3.83s/it]
                                                 
{'loss': 1.1775, 'grad_norm': 0.0009138224249182495, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [12:37<20:25,  3.83s/it]
 39%|███▊      | 201/520 [12:40<20:27,  3.85s/it]
                                                 
{'loss': 1.1981, 'grad_norm': 0.0008268642270725403, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [12:40<20:27,  3.85s/it]
 39%|███▉      | 202/520 [12:44<20:15,  3.82s/it]
                                                 
{'loss': 1.1925, 'grad_norm': 0.0009473062837039761, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [12:44<20:15,  3.82s/it]
 39%|███▉      | 203/520 [12:48<19:54,  3.77s/it]
                                                 
{'loss': 1.2421, 'grad_norm': 0.0009955725842104512, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [12:48<19:54,  3.77s/it]
 39%|███▉      | 204/520 [12:52<19:43,  3.75s/it]
                                                 
{'loss': 1.2508, 'grad_norm': 0.0009635245471003158, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [12:52<19:43,  3.75s/it]
 39%|███▉      | 205/520 [12:55<19:56,  3.80s/it]
                                                 
{'loss': 1.1995, 'grad_norm': 0.0009063719220030785, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [12:55<19:56,  3.80s/it]
 40%|███▉      | 206/520 [12:59<20:05,  3.84s/it]
                                                 
{'loss': 1.2899, 'grad_norm': 0.0009062039329816679, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [12:59<20:05,  3.84s/it]
 40%|███▉      | 207/520 [13:03<20:07,  3.86s/it]
                                                 
{'loss': 1.1749, 'grad_norm': 0.0008338250002889656, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [13:03<20:07,  3.86s/it]
 40%|████      | 208/520 [13:07<20:05,  3.87s/it]
                                                 
{'loss': 1.2747, 'grad_norm': 0.0010206410154207945, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [13:07<20:05,  3.87s/it]
 40%|████      | 209/520 [13:11<19:57,  3.85s/it]
                                                 
{'loss': 1.189, 'grad_norm': 0.0008739375593526655, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:11<19:57,  3.85s/it]
 40%|████      | 210/520 [13:15<19:35,  3.79s/it]
                                                 
{'loss': 1.2564, 'grad_norm': 0.0009421441486610964, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:15<19:35,  3.79s/it]
 41%|████      | 211/520 [13:18<19:19,  3.75s/it]
                                                 
{'loss': 1.2646, 'grad_norm': 0.0008681224086607107, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:18<19:19,  3.75s/it]
 41%|████      | 212/520 [13:22<19:07,  3.72s/it]
                                                 
{'loss': 1.2596, 'grad_norm': 0.0009160933016312563, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:22<19:07,  3.72s/it]
 41%|████      | 213/520 [13:26<18:56,  3.70s/it]
                                                 
{'loss': 1.2111, 'grad_norm': 0.0010668334590054245, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:26<18:56,  3.70s/it]
 41%|████      | 214/520 [13:29<18:48,  3.69s/it]
                                                 
{'loss': 1.201, 'grad_norm': 0.0009663081410163465, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:29<18:48,  3.69s/it]
 41%|████▏     | 215/520 [13:33<18:44,  3.69s/it]
                                                 
{'loss': 1.1347, 'grad_norm': 0.0008928221882336813, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [13:33<18:44,  3.69s/it]
 42%|████▏     | 216/520 [13:37<18:38,  3.68s/it]
                                                 
{'loss': 1.1196, 'grad_norm': 0.0008986181370422127, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [13:37<18:38,  3.68s/it]
 42%|████▏     | 217/520 [13:40<18:33,  3.67s/it]
                                                 
{'loss': 1.2558, 'grad_norm': 0.0009611517223054261, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [13:40<18:33,  3.67s/it]
 42%|████▏     | 218/520 [13:44<18:25,  3.66s/it]
                                                 
{'loss': 1.2289, 'grad_norm': 0.0010242004786472727, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [13:44<18:25,  3.66s/it]
 42%|████▏     | 219/520 [13:48<18:19,  3.65s/it]
                                                 
{'loss': 1.2376, 'grad_norm': 0.0008313000350729799, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [13:48<18:19,  3.65s/it]
 42%|████▏     | 220/520 [13:51<18:14,  3.65s/it]
                                                 
{'loss': 1.1861, 'grad_norm': 0.0008857109319049207, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [13:51<18:14,  3.65s/it]
 42%|████▎     | 221/520 [13:55<18:12,  3.65s/it]
                                                 
{'loss': 1.2314, 'grad_norm': 0.0008966009630781475, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [13:55<18:12,  3.65s/it]
 43%|████▎     | 222/520 [13:58<18:09,  3.66s/it]
                                                 
{'loss': 1.1791, 'grad_norm': 0.0008750685842275972, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [13:58<18:09,  3.66s/it]
 43%|████▎     | 223/520 [14:02<18:10,  3.67s/it]
                                                 
{'loss': 1.1682, 'grad_norm': 0.0008673283610612547, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [14:02<18:10,  3.67s/it]
 43%|████▎     | 224/520 [14:06<18:09,  3.68s/it]
                                                 
{'loss': 1.2595, 'grad_norm': 0.0008669722183534568, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [14:06<18:09,  3.68s/it]
 43%|████▎     | 225/520 [14:10<18:07,  3.69s/it]
                                                 
{'loss': 1.1767, 'grad_norm': 0.0008911741148070225, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:10<18:07,  3.69s/it]
 43%|████▎     | 226/520 [14:13<18:06,  3.69s/it]
                                                 
{'loss': 1.2708, 'grad_norm': 0.000876217152855064, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:13<18:06,  3.69s/it]
 44%|████▎     | 227/520 [14:17<17:59,  3.68s/it]
                                                 
{'loss': 1.2644, 'grad_norm': 0.0008613526399661662, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:17<17:59,  3.68s/it]
 44%|████▍     | 228/520 [14:21<17:54,  3.68s/it]
                                                 
{'loss': 1.2842, 'grad_norm': 0.001069929278165338, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:21<17:54,  3.68s/it]
 44%|████▍     | 229/520 [14:24<17:49,  3.67s/it]
                                                 
{'loss': 1.2435, 'grad_norm': 0.0008325316530389129, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:24<17:49,  3.67s/it]
 44%|████▍     | 230/520 [14:28<17:47,  3.68s/it]
                                                 
{'loss': 1.1284, 'grad_norm': 0.0008713261827134914, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:28<17:47,  3.68s/it]
 44%|████▍     | 231/520 [14:32<17:43,  3.68s/it]
                                                 
{'loss': 1.1932, 'grad_norm': 0.0008256241023155085, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:32<17:43,  3.68s/it]
 45%|████▍     | 232/520 [14:35<17:38,  3.68s/it]
                                                 
{'loss': 1.3257, 'grad_norm': 0.0009439196211518384, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [14:35<17:38,  3.68s/it]
 45%|████▍     | 233/520 [14:39<17:38,  3.69s/it]
                                                 
{'loss': 1.2169, 'grad_norm': 0.000973081087660518, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [14:39<17:38,  3.69s/it]
 45%|████▌     | 234/520 [14:43<17:48,  3.74s/it]
                                                 
{'loss': 1.1492, 'grad_norm': 0.0009174713279035883, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [14:43<17:48,  3.74s/it]
 45%|████▌     | 235/520 [14:47<17:51,  3.76s/it]
                                                 
{'loss': 1.1914, 'grad_norm': 0.0008732928228223138, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [14:47<17:51,  3.76s/it]
 45%|████▌     | 236/520 [14:51<17:54,  3.78s/it]
                                                 
{'loss': 1.2748, 'grad_norm': 0.0008328935445802758, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [14:51<17:54,  3.78s/it]
 46%|████▌     | 237/520 [14:54<17:53,  3.79s/it]
                                                 
{'loss': 1.2642, 'grad_norm': 0.0008572635514940767, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [14:54<17:53,  3.79s/it]
 46%|████▌     | 238/520 [14:58<17:51,  3.80s/it]
                                                 
{'loss': 1.2061, 'grad_norm': 0.0009052045419166341, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [14:58<17:51,  3.80s/it]
 46%|████▌     | 239/520 [15:02<17:50,  3.81s/it]
                                                 
{'loss': 1.2826, 'grad_norm': 0.0009455468570176625, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [15:02<17:50,  3.81s/it]
 46%|████▌     | 240/520 [15:06<17:51,  3.83s/it]
                                                 
{'loss': 1.0903, 'grad_norm': 0.0007906781525730577, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:06<17:51,  3.83s/it]
 46%|████▋     | 241/520 [15:10<17:47,  3.83s/it]
                                                 
{'loss': 1.1664, 'grad_norm': 0.000858704511630343, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:10<17:47,  3.83s/it]
 47%|████▋     | 242/520 [15:14<17:42,  3.82s/it]
                                                 
{'loss': 1.1824, 'grad_norm': 0.000828906211319979, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:14<17:42,  3.82s/it]
 47%|████▋     | 243/520 [15:17<17:41,  3.83s/it]
                                                 
{'loss': 1.1856, 'grad_norm': 0.000880376010741125, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:17<17:41,  3.83s/it]
 47%|████▋     | 244/520 [15:21<17:39,  3.84s/it]
                                                 
{'loss': 1.2914, 'grad_norm': 0.0009050515925426431, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:21<17:39,  3.84s/it]
 47%|████▋     | 245/520 [15:25<17:34,  3.83s/it]
                                                 
{'loss': 1.1638, 'grad_norm': 0.0009030703844879949, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:25<17:34,  3.83s/it]
 47%|████▋     | 246/520 [15:29<17:30,  3.83s/it]
                                                 
{'loss': 1.3124, 'grad_norm': 0.0009306383641897676, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:29<17:30,  3.83s/it]
 48%|████▊     | 247/520 [15:33<17:26,  3.83s/it]
                                                 
{'loss': 1.3354, 'grad_norm': 0.0009297694527927876, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:33<17:26,  3.83s/it]
 48%|████▊     | 248/520 [15:37<17:20,  3.82s/it]
                                                 
{'loss': 1.1602, 'grad_norm': 0.0008922797267441315, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [15:37<17:20,  3.82s/it]
 48%|████▊     | 249/520 [15:40<17:16,  3.83s/it]
                                                 
{'loss': 1.2558, 'grad_norm': 0.000911105132148296, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [15:40<17:16,  3.83s/it]
 48%|████▊     | 250/520 [15:44<17:04,  3.80s/it]
                                                 
{'loss': 1.1999, 'grad_norm': 0.00097316849869418, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [15:44<17:04,  3.80s/it]
 48%|████▊     | 251/520 [15:48<16:50,  3.76s/it]
                                                 
{'loss': 1.2665, 'grad_norm': 0.0008343153386617157, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [15:48<16:50,  3.76s/it]
 48%|████▊     | 252/520 [15:51<16:40,  3.74s/it]
                                                 
{'loss': 1.2095, 'grad_norm': 0.0008380178898220818, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [15:51<16:40,  3.74s/it]
 49%|████▊     | 253/520 [15:55<16:34,  3.72s/it]
                                                 
{'loss': 1.2499, 'grad_norm': 0.0009566091243813734, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [15:55<16:34,  3.72s/it]
 49%|████▉     | 254/520 [15:59<16:25,  3.70s/it]
                                                 
{'loss': 1.1884, 'grad_norm': 0.0008730992569721923, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [15:59<16:25,  3.70s/it]
 49%|████▉     | 255/520 [16:02<16:19,  3.70s/it]
                                                 
{'loss': 1.1984, 'grad_norm': 0.0009444207028747889, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [16:02<16:19,  3.70s/it]
 49%|████▉     | 256/520 [16:06<16:11,  3.68s/it]
                                                 
{'loss': 1.2519, 'grad_norm': 0.0009357778539762104, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:06<16:11,  3.68s/it]
 49%|████▉     | 257/520 [16:10<16:05,  3.67s/it]
                                                 
{'loss': 1.2286, 'grad_norm': 0.000907141970223158, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:10<16:05,  3.67s/it]
 50%|████▉     | 258/520 [16:13<16:02,  3.67s/it]
                                                 
{'loss': 1.2391, 'grad_norm': 0.0007820821244126335, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:13<16:02,  3.67s/it]
 50%|████▉     | 259/520 [16:17<15:56,  3.66s/it]
                                                 
{'loss': 1.3068, 'grad_norm': 0.000951931163766605, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:17<15:56,  3.66s/it]
 50%|█████     | 260/520 [16:21<15:51,  3.66s/it]
                                                 
{'loss': 1.2596, 'grad_norm': 0.0007548326030345746, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:21<15:51,  3.66s/it]
 50%|█████     | 261/520 [16:24<15:48,  3.66s/it]
                                                 
{'loss': 1.2075, 'grad_norm': 0.00085228442642842, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:24<15:48,  3.66s/it]
 50%|█████     | 262/520 [16:28<15:44,  3.66s/it]
                                                 
{'loss': 1.1708, 'grad_norm': 0.0008561714138569583, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:28<15:44,  3.66s/it]
 51%|█████     | 263/520 [16:32<15:39,  3.66s/it]
                                                 
{'loss': 1.2206, 'grad_norm': 0.0008447506135465498, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:32<15:39,  3.66s/it]
 51%|█████     | 264/520 [16:35<15:39,  3.67s/it]
                                                 
{'loss': 1.2613, 'grad_norm': 0.0008256883678336382, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [16:35<15:39,  3.67s/it]
 51%|█████     | 265/520 [16:39<15:37,  3.68s/it]
                                                 
{'loss': 1.1789, 'grad_norm': 0.0009810014879696079, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [16:39<15:37,  3.68s/it]
 51%|█████     | 266/520 [16:43<15:32,  3.67s/it]
                                                 
{'loss': 1.0573, 'grad_norm': 0.0007421584902777116, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [16:43<15:32,  3.67s/it]
 51%|█████▏    | 267/520 [16:46<15:30,  3.68s/it]
                                                 
{'loss': 1.1818, 'grad_norm': 0.0008292354509556741, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [16:46<15:30,  3.68s/it]
 52%|█████▏    | 268/520 [16:50<15:27,  3.68s/it]
                                                 
{'loss': 1.331, 'grad_norm': 0.0008912920748999813, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [16:50<15:27,  3.68s/it]
 52%|█████▏    | 269/520 [16:54<15:24,  3.68s/it]
                                                 
{'loss': 1.282, 'grad_norm': 0.0009007522312323337, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [16:54<15:24,  3.68s/it]
 52%|█████▏    | 270/520 [16:58<15:23,  3.69s/it]
                                                 
{'loss': 1.1591, 'grad_norm': 0.000846947397072268, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [16:58<15:23,  3.69s/it]
 52%|█████▏    | 271/520 [17:01<15:19,  3.69s/it]
                                                 
{'loss': 1.2614, 'grad_norm': 0.0009087340023147141, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [17:01<15:19,  3.69s/it]
 52%|█████▏    | 272/520 [17:05<15:18,  3.70s/it]
                                                 
{'loss': 1.1742, 'grad_norm': 0.0008347468837900717, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:05<15:18,  3.70s/it]
 52%|█████▎    | 273/520 [17:09<15:14,  3.70s/it]
                                                 
{'loss': 1.2827, 'grad_norm': 0.0008470221037078762, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:09<15:14,  3.70s/it]
 53%|█████▎    | 274/520 [17:12<15:14,  3.72s/it]
                                                 
{'loss': 1.2463, 'grad_norm': 0.0009246906332058114, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:12<15:14,  3.72s/it]
 53%|█████▎    | 275/520 [17:16<15:17,  3.74s/it]
                                                 
{'loss': 1.1882, 'grad_norm': 0.0008465250541752112, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:16<15:17,  3.74s/it]
 53%|█████▎    | 276/520 [17:20<15:20,  3.77s/it]
                                                 
{'loss': 1.2469, 'grad_norm': 0.00089497775169389, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:20<15:20,  3.77s/it]
 53%|█████▎    | 277/520 [17:24<15:22,  3.80s/it]
                                                 
{'loss': 1.2996, 'grad_norm': 0.0008080328768867475, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:24<15:22,  3.80s/it]
 53%|█████▎    | 278/520 [17:28<15:23,  3.81s/it]
                                                 
{'loss': 1.146, 'grad_norm': 0.0007916804257063824, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:28<15:23,  3.81s/it]
 54%|█████▎    | 279/520 [17:32<15:19,  3.81s/it]
                                                 
{'loss': 1.1772, 'grad_norm': 0.0009049343984390481, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:32<15:19,  3.81s/it]
 54%|█████▍    | 280/520 [17:35<15:18,  3.83s/it]
                                                 
{'loss': 1.176, 'grad_norm': 0.0009543386681254707, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [17:35<15:18,  3.83s/it]
 54%|█████▍    | 281/520 [17:39<15:11,  3.81s/it]
                                                 
{'loss': 1.2833, 'grad_norm': 0.0008975278779251666, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [17:39<15:11,  3.81s/it]
 54%|█████▍    | 282/520 [17:43<15:07,  3.81s/it]
                                                 
{'loss': 1.1622, 'grad_norm': 0.0008416377941722808, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [17:43<15:07,  3.81s/it]
 54%|█████▍    | 283/520 [17:47<14:54,  3.77s/it]
                                                 
{'loss': 1.3022, 'grad_norm': 0.0009668866159612603, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [17:47<14:54,  3.77s/it]
 55%|█████▍    | 284/520 [17:50<14:46,  3.75s/it]
                                                 
{'loss': 1.1738, 'grad_norm': 0.0009019066844787086, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [17:50<14:46,  3.75s/it]
 55%|█████▍    | 285/520 [17:54<14:39,  3.74s/it]
                                                 
{'loss': 1.1802, 'grad_norm': 0.0008676856289551404, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [17:54<14:39,  3.74s/it]
 55%|█████▌    | 286/520 [17:58<14:33,  3.73s/it]
                                                 
{'loss': 1.0585, 'grad_norm': 0.0008087570095232465, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [17:58<14:33,  3.73s/it]
 55%|█████▌    | 287/520 [18:02<14:26,  3.72s/it]
                                                 
{'loss': 1.2904, 'grad_norm': 0.0008360604149073081, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [18:02<14:26,  3.72s/it]
 55%|█████▌    | 288/520 [18:05<14:27,  3.74s/it]
                                                 
{'loss': 1.3211, 'grad_norm': 0.0008196349957762546, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:05<14:27,  3.74s/it]
 56%|█████▌    | 289/520 [18:09<14:35,  3.79s/it]
                                                 
{'loss': 1.196, 'grad_norm': 0.0008207341253726875, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:09<14:35,  3.79s/it]
 56%|█████▌    | 290/520 [18:13<14:38,  3.82s/it]
                                                 
{'loss': 1.1219, 'grad_norm': 0.0007655215990959629, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:13<14:38,  3.82s/it]
 56%|█████▌    | 291/520 [18:17<14:37,  3.83s/it]
                                                 
{'loss': 1.1615, 'grad_norm': 0.0008310387588454645, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:17<14:37,  3.83s/it]
 56%|█████▌    | 292/520 [18:21<14:36,  3.84s/it]
                                                 
{'loss': 1.2166, 'grad_norm': 0.0008194319060960096, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:21<14:36,  3.84s/it]
 56%|█████▋    | 293/520 [18:25<14:32,  3.84s/it]
                                                 
{'loss': 1.1653, 'grad_norm': 0.000832029071380735, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:25<14:32,  3.84s/it]
 57%|█████▋    | 294/520 [18:29<14:29,  3.85s/it]
                                                 
{'loss': 1.189, 'grad_norm': 0.00089105586850477, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:29<14:29,  3.85s/it]
 57%|█████▋    | 295/520 [18:32<14:26,  3.85s/it]
                                                 
{'loss': 1.2206, 'grad_norm': 0.0008156068858540923, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:32<14:26,  3.85s/it]
 57%|█████▋    | 296/520 [18:36<14:23,  3.86s/it]
                                                 
{'loss': 1.1416, 'grad_norm': 0.00085117091065196, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [18:36<14:23,  3.86s/it]
 57%|█████▋    | 297/520 [18:40<14:07,  3.80s/it]
                                                 
{'loss': 1.2697, 'grad_norm': 0.000922172698917861, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [18:40<14:07,  3.80s/it]
 57%|█████▋    | 298/520 [18:44<13:57,  3.77s/it]
                                                 
{'loss': 1.2359, 'grad_norm': 0.000777566825774334, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [18:44<13:57,  3.77s/it]
 57%|█████▊    | 299/520 [18:47<13:48,  3.75s/it]
                                                 
{'loss': 1.2581, 'grad_norm': 0.000799570462621054, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [18:47<13:48,  3.75s/it]
 58%|█████▊    | 300/520 [18:51<13:39,  3.73s/it]
                                                 
{'loss': 1.2862, 'grad_norm': 0.0008375366717905503, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [18:51<13:39,  3.73s/it]
 58%|█████▊    | 301/520 [18:55<13:32,  3.71s/it]
                                                 
{'loss': 1.2692, 'grad_norm': 0.0008353937230014906, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [18:55<13:32,  3.71s/it]
 58%|█████▊    | 302/520 [18:58<13:26,  3.70s/it]
                                                 
{'loss': 1.2642, 'grad_norm': 0.000804158652931079, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [18:58<13:26,  3.70s/it]
 58%|█████▊    | 303/520 [19:02<13:24,  3.71s/it]
                                                 
{'loss': 1.1913, 'grad_norm': 0.000913033173152988, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [19:02<13:24,  3.71s/it]
 58%|█████▊    | 304/520 [19:06<13:22,  3.71s/it]
                                                 
{'loss': 1.1743, 'grad_norm': 0.0008751465008768523, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:06<13:22,  3.71s/it]
 59%|█████▊    | 305/520 [19:10<13:16,  3.71s/it]
                                                 
{'loss': 1.2975, 'grad_norm': 0.000978941594238876, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:10<13:16,  3.71s/it]
 59%|█████▉    | 306/520 [19:13<13:14,  3.71s/it]
                                                 
{'loss': 1.2383, 'grad_norm': 0.0008526687431179773, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:13<13:14,  3.71s/it]
 59%|█████▉    | 307/520 [19:17<13:40,  3.85s/it]
                                                 
{'loss': 1.1775, 'grad_norm': 0.0007909917192085277, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:17<13:40,  3.85s/it]
 59%|█████▉    | 308/520 [19:21<13:27,  3.81s/it]
                                                 
{'loss': 1.2994, 'grad_norm': 0.0008212564027090641, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:21<13:27,  3.81s/it]
 59%|█████▉    | 309/520 [19:25<13:15,  3.77s/it]
                                                 
{'loss': 1.1832, 'grad_norm': 0.0008011326425731641, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:25<13:15,  3.77s/it]
 60%|█████▉    | 310/520 [19:29<13:06,  3.75s/it]
                                                 
{'loss': 1.1593, 'grad_norm': 0.0008426772414711472, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:29<13:06,  3.75s/it]
 60%|█████▉    | 311/520 [19:32<12:59,  3.73s/it]
                                                 
{'loss': 1.143, 'grad_norm': 0.0008086669362243245, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [19:32<12:59,  3.73s/it]
 60%|██████    | 312/520 [19:36<12:51,  3.71s/it]
                                                 
{'loss': 1.129, 'grad_norm': 0.0008023511676053161, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [19:36<12:51,  3.71s/it]
 60%|██████    | 313/520 [19:40<12:47,  3.71s/it]
                                                 
{'loss': 1.1133, 'grad_norm': 0.0007397596061754019, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [19:40<12:47,  3.71s/it]
 60%|██████    | 314/520 [19:44<13:12,  3.84s/it]
                                                 
{'loss': 1.1511, 'grad_norm': 0.0008084270330700137, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [19:44<13:12,  3.84s/it]
 61%|██████    | 315/520 [19:47<13:00,  3.81s/it]
                                                 
{'loss': 1.2263, 'grad_norm': 0.0008910586575829308, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [19:47<13:00,  3.81s/it]
 61%|██████    | 316/520 [19:52<13:16,  3.90s/it]
                                                 
{'loss': 1.1333, 'grad_norm': 0.0008556748336229067, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [19:52<13:16,  3.90s/it]
 61%|██████    | 317/520 [19:55<12:59,  3.84s/it]
                                                 
{'loss': 1.145, 'grad_norm': 0.0007455845254348017, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [19:55<12:59,  3.84s/it]
 61%|██████    | 318/520 [19:59<12:47,  3.80s/it]
                                                 
{'loss': 1.2572, 'grad_norm': 0.0008973668734563131, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [19:59<12:47,  3.80s/it]
 61%|██████▏   | 319/520 [20:03<12:59,  3.88s/it]
                                                 
{'loss': 1.14, 'grad_norm': 0.0007751141110950702, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [20:03<12:59,  3.88s/it]
 62%|██████▏   | 320/520 [20:07<12:45,  3.83s/it]
                                                 
{'loss': 1.0892, 'grad_norm': 0.0008157219035312987, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:07<12:45,  3.83s/it]
 62%|██████▏   | 321/520 [20:10<12:34,  3.79s/it]
                                                 
{'loss': 1.2765, 'grad_norm': 0.0008059210408327869, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:10<12:34,  3.79s/it]
 62%|██████▏   | 322/520 [20:14<12:25,  3.76s/it]
                                                 
{'loss': 1.1238, 'grad_norm': 0.000837338285065987, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:14<12:25,  3.76s/it]
 62%|██████▏   | 323/520 [20:18<12:16,  3.74s/it]
                                                 
{'loss': 1.1867, 'grad_norm': 0.0008054003043263963, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:18<12:16,  3.74s/it]
 62%|██████▏   | 324/520 [20:21<12:07,  3.71s/it]
                                                 
{'loss': 1.2243, 'grad_norm': 0.000842809236457948, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:21<12:07,  3.71s/it]
 62%|██████▎   | 325/520 [20:25<12:00,  3.70s/it]
                                                 
{'loss': 1.2197, 'grad_norm': 0.000882447676183897, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:25<12:00,  3.70s/it]
 63%|██████▎   | 326/520 [20:29<11:56,  3.69s/it]
                                                 
{'loss': 1.2078, 'grad_norm': 0.0008535780301336415, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:29<11:56,  3.69s/it]
 63%|██████▎   | 327/520 [20:32<11:49,  3.68s/it]
                                                 
{'loss': 1.2387, 'grad_norm': 0.0008746544493908477, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [20:32<11:49,  3.68s/it]
 63%|██████▎   | 328/520 [20:36<11:43,  3.66s/it]
                                                 
{'loss': 1.2655, 'grad_norm': 0.0008719334166765752, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [20:36<11:43,  3.66s/it]
 63%|██████▎   | 329/520 [20:40<11:42,  3.68s/it]
                                                 
{'loss': 1.1334, 'grad_norm': 0.0007168259056692257, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [20:40<11:42,  3.68s/it]
 63%|██████▎   | 330/520 [20:43<11:38,  3.68s/it]
                                                 
{'loss': 1.2158, 'grad_norm': 0.0007938578602275123, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [20:43<11:38,  3.68s/it]
 64%|██████▎   | 331/520 [20:47<11:35,  3.68s/it]
                                                 
{'loss': 1.1715, 'grad_norm': 0.0008498305274995624, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [20:47<11:35,  3.68s/it]
 64%|██████▍   | 332/520 [20:51<11:34,  3.69s/it]
                                                 
{'loss': 1.2593, 'grad_norm': 0.0007973651362104641, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [20:51<11:34,  3.69s/it]
 64%|██████▍   | 333/520 [20:55<11:31,  3.70s/it]
                                                 
{'loss': 1.3098, 'grad_norm': 0.0008830125898471809, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [20:55<11:31,  3.70s/it]
 64%|██████▍   | 334/520 [20:58<11:28,  3.70s/it]
                                                 
{'loss': 1.2258, 'grad_norm': 0.0008995086733510348, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [20:58<11:28,  3.70s/it]
 64%|██████▍   | 335/520 [21:02<11:23,  3.70s/it]
                                                 
{'loss': 1.2181, 'grad_norm': 0.0007943070608016983, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [21:02<11:23,  3.70s/it]
 65%|██████▍   | 336/520 [21:06<11:19,  3.69s/it]
                                                 
{'loss': 1.1167, 'grad_norm': 0.0008735123212150434, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:06<11:19,  3.69s/it]
 65%|██████▍   | 337/520 [21:09<11:16,  3.70s/it]
                                                 
{'loss': 1.1145, 'grad_norm': 0.0008016106345241088, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:09<11:16,  3.70s/it]
 65%|██████▌   | 338/520 [21:13<11:15,  3.71s/it]
                                                 
{'loss': 1.2303, 'grad_norm': 0.0008717819551692182, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:13<11:15,  3.71s/it]
 65%|██████▌   | 339/520 [21:17<11:10,  3.70s/it]
                                                 
{'loss': 1.177, 'grad_norm': 0.0008377898350668883, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:17<11:10,  3.70s/it]
 65%|██████▌   | 340/520 [21:21<11:06,  3.70s/it]
                                                 
{'loss': 1.157, 'grad_norm': 0.0007960569490354259, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:21<11:06,  3.70s/it]
 66%|██████▌   | 341/520 [21:24<11:03,  3.70s/it]
                                                 
{'loss': 1.1874, 'grad_norm': 0.0008853945815167423, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:24<11:03,  3.70s/it]
 66%|██████▌   | 342/520 [21:28<10:57,  3.69s/it]
                                                 
{'loss': 1.2305, 'grad_norm': 0.0009785729734216378, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:28<10:57,  3.69s/it]
 66%|██████▌   | 343/520 [21:32<10:54,  3.70s/it]
                                                 
{'loss': 1.1851, 'grad_norm': 0.0007086141033941644, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [21:32<10:54,  3.70s/it]
 66%|██████▌   | 344/520 [21:35<10:49,  3.69s/it]
                                                 
{'loss': 1.1404, 'grad_norm': 0.0007387164984841032, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [21:35<10:49,  3.69s/it]
 66%|██████▋   | 345/520 [21:39<10:45,  3.69s/it]
                                                 
{'loss': 1.2455, 'grad_norm': 0.0008349844210000557, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [21:39<10:45,  3.69s/it]
 67%|██████▋   | 346/520 [21:43<10:40,  3.68s/it]
                                                 
{'loss': 1.1976, 'grad_norm': 0.0008181258736939009, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [21:43<10:40,  3.68s/it]
 67%|██████▋   | 347/520 [21:46<10:36,  3.68s/it]
                                                 
{'loss': 1.1624, 'grad_norm': 0.000751293508379836, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [21:46<10:36,  3.68s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [21:50<10:33,  3.68s/it]
                                                 
{'loss': 1.1195, 'grad_norm': 0.0009673287360138676, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [21:50<10:33,  3.68s/it]
 67%|██████▋   | 349/520 [21:54<10:30,  3.69s/it]
                                                 
{'loss': 1.1544, 'grad_norm': 0.0008217254163744408, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [21:54<10:30,  3.69s/it]
 67%|██████▋   | 350/520 [21:57<10:26,  3.69s/it]
                                                 
{'loss': 1.198, 'grad_norm': 0.0008248291034825256, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [21:57<10:26,  3.69s/it]
 68%|██████▊   | 351/520 [22:01<10:24,  3.69s/it]
                                                 
{'loss': 1.1037, 'grad_norm': 0.0007635817940211965, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [22:01<10:24,  3.69s/it]
 68%|██████▊   | 352/520 [22:05<10:19,  3.69s/it]
                                                 
{'loss': 1.2246, 'grad_norm': 0.0007839317681087829, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [22:05<10:19,  3.69s/it]
 68%|██████▊   | 353/520 [22:09<10:18,  3.70s/it]
                                                 
{'loss': 1.164, 'grad_norm': 0.0007001251765233179, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:09<10:18,  3.70s/it]
 68%|██████▊   | 354/520 [22:12<10:14,  3.70s/it]
                                                 
{'loss': 1.2696, 'grad_norm': 0.0007815836971490103, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:12<10:14,  3.70s/it]
 68%|██████▊   | 355/520 [22:16<10:07,  3.68s/it]
                                                 
{'loss': 1.1679, 'grad_norm': 0.0008082400153136906, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:16<10:07,  3.68s/it]
 68%|██████▊   | 356/520 [22:20<10:04,  3.69s/it]
                                                 
{'loss': 1.1705, 'grad_norm': 0.0008228787700959886, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:20<10:04,  3.69s/it]
 69%|██████▊   | 357/520 [22:23<09:58,  3.67s/it]
                                                 
{'loss': 1.2013, 'grad_norm': 0.0007925838674693638, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:23<09:58,  3.67s/it]
 69%|██████▉   | 358/520 [22:27<09:55,  3.68s/it]
                                                 
{'loss': 1.1315, 'grad_norm': 0.0008142935986232956, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:27<09:55,  3.68s/it]
 69%|██████▉   | 359/520 [22:31<09:53,  3.68s/it]
                                                 
{'loss': 1.2061, 'grad_norm': 0.000829460546446095, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [22:31<09:53,  3.68s/it]
 69%|██████▉   | 360/520 [22:34<09:50,  3.69s/it]
                                                 
{'loss': 1.2273, 'grad_norm': 0.0008682865891467183, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [22:34<09:50,  3.69s/it]
 69%|██████▉   | 361/520 [22:38<09:49,  3.71s/it]
                                                 
{'loss': 1.2256, 'grad_norm': 0.0007382092568638627, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [22:38<09:49,  3.71s/it]
 70%|██████▉   | 362/520 [22:42<09:50,  3.74s/it]
                                                 
{'loss': 1.1834, 'grad_norm': 0.0008792823752577819, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [22:42<09:50,  3.74s/it]
 70%|██████▉   | 363/520 [22:46<09:50,  3.76s/it]
                                                 
{'loss': 1.2244, 'grad_norm': 0.0008350880218370427, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [22:46<09:50,  3.76s/it]
 70%|███████   | 364/520 [22:49<09:49,  3.78s/it]
                                                 
{'loss': 1.2545, 'grad_norm': 0.0008301556299740614, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [22:49<09:49,  3.78s/it]
 70%|███████   | 365/520 [22:53<09:47,  3.79s/it]
                                                 
{'loss': 1.2662, 'grad_norm': 0.0008185155040941783, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [22:53<09:47,  3.79s/it]
 70%|███████   | 366/520 [22:57<09:45,  3.80s/it]
                                                 
{'loss': 1.2328, 'grad_norm': 0.0007936157616616551, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [22:57<09:45,  3.80s/it]
 71%|███████   | 367/520 [23:01<09:42,  3.81s/it]
                                                 
{'loss': 1.2258, 'grad_norm': 0.0008388414474191501, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [23:01<09:42,  3.81s/it]
 71%|███████   | 368/520 [23:05<09:40,  3.82s/it]
                                                 
{'loss': 1.0855, 'grad_norm': 0.0008327672145839597, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [23:05<09:40,  3.82s/it]
 71%|███████   | 369/520 [23:09<09:37,  3.82s/it]
                                                 
{'loss': 1.2016, 'grad_norm': 0.0007434439398962111, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:09<09:37,  3.82s/it]
 71%|███████   | 370/520 [23:12<09:33,  3.83s/it]
                                                 
{'loss': 1.1422, 'grad_norm': 0.0007784178747905046, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:12<09:33,  3.83s/it]
 71%|███████▏  | 371/520 [23:16<09:30,  3.83s/it]
                                                 
{'loss': 1.1403, 'grad_norm': 0.0008736634332951631, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:16<09:30,  3.83s/it]
 72%|███████▏  | 372/520 [23:20<09:26,  3.83s/it]
                                                 
{'loss': 1.2823, 'grad_norm': 0.0007480325897124572, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:20<09:26,  3.83s/it]
 72%|███████▏  | 373/520 [23:24<09:24,  3.84s/it]
                                                 
{'loss': 1.1689, 'grad_norm': 0.0008739316856266908, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:24<09:24,  3.84s/it]
 72%|███████▏  | 374/520 [23:28<09:20,  3.84s/it]
                                                 
{'loss': 1.2298, 'grad_norm': 0.000853951784021672, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:28<09:20,  3.84s/it]
 72%|███████▏  | 375/520 [23:32<09:17,  3.84s/it]
                                                 
{'loss': 1.1412, 'grad_norm': 0.0008076611026424229, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [23:32<09:17,  3.84s/it]
 72%|███████▏  | 376/520 [23:35<09:12,  3.84s/it]
                                                 
{'loss': 1.2484, 'grad_norm': 0.0007883210865677144, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [23:35<09:12,  3.84s/it]
 72%|███████▎  | 377/520 [23:39<09:08,  3.84s/it]
                                                 
{'loss': 1.1905, 'grad_norm': 0.0008669268565031147, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [23:39<09:08,  3.84s/it]
 73%|███████▎  | 378/520 [23:43<09:04,  3.84s/it]
                                                 
{'loss': 1.2463, 'grad_norm': 0.0007786486346641694, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [23:43<09:04,  3.84s/it]
 73%|███████▎  | 379/520 [23:47<08:59,  3.83s/it]
                                                 
{'loss': 1.2191, 'grad_norm': 0.0007978583491180279, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [23:47<08:59,  3.83s/it]
 73%|███████▎  | 380/520 [23:51<08:56,  3.83s/it]
                                                 
{'loss': 1.2562, 'grad_norm': 0.0008544852180473858, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [23:51<08:56,  3.83s/it]
 73%|███████▎  | 381/520 [23:55<08:52,  3.83s/it]
                                                 
{'loss': 1.2231, 'grad_norm': 0.0007832567025179858, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [23:55<08:52,  3.83s/it]
 73%|███████▎  | 382/520 [23:58<08:48,  3.83s/it]
                                                 
{'loss': 1.2199, 'grad_norm': 0.0007750545323217298, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [23:58<08:48,  3.83s/it]
 74%|███████▎  | 383/520 [24:02<08:43,  3.82s/it]
                                                 
{'loss': 1.0616, 'grad_norm': 0.0008820196382492979, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [24:02<08:43,  3.82s/it]
 74%|███████▍  | 384/520 [24:06<08:38,  3.81s/it]
                                                 
{'loss': 1.2689, 'grad_norm': 0.0007498773588755908, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [24:06<08:38,  3.81s/it]
 74%|███████▍  | 385/520 [24:10<08:35,  3.82s/it]
                                                 
{'loss': 1.206, 'grad_norm': 0.0007551131634607304, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:10<08:35,  3.82s/it]
 74%|███████▍  | 386/520 [24:14<08:28,  3.79s/it]
                                                 
{'loss': 1.1567, 'grad_norm': 0.0007211021614860548, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:14<08:28,  3.79s/it]
 74%|███████▍  | 387/520 [24:17<08:21,  3.77s/it]
                                                 
{'loss': 1.2842, 'grad_norm': 0.0008159014311277256, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:17<08:21,  3.77s/it]
 75%|███████▍  | 388/520 [24:21<08:13,  3.74s/it]
                                                 
{'loss': 1.1072, 'grad_norm': 0.0007747531074633244, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:21<08:13,  3.74s/it]
 75%|███████▍  | 389/520 [24:25<08:07,  3.72s/it]
                                                 
{'loss': 1.1583, 'grad_norm': 0.0009248345869112979, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}
+
 75%|███████▍  | 389/520 [24:25<08:07,  3.72s/it]
 75%|███████▌  | 390/520 [24:28<08:01,  3.71s/it]
                                                 
{'loss': 1.2319, 'grad_norm': 0.0008248607797807652, 'learning_rate': 0.031074166151605298, 'epoch': 0.75}
+
 75%|███████▌  | 390/520 [24:28<08:01,  3.71s/it]
 75%|███████▌  | 391/520 [24:32<07:58,  3.71s/it]
                                                 
{'loss': 1.2958, 'grad_norm': 0.0008336547570292818, 'learning_rate': 0.030623894397145836, 'epoch': 0.75}
+
 75%|███████▌  | 391/520 [24:32<07:58,  3.71s/it]
 75%|███████▌  | 392/520 [24:36<07:56,  3.72s/it]
                                                 
{'loss': 1.1222, 'grad_norm': 0.0008084840295178099, 'learning_rate': 0.03017631819139273, 'epoch': 0.75}
+
 75%|███████▌  | 392/520 [24:36<07:56,  3.72s/it]
 76%|███████▌  | 393/520 [24:40<07:51,  3.71s/it]
                                                 
{'loss': 1.128, 'grad_norm': 0.0007220015504010789, 'learning_rate': 0.029731454924533086, 'epoch': 0.76}
+
 76%|███████▌  | 393/520 [24:40<07:51,  3.71s/it]
 76%|███████▌  | 394/520 [24:43<07:45,  3.70s/it]
                                                 
{'loss': 1.1869, 'grad_norm': 0.000839723957325972, 'learning_rate': 0.029289321881345254, 'epoch': 0.76}
+
 76%|███████▌  | 394/520 [24:43<07:45,  3.70s/it]
 76%|███████▌  | 395/520 [24:47<07:43,  3.70s/it]
                                                 
{'loss': 1.1506, 'grad_norm': 0.0008727432670039828, 'learning_rate': 0.028849936240527008, 'epoch': 0.76}
+
 76%|███████▌  | 395/520 [24:47<07:43,  3.70s/it]
 76%|███████▌  | 396/520 [24:51<07:43,  3.74s/it]
                                                 
{'loss': 1.2318, 'grad_norm': 0.0008701459636762614, 'learning_rate': 0.028413315074028157, 'epoch': 0.76}
+
 76%|███████▌  | 396/520 [24:51<07:43,  3.74s/it]
 76%|███████▋  | 397/520 [24:54<07:39,  3.73s/it]
                                                 
{'loss': 1.2038, 'grad_norm': 0.0008120555296137997, 'learning_rate': 0.027979475346387363, 'epoch': 0.76}
+
 76%|███████▋  | 397/520 [24:54<07:39,  3.73s/it]
 77%|███████▋  | 398/520 [24:58<07:34,  3.72s/it]
                                                 
{'loss': 1.1999, 'grad_norm': 0.0008657316807012219, 'learning_rate': 0.027548433914072735, 'epoch': 0.77}
+
 77%|███████▋  | 398/520 [24:58<07:34,  3.72s/it]
 77%|███████▋  | 399/520 [25:02<07:31,  3.73s/it]
                                                 
{'loss': 1.1672, 'grad_norm': 0.0007874027904184501, 'learning_rate': 0.027120207524827168, 'epoch': 0.77}
+
 77%|███████▋  | 399/520 [25:02<07:31,  3.73s/it]
 77%|███████▋  | 400/520 [25:06<07:26,  3.72s/it]
                                                 
{'loss': 1.1958, 'grad_norm': 0.0007420105460811616, 'learning_rate': 0.02669481281701739, 'epoch': 0.77}
+
 77%|███████▋  | 400/520 [25:06<07:26,  3.72s/it]
 77%|███████▋  | 401/520 [25:09<07:21,  3.71s/it]
                                                 
{'loss': 1.0433, 'grad_norm': 0.0008801184071867263, 'learning_rate': 0.026272266318987603, 'epoch': 0.77}
+
 77%|███████▋  | 401/520 [25:09<07:21,  3.71s/it]
 77%|███████▋  | 402/520 [25:13<07:17,  3.71s/it]
                                                 
{'loss': 1.1663, 'grad_norm': 0.0008279727028498112, 'learning_rate': 0.02585258444841733, 'epoch': 0.77}
+
 77%|███████▋  | 402/520 [25:13<07:17,  3.71s/it]
 78%|███████▊  | 403/520 [25:17<07:11,  3.69s/it]
                                                 
{'loss': 1.1888, 'grad_norm': 0.0008996014243582944, 'learning_rate': 0.025435783511683442, 'epoch': 0.78}
+
 78%|███████▊  | 403/520 [25:17<07:11,  3.69s/it]
 78%|███████▊  | 404/520 [25:20<07:08,  3.70s/it]
                                                 
{'loss': 1.1004, 'grad_norm': 0.0009433303909752815, 'learning_rate': 0.02502187970322657, 'epoch': 0.78}
+
 78%|███████▊  | 404/520 [25:20<07:08,  3.70s/it]
 78%|███████▊  | 405/520 [25:24<07:05,  3.70s/it]
                                                 
{'loss': 1.1778, 'grad_norm': 0.000813893158650642, 'learning_rate': 0.02461088910492202, 'epoch': 0.78}
+
 78%|███████▊  | 405/520 [25:24<07:05,  3.70s/it]
 78%|███████▊  | 406/520 [25:28<07:02,  3.70s/it]
                                                 
{'loss': 1.105, 'grad_norm': 0.0009896461092877842, 'learning_rate': 0.02420282768545469, 'epoch': 0.78}
+
 78%|███████▊  | 406/520 [25:28<07:02,  3.70s/it]
 78%|███████▊  | 407/520 [25:31<06:58,  3.70s/it]
                                                 
{'loss': 1.2778, 'grad_norm': 0.0008456555821581969, 'learning_rate': 0.02379771129969892, 'epoch': 0.78}
+
 78%|███████▊  | 407/520 [25:31<06:58,  3.70s/it]
 78%|███████▊  | 408/520 [25:35<06:55,  3.71s/it]
                                                 
{'loss': 1.1815, 'grad_norm': 0.0008903054498920023, 'learning_rate': 0.023395555688102213, 'epoch': 0.78}
+
 78%|███████▊  | 408/520 [25:35<06:55,  3.71s/it]
 79%|███████▊  | 409/520 [25:39<06:51,  3.71s/it]
                                                 
{'loss': 1.3039, 'grad_norm': 0.000885810806683387, 'learning_rate': 0.02299637647607372, 'epoch': 0.79}
+
 79%|███████▊  | 409/520 [25:39<06:51,  3.71s/it]
 79%|███████▉  | 410/520 [25:43<06:47,  3.70s/it]
                                                 
{'loss': 1.0373, 'grad_norm': 0.0008025063565294821, 'learning_rate': 0.022600189173377264, 'epoch': 0.79}
+
 79%|███████▉  | 410/520 [25:43<06:47,  3.70s/it]
 79%|███████▉  | 411/520 [25:46<06:43,  3.70s/it]
                                                 
{'loss': 1.2808, 'grad_norm': 0.0009111135083882514, 'learning_rate': 0.022207009173528525, 'epoch': 0.79}
+
 79%|███████▉  | 411/520 [25:46<06:43,  3.70s/it]
 79%|███████▉  | 412/520 [25:50<06:39,  3.70s/it]
                                                 
{'loss': 1.1866, 'grad_norm': 0.0008308242213293154, 'learning_rate': 0.02181685175319702, 'epoch': 0.79}
+
 79%|███████▉  | 412/520 [25:50<06:39,  3.70s/it]
 79%|███████▉  | 413/520 [25:54<06:35,  3.70s/it]
                                                 
{'loss': 1.1982, 'grad_norm': 0.0007986485328236338, 'learning_rate': 0.021429732071612653, 'epoch': 0.79}
+
 79%|███████▉  | 413/520 [25:54<06:35,  3.70s/it]
 80%|███████▉  | 414/520 [25:57<06:32,  3.70s/it]
                                                 
{'loss': 1.0012, 'grad_norm': 0.0006778020094542154, 'learning_rate': 0.02104566516997647, 'epoch': 0.8}
+
 80%|███████▉  | 414/520 [25:57<06:32,  3.70s/it]
 80%|███████▉  | 415/520 [26:01<06:29,  3.71s/it]
                                                 
{'loss': 1.1724, 'grad_norm': 0.0007630902271984882, 'learning_rate': 0.020664665970876496, 'epoch': 0.8}
+
 80%|███████▉  | 415/520 [26:01<06:29,  3.71s/it]
 80%|████████  | 416/520 [26:05<06:25,  3.70s/it]
                                                 
{'loss': 1.086, 'grad_norm': 0.0008819756411564103, 'learning_rate': 0.020286749277707784, 'epoch': 0.8}
+
 80%|████████  | 416/520 [26:05<06:25,  3.70s/it]
 80%|████████  | 417/520 [26:08<06:21,  3.70s/it]
                                                 
{'loss': 1.2437, 'grad_norm': 0.0008374692691333609, 'learning_rate': 0.019911929774097215, 'epoch': 0.8}
+
 80%|████████  | 417/520 [26:08<06:21,  3.70s/it]
 80%|████████  | 418/520 [26:12<06:17,  3.70s/it]
                                                 
{'loss': 1.2354, 'grad_norm': 0.0007751965945991629, 'learning_rate': 0.019540222023333165, 'epoch': 0.8}
+
 80%|████████  | 418/520 [26:12<06:17,  3.70s/it]
 81%|████████  | 419/520 [26:16<06:13,  3.70s/it]
                                                 
{'loss': 1.2262, 'grad_norm': 0.000930858036016473, 'learning_rate': 0.01917164046779948, 'epoch': 0.81}
+
 81%|████████  | 419/520 [26:16<06:13,  3.70s/it]
 81%|████████  | 420/520 [26:20<06:08,  3.68s/it]
                                                 
{'loss': 1.1179, 'grad_norm': 0.0008517771357075663, 'learning_rate': 0.018806199428414352, 'epoch': 0.81}
+
 81%|████████  | 420/520 [26:20<06:08,  3.68s/it]
 81%|████████  | 421/520 [26:23<06:04,  3.68s/it]
                                                 
{'loss': 1.0556, 'grad_norm': 0.0008526804169636902, 'learning_rate': 0.018443913104073985, 'epoch': 0.81}
+
 81%|████████  | 421/520 [26:23<06:04,  3.68s/it]
 81%|████████  | 422/520 [26:27<05:59,  3.67s/it]
                                                 
{'loss': 1.1723, 'grad_norm': 0.0008585286686565852, 'learning_rate': 0.01808479557110081, 'epoch': 0.81}
+
 81%|████████  | 422/520 [26:27<05:59,  3.67s/it]
 81%|████████▏ | 423/520 [26:31<05:56,  3.68s/it]
                                                 
{'loss': 1.1465, 'grad_norm': 0.0008931731000442683, 'learning_rate': 0.017728860782696667, 'epoch': 0.81}
+
 81%|████████▏ | 423/520 [26:31<05:56,  3.68s/it]
 82%|████████▏ | 424/520 [26:34<05:53,  3.68s/it]
                                                 
{'loss': 1.2856, 'grad_norm': 0.0008069265064794872, 'learning_rate': 0.017376122568400532, 'epoch': 0.82}
+
 82%|████████▏ | 424/520 [26:34<05:53,  3.68s/it]
 82%|████████▏ | 425/520 [26:38<05:54,  3.73s/it]
                                                 
{'loss': 1.1676, 'grad_norm': 0.0008193841916340304, 'learning_rate': 0.017026594633551252, 'epoch': 0.82}
+
 82%|████████▏ | 425/520 [26:38<05:54,  3.73s/it]
 82%|████████▏ | 426/520 [26:42<05:53,  3.76s/it]
                                                 
{'loss': 1.1967, 'grad_norm': 0.0010472014377934475, 'learning_rate': 0.01668029055875512, 'epoch': 0.82}
+
 82%|████████▏ | 426/520 [26:42<05:53,  3.76s/it]
 82%|████████▏ | 427/520 [26:46<05:51,  3.78s/it]
                                                 
{'loss': 1.1017, 'grad_norm': 0.0008026648543301461, 'learning_rate': 0.016337223799358026, 'epoch': 0.82}
+
 82%|████████▏ | 427/520 [26:46<05:51,  3.78s/it]
 82%|████████▏ | 428/520 [26:49<05:47,  3.78s/it]
                                                 
{'loss': 1.0844, 'grad_norm': 0.0008754366945244988, 'learning_rate': 0.01599740768492286, 'epoch': 0.82}
+
 82%|████████▏ | 428/520 [26:49<05:47,  3.78s/it]
 82%|████████▎ | 429/520 [26:53<05:44,  3.79s/it]
                                                 
{'loss': 1.1845, 'grad_norm': 0.0008226745963536027, 'learning_rate': 0.015660855418711452, 'epoch': 0.82}
+
 82%|████████▎ | 429/520 [26:53<05:44,  3.79s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2076 > 2048). Running this sequence through the model will result in indexing errors
+
 83%|████████▎ | 430/520 [26:57<05:41,  3.79s/it]
                                                 
{'loss': 1.1781, 'grad_norm': 0.0007802251877995353, 'learning_rate': 0.015327580077171589, 'epoch': 0.83}
+
 83%|████████▎ | 430/520 [26:57<05:41,  3.79s/it]
 83%|████████▎ | 431/520 [27:01<05:38,  3.80s/it]
                                                 
{'loss': 1.1695, 'grad_norm': 0.000828757431388498, 'learning_rate': 0.014997594609429088, 'epoch': 0.83}
+
 83%|████████▎ | 431/520 [27:01<05:38,  3.80s/it]
 83%|████████▎ | 432/520 [27:05<05:34,  3.80s/it]
                                                 
{'loss': 1.0896, 'grad_norm': 0.0008476511423660411, 'learning_rate': 0.01467091183678444, 'epoch': 0.83}
+
 83%|████████▎ | 432/520 [27:05<05:34,  3.80s/it]
 83%|████████▎ | 433/520 [27:08<05:29,  3.79s/it]
                                                 
{'loss': 1.2267, 'grad_norm': 0.0008176544907625566, 'learning_rate': 0.014347544452214867, 'epoch': 0.83}
+
 83%|████████▎ | 433/520 [27:08<05:29,  3.79s/it]
 83%|████████▎ | 434/520 [27:12<05:25,  3.79s/it]
                                                 
{'loss': 0.9699, 'grad_norm': 0.0008255815069848533, 'learning_rate': 0.014027505019880971, 'epoch': 0.83}
+
 83%|████████▎ | 434/520 [27:12<05:25,  3.79s/it]
 84%|████████▎ | 435/520 [27:16<05:25,  3.83s/it]
                                                 
{'loss': 1.2623, 'grad_norm': 0.0009078494703712409, 'learning_rate': 0.013710805974638696, 'epoch': 0.84}
+
 84%|████████▎ | 435/520 [27:16<05:25,  3.83s/it]
 84%|████████▍ | 436/520 [27:20<05:27,  3.90s/it]
                                                 
{'loss': 1.0604, 'grad_norm': 0.0008494530321492785, 'learning_rate': 0.01339745962155613, 'epoch': 0.84}
+
 84%|████████▍ | 436/520 [27:20<05:27,  3.90s/it]
 84%|████████▍ | 437/520 [27:24<05:23,  3.89s/it]
                                                 
{'loss': 1.2856, 'grad_norm': 0.0008515880574511693, 'learning_rate': 0.01308747813543536, 'epoch': 0.84}
+
 84%|████████▍ | 437/520 [27:24<05:23,  3.89s/it]
 84%|████████▍ | 438/520 [27:28<05:17,  3.87s/it]
                                                 
{'loss': 1.097, 'grad_norm': 0.000828657411005991, 'learning_rate': 0.012780873560339467, 'epoch': 0.84}
+
 84%|████████▍ | 438/520 [27:28<05:17,  3.87s/it]
 84%|████████▍ | 439/520 [27:32<05:11,  3.85s/it]
                                                 
{'loss': 1.1503, 'grad_norm': 0.0006916731983654988, 'learning_rate': 0.012477657809124632, 'epoch': 0.84}
+
 84%|████████▍ | 439/520 [27:32<05:11,  3.85s/it]
 85%|████████▍ | 440/520 [27:36<05:06,  3.83s/it]
                                                 
{'loss': 1.1362, 'grad_norm': 0.0008956150436428211, 'learning_rate': 0.012177842662977134, 'epoch': 0.85}
+
 85%|████████▍ | 440/520 [27:36<05:06,  3.83s/it]
 85%|████████▍ | 441/520 [27:39<05:01,  3.81s/it]
                                                 
{'loss': 1.1705, 'grad_norm': 0.000785564702579343, 'learning_rate': 0.01188143977095576, 'epoch': 0.85}
+
 85%|████████▍ | 441/520 [27:39<05:01,  3.81s/it]
 85%|████████▌ | 442/520 [27:43<04:57,  3.81s/it]
                                                 
{'loss': 1.1952, 'grad_norm': 0.0009115164512311156, 'learning_rate': 0.011588460649539035, 'epoch': 0.85}
+
 85%|████████▌ | 442/520 [27:43<04:57,  3.81s/it]
 85%|████████▌ | 443/520 [27:47<04:53,  3.81s/it]
                                                 
{'loss': 1.213, 'grad_norm': 0.0008096995175558625, 'learning_rate': 0.011298916682177829, 'epoch': 0.85}
+
 85%|████████▌ | 443/520 [27:47<04:53,  3.81s/it]
 85%|████████▌ | 444/520 [27:51<04:49,  3.81s/it]
                                                 
{'loss': 1.1775, 'grad_norm': 0.000747503523064445, 'learning_rate': 0.011012819118853146, 'epoch': 0.85}
+
 85%|████████▌ | 444/520 [27:51<04:49,  3.81s/it]
 86%|████████▌ | 445/520 [27:55<04:44,  3.80s/it]
                                                 
{'loss': 1.1089, 'grad_norm': 0.0008512617338313648, 'learning_rate': 0.01073017907563887, 'epoch': 0.86}
+
 86%|████████▌ | 445/520 [27:55<04:44,  3.80s/it]
 86%|████████▌ | 446/520 [27:58<04:41,  3.80s/it]
                                                 
{'loss': 1.2447, 'grad_norm': 0.0007548490435289269, 'learning_rate': 0.010451007534269908, 'epoch': 0.86}
+
 86%|████████▌ | 446/520 [27:58<04:41,  3.80s/it]
 86%|████████▌ | 447/520 [28:02<04:37,  3.80s/it]
                                                 
{'loss': 1.1784, 'grad_norm': 0.0008135849690905176, 'learning_rate': 0.010175315341715598, 'epoch': 0.86}
+
 86%|████████▌ | 447/520 [28:02<04:37,  3.80s/it]
 86%|████████▌ | 448/520 [28:06<04:34,  3.81s/it]
                                                 
{'loss': 1.1674, 'grad_norm': 0.0009248303206233663, 'learning_rate': 0.009903113209758098, 'epoch': 0.86}
+
 86%|████████▌ | 448/520 [28:06<04:34,  3.81s/it]
 86%|████████▋ | 449/520 [28:10<04:29,  3.80s/it]
                                                 
{'loss': 1.2035, 'grad_norm': 0.0008292729166579809, 'learning_rate': 0.009634411714576352, 'epoch': 0.86}
+
 86%|████████▋ | 449/520 [28:10<04:29,  3.80s/it]
 87%|████████▋ | 450/520 [28:13<04:25,  3.79s/it]
                                                 
{'loss': 1.2055, 'grad_norm': 0.0008515780433574366, 'learning_rate': 0.009369221296335007, 'epoch': 0.87}
+
 87%|████████▋ | 450/520 [28:13<04:25,  3.79s/it]
 87%|████████▋ | 451/520 [28:17<04:21,  3.79s/it]
                                                 
{'loss': 1.2008, 'grad_norm': 0.0008353060305996569, 'learning_rate': 0.009107552258778906, 'epoch': 0.87}
+
 87%|████████▋ | 451/520 [28:17<04:21,  3.79s/it]
 87%|████████▋ | 452/520 [28:21<04:17,  3.79s/it]
                                                 
{'loss': 1.2523, 'grad_norm': 0.0007950330603018758, 'learning_rate': 0.008849414768832687, 'epoch': 0.87}
+
 87%|████████▋ | 452/520 [28:21<04:17,  3.79s/it]
 87%|████████▋ | 453/520 [28:25<04:13,  3.78s/it]
                                                 
{'loss': 1.2221, 'grad_norm': 0.0007844354035258398, 'learning_rate': 0.008594818856205699, 'epoch': 0.87}
+
 87%|████████▋ | 453/520 [28:25<04:13,  3.78s/it]
 87%|████████▋ | 454/520 [28:29<04:10,  3.79s/it]
                                                 
{'loss': 1.113, 'grad_norm': 0.0008150279377949262, 'learning_rate': 0.00834377441300238, 'epoch': 0.87}
+
 87%|████████▋ | 454/520 [28:29<04:10,  3.79s/it]
 88%|████████▊ | 455/520 [28:32<04:06,  3.79s/it]
                                                 
{'loss': 1.2473, 'grad_norm': 0.0008205291369071589, 'learning_rate': 0.008096291193337934, 'epoch': 0.88}
+
 88%|████████▊ | 455/520 [28:32<04:06,  3.79s/it]
 88%|████████▊ | 456/520 [28:36<04:03,  3.80s/it]
                                                 
{'loss': 1.1864, 'grad_norm': 0.000832636360611624, 'learning_rate': 0.007852378812959226, 'epoch': 0.88}
+
 88%|████████▊ | 456/520 [28:36<04:03,  3.80s/it]
 88%|████████▊ | 457/520 [28:40<03:59,  3.80s/it]
                                                 
{'loss': 1.1318, 'grad_norm': 0.0007465971867570751, 'learning_rate': 0.007612046748871327, 'epoch': 0.88}
+
 88%|████████▊ | 457/520 [28:40<03:59,  3.80s/it]
 88%|████████▊ | 458/520 [28:44<03:55,  3.80s/it]
                                                 
{'loss': 1.3073, 'grad_norm': 0.0009125076344618415, 'learning_rate': 0.007375304338969136, 'epoch': 0.88}
+
 88%|████████▊ | 458/520 [28:44<03:55,  3.80s/it]
 88%|████████▊ | 459/520 [28:48<03:51,  3.80s/it]
                                                 
{'loss': 1.23, 'grad_norm': 0.0008091944639391444, 'learning_rate': 0.007142160781674645, 'epoch': 0.88}
+
 88%|████████▊ | 459/520 [28:48<03:51,  3.80s/it]
 88%|████████▊ | 460/520 [28:51<03:47,  3.80s/it]
                                                 
{'loss': 1.1236, 'grad_norm': 0.0008179668647457307, 'learning_rate': 0.006912625135579587, 'epoch': 0.88}
+
 88%|████████▊ | 460/520 [28:51<03:47,  3.80s/it]
 89%|████████▊ | 461/520 [28:55<03:44,  3.80s/it]
                                                 
{'loss': 1.2162, 'grad_norm': 0.0006670075591484949, 'learning_rate': 0.0066867063190933496, 'epoch': 0.89}
+
 89%|████████▊ | 461/520 [28:55<03:44,  3.80s/it]
 89%|████████▉ | 462/520 [28:59<03:40,  3.79s/it]
                                                 
{'loss': 1.2916, 'grad_norm': 0.0008281837417988649, 'learning_rate': 0.006464413110096601, 'epoch': 0.89}
+
 89%|████████▉ | 462/520 [28:59<03:40,  3.79s/it]
 89%|████████▉ | 463/520 [29:03<03:36,  3.79s/it]
                                                 
{'loss': 1.0893, 'grad_norm': 0.0008491910721560469, 'learning_rate': 0.006245754145600091, 'epoch': 0.89}
+
 89%|████████▉ | 463/520 [29:03<03:36,  3.79s/it]
 89%|████████▉ | 464/520 [29:06<03:30,  3.75s/it]
                                                 
{'loss': 1.2204, 'grad_norm': 0.0008451749904476141, 'learning_rate': 0.006030737921409169, 'epoch': 0.89}
+
 89%|████████▉ | 464/520 [29:06<03:30,  3.75s/it]
 89%|████████▉ | 465/520 [29:10<03:25,  3.73s/it]
                                                 
{'loss': 1.3276, 'grad_norm': 0.0008622497445055682, 'learning_rate': 0.005819372791793654, 'epoch': 0.89}
+
 89%|████████▉ | 465/520 [29:10<03:25,  3.73s/it]
 90%|████████▉ | 466/520 [29:14<03:20,  3.71s/it]
                                                 
{'loss': 1.2179, 'grad_norm': 0.0007580164300607254, 'learning_rate': 0.005611666969163243, 'epoch': 0.9}
+
 90%|████████▉ | 466/520 [29:14<03:20,  3.71s/it]
 90%|████████▉ | 467/520 [29:18<03:16,  3.70s/it]
                                                 
{'loss': 1.1828, 'grad_norm': 0.000780666692973717, 'learning_rate': 0.005407628523748398, 'epoch': 0.9}
+
 90%|████████▉ | 467/520 [29:18<03:16,  3.70s/it]
 90%|█████████ | 468/520 [29:21<03:12,  3.70s/it]
                                                 
{'loss': 1.1854, 'grad_norm': 0.0009540482852192156, 'learning_rate': 0.00520726538328683, 'epoch': 0.9}
+
 90%|█████████ | 468/520 [29:21<03:12,  3.70s/it]
 90%|█████████ | 469/520 [29:25<03:08,  3.69s/it]
                                                 
{'loss': 1.2506, 'grad_norm': 0.0008994296950317519, 'learning_rate': 0.005010585332715401, 'epoch': 0.9}
+
 90%|█████████ | 469/520 [29:25<03:08,  3.69s/it]
 90%|█████████ | 470/520 [29:29<03:04,  3.68s/it]
                                                 
{'loss': 1.1227, 'grad_norm': 0.0007664653106057077, 'learning_rate': 0.004817596013867765, 'epoch': 0.9}
+
 90%|█████████ | 470/520 [29:29<03:04,  3.68s/it]
 91%|█████████ | 471/520 [29:32<03:00,  3.69s/it]
                                                 
{'loss': 1.1486, 'grad_norm': 0.0008941268365293736, 'learning_rate': 0.004628304925177318, 'epoch': 0.91}
+
 91%|█████████ | 471/520 [29:32<03:00,  3.69s/it]
 91%|█████████ | 472/520 [29:36<02:57,  3.70s/it]
                                                 
{'loss': 1.1172, 'grad_norm': 0.0007924799482620485, 'learning_rate': 0.004442719421385921, 'epoch': 0.91}
+
 91%|█████████ | 472/520 [29:36<02:57,  3.70s/it]
 91%|█████████ | 473/520 [29:40<02:53,  3.69s/it]
                                                 
{'loss': 1.1923, 'grad_norm': 0.0008588249443362459, 'learning_rate': 0.004260846713258193, 'epoch': 0.91}
+
 91%|█████████ | 473/520 [29:40<02:53,  3.69s/it]
 91%|█████████ | 474/520 [29:43<02:50,  3.70s/it]
                                                 
{'loss': 1.2175, 'grad_norm': 0.0008070992574055112, 'learning_rate': 0.004082693867301224, 'epoch': 0.91}
+
 91%|█████████ | 474/520 [29:43<02:50,  3.70s/it]
 91%|█████████▏| 475/520 [29:47<02:46,  3.70s/it]
                                                 
{'loss': 1.1284, 'grad_norm': 0.000809100559749551, 'learning_rate': 0.003908267805490051, 'epoch': 0.91}
+
 91%|█████████▏| 475/520 [29:47<02:46,  3.70s/it]
 92%|█████████▏| 476/520 [29:51<02:42,  3.69s/it]
                                                 
{'loss': 1.1738, 'grad_norm': 0.0008654396389442969, 'learning_rate': 0.003737575304998797, 'epoch': 0.92}
+
 92%|█████████▏| 476/520 [29:51<02:42,  3.69s/it]
 92%|█████████▏| 477/520 [29:54<02:38,  3.70s/it]
                                                 
{'loss': 1.1665, 'grad_norm': 0.0008989376666599901, 'learning_rate': 0.003570622997937234, 'epoch': 0.92}
+
 92%|█████████▏| 477/520 [29:54<02:38,  3.70s/it]
 92%|█████████▏| 478/520 [29:58<02:34,  3.69s/it]
                                                 
{'loss': 1.1039, 'grad_norm': 0.0008050926913155507, 'learning_rate': 0.00340741737109318, 'epoch': 0.92}
+
 92%|█████████▏| 478/520 [29:58<02:34,  3.69s/it]
 92%|█████████▏| 479/520 [30:02<02:31,  3.68s/it]
                                                 
{'loss': 1.185, 'grad_norm': 0.0008758240434861823, 'learning_rate': 0.003247964765680389, 'epoch': 0.92}
+
 92%|█████████▏| 479/520 [30:02<02:31,  3.68s/it]
 92%|█████████▏| 480/520 [30:06<02:27,  3.69s/it]
                                                 
{'loss': 1.2026, 'grad_norm': 0.000776106749590503, 'learning_rate': 0.0030922713770922153, 'epoch': 0.92}
+
 92%|█████████▏| 480/520 [30:06<02:27,  3.69s/it]
 92%|█████████▎| 481/520 [30:09<02:23,  3.69s/it]
                                                 
{'loss': 1.1942, 'grad_norm': 0.0007738123459334898, 'learning_rate': 0.0029403432546609046, 'epoch': 0.93}
+
 92%|█████████▎| 481/520 [30:09<02:23,  3.69s/it]
 93%|█████████▎| 482/520 [30:13<02:20,  3.69s/it]
                                                 
{'loss': 1.2102, 'grad_norm': 0.0007721300750546789, 'learning_rate': 0.0027921863014225504, 'epoch': 0.93}
+
 93%|█████████▎| 482/520 [30:13<02:20,  3.69s/it]
 93%|█████████▎| 483/520 [30:17<02:16,  3.70s/it]
                                                 
{'loss': 1.1839, 'grad_norm': 0.0008428985842876483, 'learning_rate': 0.002647806273887665, 'epoch': 0.93}
+
 93%|█████████▎| 483/520 [30:17<02:16,  3.70s/it]
 93%|█████████▎| 484/520 [30:20<02:13,  3.70s/it]
                                                 
{'loss': 1.1885, 'grad_norm': 0.0008338504955864701, 'learning_rate': 0.0025072087818176383, 'epoch': 0.93}
+
 93%|█████████▎| 484/520 [30:20<02:13,  3.70s/it]
 93%|█████████▎| 485/520 [30:24<02:09,  3.70s/it]
                                                 
{'loss': 1.1431, 'grad_norm': 0.0008166841379283582, 'learning_rate': 0.002370399288006664, 'epoch': 0.93}
+
 93%|█████████▎| 485/520 [30:24<02:09,  3.70s/it]
 93%|█████████▎| 486/520 [30:28<02:05,  3.70s/it]
                                                 
{'loss': 1.2618, 'grad_norm': 0.0008517600689568233, 'learning_rate': 0.0022373831080695463, 'epoch': 0.93}
+
 93%|█████████▎| 486/520 [30:28<02:05,  3.70s/it]
 94%|█████████▎| 487/520 [30:31<02:02,  3.73s/it]
                                                 
{'loss': 1.1198, 'grad_norm': 0.0008073967023695956, 'learning_rate': 0.0021081654102351635, 'epoch': 0.94}
+
 94%|█████████▎| 487/520 [30:31<02:02,  3.73s/it]
 94%|█████████▍| 488/520 [30:35<01:58,  3.72s/it]
                                                 
{'loss': 1.0654, 'grad_norm': 0.0008897104540378354, 'learning_rate': 0.0019827512151456175, 'epoch': 0.94}
+
 94%|█████████▍| 488/520 [30:35<01:58,  3.72s/it]
 94%|█████████▍| 489/520 [30:39<01:54,  3.71s/it]
                                                 
{'loss': 1.2146, 'grad_norm': 0.0007146661508777825, 'learning_rate': 0.0018611453956612345, 'epoch': 0.94}
+
 94%|█████████▍| 489/520 [30:39<01:54,  3.71s/it]
 94%|█████████▍| 490/520 [30:43<01:51,  3.71s/it]
                                                 
{'loss': 1.1894, 'grad_norm': 0.0008404780039447094, 'learning_rate': 0.0017433526766711727, 'epoch': 0.94}
+
 94%|█████████▍| 490/520 [30:43<01:51,  3.71s/it]
 94%|█████████▍| 491/520 [30:46<01:47,  3.70s/it]
                                                 
{'loss': 1.1512, 'grad_norm': 0.0008638294239676733, 'learning_rate': 0.0016293776349098677, 'epoch': 0.94}
+
 94%|█████████▍| 491/520 [30:46<01:47,  3.70s/it]
 95%|█████████▍| 492/520 [30:50<01:43,  3.69s/it]
                                                 
{'loss': 1.2612, 'grad_norm': 0.0008653794068588022, 'learning_rate': 0.0015192246987791981, 'epoch': 0.95}
+
 95%|█████████▍| 492/520 [30:50<01:43,  3.69s/it]
 95%|█████████▍| 493/520 [30:54<01:39,  3.70s/it]
                                                 
{'loss': 1.2212, 'grad_norm': 0.0008358039912215987, 'learning_rate': 0.0014128981481764114, 'epoch': 0.95}
+
 95%|█████████▍| 493/520 [30:54<01:39,  3.70s/it]
 95%|█████████▌| 494/520 [30:57<01:36,  3.70s/it]
                                                 
{'loss': 1.1915, 'grad_norm': 0.0008019533556097153, 'learning_rate': 0.0013104021143278911, 'epoch': 0.95}
+
 95%|█████████▌| 494/520 [30:57<01:36,  3.70s/it]
 95%|█████████▌| 495/520 [31:01<01:32,  3.69s/it]
                                                 
{'loss': 1.1671, 'grad_norm': 0.0008737589691262165, 'learning_rate': 0.0012117405796285285, 'epoch': 0.95}
+
 95%|█████████▌| 495/520 [31:01<01:32,  3.69s/it]
 95%|█████████▌| 496/520 [31:05<01:28,  3.69s/it]
                                                 
{'loss': 1.0944, 'grad_norm': 0.000834235639913945, 'learning_rate': 0.0011169173774871477, 'epoch': 0.95}
+
 95%|█████████▌| 496/520 [31:05<01:28,  3.69s/it]
 96%|█████████▌| 497/520 [31:08<01:24,  3.69s/it]
                                                 
{'loss': 1.1421, 'grad_norm': 0.0007383312425031822, 'learning_rate': 0.0010259361921774012, 'epoch': 0.96}
+
 96%|█████████▌| 497/520 [31:08<01:24,  3.69s/it]
 96%|█████████▌| 498/520 [31:12<01:21,  3.69s/it]
                                                 
{'loss': 1.1639, 'grad_norm': 0.000829206648339827, 'learning_rate': 0.000938800558694719, 'epoch': 0.96}
+
 96%|█████████▌| 498/520 [31:12<01:21,  3.69s/it]
 96%|█████████▌| 499/520 [31:16<01:17,  3.69s/it]
                                                 
{'loss': 1.2813, 'grad_norm': 0.0007956669120223702, 'learning_rate': 0.0008555138626189618, 'epoch': 0.96}
+
 96%|█████████▌| 499/520 [31:16<01:17,  3.69s/it]
 96%|█████████▌| 500/520 [31:19<01:13,  3.67s/it]
                                                 
{'loss': 1.2829, 'grad_norm': 0.000985774066769775, 'learning_rate': 0.0007760793399827937, 'epoch': 0.96}
+
 96%|█████████▌| 500/520 [31:19<01:13,  3.67s/it]
 96%|█████████▋| 501/520 [31:23<01:09,  3.66s/it]
                                                 
{'loss': 1.1866, 'grad_norm': 0.0008873653215458743, 'learning_rate': 0.000700500077146038, 'epoch': 0.96}
+
 96%|█████████▋| 501/520 [31:23<01:09,  3.66s/it]
 97%|█████████▋| 502/520 [31:27<01:05,  3.66s/it]
                                                 
{'loss': 1.1992, 'grad_norm': 0.000803165741306542, 'learning_rate': 0.0006287790106757397, 'epoch': 0.97}
+
 97%|█████████▋| 502/520 [31:27<01:05,  3.66s/it]
 97%|█████████▋| 503/520 [31:30<01:02,  3.67s/it]
                                                 
{'loss': 1.1773, 'grad_norm': 0.0008644889102500898, 'learning_rate': 0.0005609189272320237, 'epoch': 0.97}
+
 97%|█████████▋| 503/520 [31:30<01:02,  3.67s/it]
 97%|█████████▋| 504/520 [31:34<00:58,  3.68s/it]
                                                 
{'loss': 1.1933, 'grad_norm': 0.0009601909231002968, 'learning_rate': 0.000496922463459859, 'epoch': 0.97}
+
 97%|█████████▋| 504/520 [31:34<00:58,  3.68s/it]
 97%|█████████▋| 505/520 [31:38<00:55,  3.67s/it]
                                                 
{'loss': 1.2319, 'grad_norm': 0.0008545070102187552, 'learning_rate': 0.0004367921058866187, 'epoch': 0.97}
+
 97%|█████████▋| 505/520 [31:38<00:55,  3.67s/it]
 97%|█████████▋| 506/520 [31:41<00:51,  3.67s/it]
                                                 
{'loss': 1.1557, 'grad_norm': 0.000858139222197754, 'learning_rate': 0.0003805301908254455, 'epoch': 0.97}
+
 97%|█████████▋| 506/520 [31:41<00:51,  3.67s/it]
 98%|█████████▊| 507/520 [31:45<00:47,  3.67s/it]
                                                 
{'loss': 1.3231, 'grad_norm': 0.0007994581696170612, 'learning_rate': 0.0003281389042844918, 'epoch': 0.97}
+
 98%|█████████▊| 507/520 [31:45<00:47,  3.67s/it]
 98%|█████████▊| 508/520 [31:49<00:44,  3.68s/it]
                                                 
{'loss': 1.2736, 'grad_norm': 0.0008531476261603529, 'learning_rate': 0.00027962028188198705, 'epoch': 0.98}
+
 98%|█████████▊| 508/520 [31:49<00:44,  3.68s/it]
 98%|█████████▊| 509/520 [31:52<00:40,  3.68s/it]
                                                 
{'loss': 1.2419, 'grad_norm': 0.0008036157003883641, 'learning_rate': 0.00023497620876711256, 'epoch': 0.98}
+
 98%|█████████▊| 509/520 [31:52<00:40,  3.68s/it]
 98%|█████████▊| 510/520 [31:56<00:36,  3.68s/it]
                                                 
{'loss': 1.1902, 'grad_norm': 0.0008393776899569102, 'learning_rate': 0.00019420841954681523, 'epoch': 0.98}
+
 98%|█████████▊| 510/520 [31:56<00:36,  3.68s/it]
 98%|█████████▊| 511/520 [32:00<00:33,  3.68s/it]
                                                 
{'loss': 1.1552, 'grad_norm': 0.0008192325373959474, 'learning_rate': 0.00015731849821833956, 'epoch': 0.98}
+
 98%|█████████▊| 511/520 [32:00<00:33,  3.68s/it]
 98%|█████████▊| 512/520 [32:04<00:29,  3.68s/it]
                                                 
{'loss': 1.0489, 'grad_norm': 0.0008351269199725249, 'learning_rate': 0.00012430787810776555, 'epoch': 0.98}
+
 98%|█████████▊| 512/520 [32:04<00:29,  3.68s/it]
 99%|█████████▊| 513/520 [32:07<00:25,  3.69s/it]
                                                 
{'loss': 1.2516, 'grad_norm': 0.0009202138426680274, 'learning_rate': 9.517784181422018e-05, 'epoch': 0.99}
+
 99%|█████████▊| 513/520 [32:07<00:25,  3.69s/it]
 99%|█████████▉| 514/520 [32:11<00:22,  3.68s/it]
                                                 
{'loss': 1.2149, 'grad_norm': 0.0007720897731468561, 'learning_rate': 6.992952116013917e-05, 'epoch': 0.99}
+
 99%|█████████▉| 514/520 [32:11<00:22,  3.68s/it]
 99%|█████████▉| 515/520 [32:15<00:18,  3.67s/it]
                                                 
{'loss': 1.264, 'grad_norm': 0.0009663493444233865, 'learning_rate': 4.856389714723575e-05, 'epoch': 0.99}
+
 99%|█████████▉| 515/520 [32:15<00:18,  3.67s/it]
 99%|█████████▉| 516/520 [32:18<00:14,  3.66s/it]
                                                 
{'loss': 1.182, 'grad_norm': 0.0008307641532366346, 'learning_rate': 3.108179991837545e-05, 'epoch': 0.99}
+
 99%|█████████▉| 516/520 [32:18<00:14,  3.66s/it]
 99%|█████████▉| 517/520 [32:22<00:10,  3.64s/it]
                                                 
{'loss': 1.2185, 'grad_norm': 0.0008240325555137658, 'learning_rate': 1.7483908725357544e-05, 'epoch': 0.99}
+
 99%|█████████▉| 517/520 [32:22<00:10,  3.64s/it]
100%|█████████▉| 518/520 [32:25<00:07,  3.62s/it]
                                                 
{'loss': 1.1853, 'grad_norm': 0.0008801275682810427, 'learning_rate': 7.770751902513862e-06, 'epoch': 1.0}
+
100%|█████████▉| 518/520 [32:25<00:07,  3.62s/it]
100%|█████████▉| 519/520 [32:29<00:03,  3.62s/it]
                                                 
{'loss': 1.1853, 'grad_norm': 0.0008133024113494941, 'learning_rate': 1.9427068461808086e-06, 'epoch': 1.0}
+
100%|█████████▉| 519/520 [32:29<00:03,  3.62s/it]
100%|██████████| 520/520 [32:33<00:00,  3.88s/it]
                                                 
{'loss': 1.1923, 'grad_norm': 0.0007577466748367517, 'learning_rate': 0.0, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:33<00:00,  3.88s/it]
                                                 
{'train_runtime': 1953.9597, 'train_samples_per_second': 34.048, 'train_steps_per_second': 0.266, 'train_loss': 1.2446195717041308, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:33<00:00,  3.88s/it]
100%|██████████| 520/520 [32:33<00:00,  3.76s/it]
+[2025-10-16 10:10:03,518] [INFO] [launch.py:348:main] Process 2496549 exits successfully.
+[2025-10-16 10:10:04,519] [INFO] [launch.py:348:main] Process 2496550 exits successfully.
+[2025-10-16 10:10:04,520] [INFO] [launch.py:348:main] Process 2496548 exits successfully.
+[2025-10-16 10:10:04,520] [INFO] [launch.py:348:main] Process 2496552 exits successfully.
+[2025-10-16 10:10:04,521] [INFO] [launch.py:348:main] Process 2496546 exits successfully.
+[2025-10-16 10:10:05,522] [INFO] [launch.py:348:main] Process 2496551 exits successfully.
+[2025-10-16 10:10:05,523] [INFO] [launch.py:348:main] Process 2496547 exits successfully.
+[2025-10-16 10:10:08,526] [INFO] [launch.py:348:main] Process 2496545 exits successfully.
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_0.5_2e-1_connector-3.0_0.5_2e-1_ablation_20251016_092835.log
+Timestamp: 2025-10-16 10:10:11
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation_20251016_101011.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation_20251016_101011.log
new file mode 100644
index 0000000000000000000000000000000000000000..c8e3d2ef3b36008e3bf4444a5a87ce87e62db0a4
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation_20251016_101011.log
@@ -0,0 +1,849 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation_20251016_101011.log
+Timestamp: 2025-10-16 10:10:11
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 10:10:13,787] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:17,192] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 10:10:17,193] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 3.0 --temperature_attn_text 2.3 --temperature_mlp_text 2.3 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 3.0 --temperature_attn_vision 2.3 --temperature_mlp_vision 2.3 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 3.0 --temperature_connector 2.3 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 10:10:19,763] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:20,803] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 10:10:20,804] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 10:10:20,804] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 10:10:20,804] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 10:10:20,804] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 10:10:20,804] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 10:10:20,804] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 10:10:20,806] [INFO] [launch.py:253:main] process 2516080 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,808] [INFO] [launch.py:253:main] process 2516081 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,810] [INFO] [launch.py:253:main] process 2516082 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,812] [INFO] [launch.py:253:main] process 2516083 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,814] [INFO] [launch.py:253:main] process 2516084 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,816] [INFO] [launch.py:253:main] process 2516085 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,817] [INFO] [launch.py:253:main] process 2516086 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 10:10:20,819] [INFO] [launch.py:253:main] process 2516087 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 10:10:27,635] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,825] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,826] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,827] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,839] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,852] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,852] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:27,858] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 10:10:28,159] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,235] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,235] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 10:10:28,236] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,238] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,255] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,264] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,265] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 10:10:28,270] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 2.3, 'temperature_mlp': 2.3, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 2.3, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 2.3,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 2.3,
+    "temperature_mlp": 2.3,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2516080:2516080 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516080:2516080 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516080:2516080 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516080:2516080 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516080:2516080 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516080:2516080 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2516083:2516083 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516083:2516083 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516083:2516083 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516083:2516083 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516083:2516083 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516083:2516083 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516082:2516082 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516082:2516082 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516082:2516082 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516082:2516082 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516082:2516082 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516082:2516082 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516084:2516084 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516084:2516084 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516084:2516084 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516084:2516084 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516084:2516084 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516084:2516084 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516086:2516086 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516086:2516086 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516086:2516086 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516086:2516086 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516086:2516086 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516086:2516086 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516087:2516087 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516087:2516087 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516087:2516087 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516087:2516087 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516087:2516087 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516087:2516087 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516085:2516085 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516085:2516085 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516085:2516085 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516085:2516085 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516085:2516085 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516085:2516085 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516081:2516081 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2516081:2516081 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516081:2516081 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516081:2516081 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2516081:2516081 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2516081:2516081 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO ncclCommInitRank comm 0x55bebd16c530 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO ncclCommInitRank comm 0x562e8ab74030 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO ncclCommInitRank comm 0x5592f3b61f70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO ncclCommInitRank comm 0x55bf4cb56690 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO ncclCommInitRank comm 0x55c350e05b40 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO ncclCommInitRank comm 0x5583c54a1980 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO ncclCommInitRank comm 0x55d2457d7d60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO ncclCommInitRank comm 0x55bd1439f120 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xa201222bf068c93 - Init START
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO comm 0x55bebd16c530 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO comm 0x55d2457d7d60 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO comm 0x55bf4cb56690 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO comm 0x562e8ab74030 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO comm 0x5592f3b61f70 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO comm 0x55c350e05b40 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO comm 0x5583c54a1980 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO comm 0x55bd1439f120 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516087:2517651 [7] NCCL INFO ncclCommInitRank comm 0x562e8ab74030 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516084:2517649 [4] NCCL INFO ncclCommInitRank comm 0x55c350e05b40 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516086:2517650 [6] NCCL INFO ncclCommInitRank comm 0x55bebd16c530 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516080:2517646 [0] NCCL INFO ncclCommInitRank comm 0x55bd1439f120 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516082:2517648 [2] NCCL INFO ncclCommInitRank comm 0x55bf4cb56690 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516081:2517653 [1] NCCL INFO ncclCommInitRank comm 0x5583c54a1980 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516085:2517652 [5] NCCL INFO ncclCommInitRank comm 0x5592f3b61f70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xa201222bf068c93 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2516083:2517647 [3] NCCL INFO ncclCommInitRank comm 0x55d2457d7d60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xa201222bf068c93 - Init COMPLETE
+[2025-10-16 10:11:16,258] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[E ProcessGroupNCCL.cpp:474] [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800829 milliseconds before timing out.
+[E ProcessGroupNCCL.cpp:474] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800866 milliseconds before timing out.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+[E ProcessGroupNCCL.cpp:474] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800890 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2516081:2517681 [1] NCCL INFO [Service thread] Connection closed by localRank 1
+ywang29-vrdb-test2-worker-0:2516083:2517679 [3] NCCL INFO [Service thread] Connection closed by localRank 3
+ywang29-vrdb-test2-worker-0:2516083:2517201 [3] NCCL INFO comm 0x55d2457d7d60 rank 3 nranks 8 cudaDev 3 busId 201d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800866 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 3] NCCL watchdog thread terminated with exception: [Rank 3] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800866 milliseconds before timing out.
+ywang29-vrdb-test2-worker-0:2516081:2517199 [1] NCCL INFO comm 0x5583c54a1980 rank 1 nranks 8 cudaDev 1 busId 101d0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 1] NCCL watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800829 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 1] NCCL watchdog thread terminated with exception: [Rank 1] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800829 milliseconds before timing out.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+ywang29-vrdb-test2-worker-0:2516082:2517672 [2] NCCL INFO [Service thread] Connection closed by localRank 2
+ywang29-vrdb-test2-worker-0:2516082:2517195 [2] NCCL INFO comm 0x55bf4cb56690 rank 2 nranks 8 cudaDev 2 busId 201c0 - Abort COMPLETE
+[E ProcessGroupNCCL.cpp:488] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
+[E ProcessGroupNCCL.cpp:494] To avoid data inconsistency, we are taking the entire process down.
+[E ProcessGroupNCCL.cpp:915] [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800890 milliseconds before timing out.
+terminate called after throwing an instance of 'std::runtime_error'
+  what():  [Rank 2] NCCL watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=1858, OpType=BROADCAST, NumelIn=677376, NumelOut=677376, Timeout(ms)=1800000) ran for 1800890 milliseconds before timing out.
+[2025-10-16 16:19:29,709] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+[2025-10-16 16:19:39,498] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516080
+[2025-10-16 16:19:39,876] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516081
+[2025-10-16 16:19:39,876] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516082
+[2025-10-16 16:19:39,879] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516083
+[2025-10-16 16:19:40,013] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516084
+[2025-10-16 16:19:40,429] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516085
+[2025-10-16 16:19:40,846] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516086
+[2025-10-16 16:19:41,263] [INFO] [launch.py:316:sigkill_handler] Killing subprocess 2516087
+[2025-10-16 16:19:41,681] [ERROR] [launch.py:322:sigkill_handler] ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '3.0', '--temperature_attn_text', '2.3', '--temperature_mlp_text', '2.3', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '3.0', '--temperature_attn_vision', '2.3', '--temperature_mlp_vision', '2.3', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '3.0', '--temperature_connector', '2.3', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1'] exits with return code = -6
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-3.0_2.3_2e-1_connector-3.0_2.3_2e-1_ablation_20251016_101011.log
+Timestamp: 2025-10-16 16:19:43
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation_20251016_161943.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation_20251016_161943.log
new file mode 100644
index 0000000000000000000000000000000000000000..51709fe7cdce36fbbeca9eb20c512db831ad28a1
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation_20251016_161943.log
@@ -0,0 +1,2312 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation_20251016_161943.log
+Timestamp: 2025-10-16 16:19:43
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 16:19:45,823] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:49,140] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 16:19:49,142] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 5.0 --temperature_attn_text 2.5 --temperature_mlp_text 2.5 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 5.0 --temperature_attn_vision 2.5 --temperature_mlp_vision 2.5 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 5.0 --temperature_connector 2.5 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 16:19:51,717] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:52,768] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 16:19:52,768] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 16:19:52,768] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 16:19:52,768] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 16:19:52,768] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 16:19:52,768] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 16:19:52,768] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 16:19:52,771] [INFO] [launch.py:253:main] process 2523002 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,773] [INFO] [launch.py:253:main] process 2523003 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,774] [INFO] [launch.py:253:main] process 2523004 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,776] [INFO] [launch.py:253:main] process 2523005 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,778] [INFO] [launch.py:253:main] process 2523006 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,780] [INFO] [launch.py:253:main] process 2523007 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,782] [INFO] [launch.py:253:main] process 2523008 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:19:52,784] [INFO] [launch.py:253:main] process 2523009 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.5', '--temperature_mlp_text', '2.5', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.5', '--temperature_mlp_vision', '2.5', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.5', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 16:19:59,627] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,629] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,690] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,704] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,704] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,707] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,710] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:19:59,712] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:20:00,185] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,185] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,185] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,185] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,186] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,186] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,186] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:20:00,186] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 16:20:00,186] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 2.5, 'temperature_mlp': 2.5, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 2.5, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 2.5,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 2.5,
+    "temperature_mlp": 2.5,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2523002:2523002 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523002:2523002 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523002:2523002 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523002:2523002 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523002:2523002 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2523002:2523002 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2523006:2523006 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523006:2523006 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523006:2523006 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523006:2523006 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523006:2523006 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523006:2523006 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Using network Socket
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2523007:2523007 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523007:2523007 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523007:2523007 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523007:2523007 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523007:2523007 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523007:2523007 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2523005:2523005 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523005:2523005 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523005:2523005 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523005:2523005 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523005:2523005 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523005:2523005 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2523009:2523009 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523009:2523009 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523009:2523009 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523009:2523009 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523009:2523009 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523009:2523009 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2523004:2523004 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523004:2523004 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523004:2523004 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523004:2523004 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523004:2523004 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523004:2523004 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2523008:2523008 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523008:2523008 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523008:2523008 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523008:2523008 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523008:2523008 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523008:2523008 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523003:2523003 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2523003:2523003 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523003:2523003 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523003:2523003 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2523003:2523003 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2523003:2523003 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO ncclCommInitRank comm 0x55f15a736dc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO ncclCommInitRank comm 0x55bdc5ec2410 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO ncclCommInitRank comm 0x5567ddb8c210 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO ncclCommInitRank comm 0x5645871a7240 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO ncclCommInitRank comm 0x55defe25f380 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO ncclCommInitRank comm 0x5570017bd560 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO ncclCommInitRank comm 0x559abe55ad50 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO ncclCommInitRank comm 0x55f8c6cc1e00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xc15495574ba96372 - Init START
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO comm 0x5570017bd560 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO comm 0x55bdc5ec2410 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO comm 0x55f15a736dc0 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO comm 0x55f8c6cc1e00 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO comm 0x5567ddb8c210 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO comm 0x5645871a7240 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO comm 0x55defe25f380 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO comm 0x559abe55ad50 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523006:2524718 [4] NCCL INFO ncclCommInitRank comm 0x559abe55ad50 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523008:2524726 [6] NCCL INFO ncclCommInitRank comm 0x55defe25f380 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523004:2524725 [2] NCCL INFO ncclCommInitRank comm 0x55bdc5ec2410 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523002:2524717 [0] NCCL INFO ncclCommInitRank comm 0x55f8c6cc1e00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523005:2524723 [3] NCCL INFO ncclCommInitRank comm 0x5570017bd560 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523003:2524727 [1] NCCL INFO ncclCommInitRank comm 0x55f15a736dc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523007:2524719 [5] NCCL INFO ncclCommInitRank comm 0x5645871a7240 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xc15495574ba96372 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2523009:2524724 [7] NCCL INFO ncclCommInitRank comm 0x5567ddb8c210 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xc15495574ba96372 - Init COMPLETE
+[2025-10-16 16:20:45,169] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 16:20:46,906] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=5.000000
+Pre-training init connector._connector.0.scores: Mean=5.000005
+Pre-training init connector._connector.2.scores: Mean=4.999970
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 16:21:05,159 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 16:21:05,169 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO ncclCommInitRank comm 0x7ff00006aae0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO ncclCommInitRank comm 0x7f182006aa80 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO ncclCommInitRank comm 0x7f39d806b0a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO ncclCommInitRank comm 0x7f9a1c06b300 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO ncclCommInitRank comm 0x7f44f406b020 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO ncclCommInitRank comm 0x7fe07c06ac30 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO ncclCommInitRank comm 0x7fd6b006a690 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO ncclCommInitRank comm 0x7fb77006b210 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xeeed4a9e834f014e - Init START
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO comm 0x7ff00006aae0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO comm 0x7f182006aa80 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO comm 0x7f44f406b020 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO comm 0x7fe07c06ac30 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO comm 0x7f9a1c06b300 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO comm 0x7fd6b006a690 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO comm 0x7f39d806b0a0 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO comm 0x7fb77006b210 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2523005:2529776 [3] NCCL INFO ncclCommInitRank comm 0x7f39d806b0a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523002:2529769 [0] NCCL INFO ncclCommInitRank comm 0x7fb77006b210 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523004:2529773 [2] NCCL INFO ncclCommInitRank comm 0x7ff00006aae0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523003:2529772 [1] NCCL INFO ncclCommInitRank comm 0x7f182006aa80 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523008:2529771 [6] NCCL INFO ncclCommInitRank comm 0x7fe07c06ac30 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523009:2529775 [7] NCCL INFO ncclCommInitRank comm 0x7fd6b006a690 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523007:2529774 [5] NCCL INFO ncclCommInitRank comm 0x7f44f406b020 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xeeed4a9e834f014e - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2523006:2529770 [4] NCCL INFO ncclCommInitRank comm 0x7f9a1c06b300 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xeeed4a9e834f014e - Init COMPLETE
+
  0%|          | 1/520 [00:14<2:03:31, 14.28s/it]
                                                 
{'loss': 2.2894, 'grad_norm': 0.027402727813854326, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:14<2:03:31, 14.28s/it]
  0%|          | 2/520 [00:17<1:09:36,  8.06s/it]
                                                 
{'loss': 2.2411, 'grad_norm': 0.028359307088042747, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:17<1:09:36,  8.06s/it]
  1%|          | 3/520 [00:21<52:14,  6.06s/it]  
                                               
{'loss': 2.4151, 'grad_norm': 0.032556288440069074, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:21<52:14,  6.06s/it]
  1%|          | 4/520 [00:25<44:15,  5.15s/it]
                                               
{'loss': 1.7891, 'grad_norm': 0.011439645489641917, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:25<44:15,  5.15s/it]
  1%|          | 5/520 [00:29<39:49,  4.64s/it]
                                               
{'loss': 1.7753, 'grad_norm': 0.008088337427565727, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:29<39:49,  4.64s/it]
  1%|          | 6/520 [00:32<37:09,  4.34s/it]
                                               
{'loss': 1.5441, 'grad_norm': 0.006630398754448768, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:32<37:09,  4.34s/it]
  1%|▏         | 7/520 [00:36<35:18,  4.13s/it]
                                               
{'loss': 1.5652, 'grad_norm': 0.00923383344225508, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:36<35:18,  4.13s/it]
  2%|▏         | 8/520 [00:40<35:47,  4.19s/it]
                                               
{'loss': 1.5739, 'grad_norm': 0.006389834318106741, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:40<35:47,  4.19s/it]
  2%|▏         | 9/520 [00:45<35:57,  4.22s/it]
                                               
{'loss': 1.5821, 'grad_norm': 0.0035180361262384698, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:45<35:57,  4.22s/it]
  2%|▏         | 10/520 [00:48<34:31,  4.06s/it]
                                                
{'loss': 1.4158, 'grad_norm': 0.0038190788541624253, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:48<34:31,  4.06s/it]
  2%|▏         | 11/520 [00:52<33:45,  3.98s/it]
                                                
{'loss': 1.4944, 'grad_norm': 0.0047356892490299, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [00:52<33:45,  3.98s/it]
  2%|▏         | 12/520 [00:56<33:04,  3.91s/it]
                                                
{'loss': 1.4159, 'grad_norm': 0.00368931962770805, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [00:56<33:04,  3.91s/it][2025-10-16 16:22:10,678] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:00<34:09,  4.04s/it]
                                                
{'loss': 1.4251, 'grad_norm': 0.0025514366059197296, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:00<34:09,  4.04s/it]
  3%|▎         | 14/520 [01:04<33:11,  3.94s/it]
                                                
{'loss': 1.4711, 'grad_norm': 0.0030278371133849852, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:04<33:11,  3.94s/it]
  3%|▎         | 15/520 [01:08<32:32,  3.87s/it]
                                                
{'loss': 1.4573, 'grad_norm': 0.002710120661164573, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:08<32:32,  3.87s/it]
  3%|▎         | 16/520 [01:11<32:01,  3.81s/it]
                                                
{'loss': 1.4117, 'grad_norm': 0.002387082526938454, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:11<32:01,  3.81s/it]
  3%|▎         | 17/520 [01:15<31:42,  3.78s/it]
                                                
{'loss': 1.5, 'grad_norm': 0.0026348634650780094, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:15<31:42,  3.78s/it]
  3%|▎         | 18/520 [01:19<31:22,  3.75s/it]
                                                
{'loss': 1.3519, 'grad_norm': 0.002255251422328421, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:19<31:22,  3.75s/it]
  4%|▎         | 19/520 [01:22<31:11,  3.74s/it]
                                                
{'loss': 1.4055, 'grad_norm': 0.0019330492295537602, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:22<31:11,  3.74s/it]
  4%|▍         | 20/520 [01:26<31:01,  3.72s/it]
                                                
{'loss': 1.3441, 'grad_norm': 0.0023971679151212643, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:26<31:01,  3.72s/it]
  4%|▍         | 21/520 [01:30<31:02,  3.73s/it]
                                                
{'loss': 1.3974, 'grad_norm': 0.002203002048660798, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:30<31:02,  3.73s/it]
  4%|▍         | 22/520 [01:34<30:51,  3.72s/it]
                                                
{'loss': 1.494, 'grad_norm': 0.0020483246199332946, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:34<30:51,  3.72s/it]
  4%|▍         | 23/520 [01:37<30:43,  3.71s/it]
                                                
{'loss': 1.4281, 'grad_norm': 0.0017727410622117386, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:37<30:43,  3.71s/it]
  5%|▍         | 24/520 [01:41<30:35,  3.70s/it]
                                                
{'loss': 1.3791, 'grad_norm': 0.0018113899596769697, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [01:41<30:35,  3.70s/it]
  5%|▍         | 25/520 [01:45<30:27,  3.69s/it]
                                                
{'loss': 1.4307, 'grad_norm': 0.0020880062897188968, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [01:45<30:27,  3.69s/it]
  5%|▌         | 26/520 [01:48<30:26,  3.70s/it]
                                                
{'loss': 1.4112, 'grad_norm': 0.0018519725033277607, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [01:48<30:26,  3.70s/it]
  5%|▌         | 27/520 [01:52<30:18,  3.69s/it]
                                                
{'loss': 1.3265, 'grad_norm': 0.0020022112476405654, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [01:52<30:18,  3.69s/it]
  5%|▌         | 28/520 [01:56<30:16,  3.69s/it]
                                                
{'loss': 1.333, 'grad_norm': 0.001964773658041951, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [01:56<30:16,  3.69s/it]
  6%|▌         | 29/520 [01:59<30:13,  3.69s/it]
                                                
{'loss': 1.3573, 'grad_norm': 0.0017937852816791167, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [01:59<30:13,  3.69s/it]
  6%|▌         | 30/520 [02:03<30:23,  3.72s/it]
                                                
{'loss': 1.4509, 'grad_norm': 0.0016226065640172305, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:03<30:23,  3.72s/it]
  6%|▌         | 31/520 [02:07<30:32,  3.75s/it]
                                                
{'loss': 1.3325, 'grad_norm': 0.0016560691962547284, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:07<30:32,  3.75s/it]
  6%|▌         | 32/520 [02:11<30:22,  3.74s/it]
                                                
{'loss': 1.3105, 'grad_norm': 0.0017601936388892663, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:11<30:22,  3.74s/it]
  6%|▋         | 33/520 [02:14<30:13,  3.72s/it]
                                                
{'loss': 1.3367, 'grad_norm': 0.0017234566587645387, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:14<30:13,  3.72s/it]
  7%|▋         | 34/520 [02:18<30:12,  3.73s/it]
                                                
{'loss': 1.3238, 'grad_norm': 0.0017538933124099034, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:18<30:12,  3.73s/it]
  7%|▋         | 35/520 [02:22<30:19,  3.75s/it]
                                                
{'loss': 1.3405, 'grad_norm': 0.0021262668529726234, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:22<30:19,  3.75s/it]
  7%|▋         | 36/520 [02:26<30:25,  3.77s/it]
                                                
{'loss': 1.4343, 'grad_norm': 0.0016500732049230325, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:26<30:25,  3.77s/it]
  7%|▋         | 37/520 [02:30<30:27,  3.78s/it]
                                                
{'loss': 1.4253, 'grad_norm': 0.0015982854957311895, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:30<30:27,  3.78s/it]
  7%|▋         | 38/520 [02:33<30:27,  3.79s/it]
                                                
{'loss': 1.5074, 'grad_norm': 0.0017084145739177507, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:33<30:27,  3.79s/it]
  8%|▊         | 39/520 [02:37<30:17,  3.78s/it]
                                                
{'loss': 1.3563, 'grad_norm': 0.0018013409065585695, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [02:37<30:17,  3.78s/it]
  8%|▊         | 40/520 [02:41<29:55,  3.74s/it]
                                                
{'loss': 1.3875, 'grad_norm': 0.0016304384317021442, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [02:41<29:55,  3.74s/it]
  8%|▊         | 41/520 [02:45<29:45,  3.73s/it]
                                                
{'loss': 1.3657, 'grad_norm': 0.0016470087752036484, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [02:45<29:45,  3.73s/it]
  8%|▊         | 42/520 [02:48<29:36,  3.72s/it]
                                                
{'loss': 1.3854, 'grad_norm': 0.0021606694453158715, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [02:48<29:36,  3.72s/it]
  8%|▊         | 43/520 [02:52<29:28,  3.71s/it]
                                                
{'loss': 1.3258, 'grad_norm': 0.0015085228788120411, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [02:52<29:28,  3.71s/it]
  8%|▊         | 44/520 [02:56<29:36,  3.73s/it]
                                                
{'loss': 1.4204, 'grad_norm': 0.0015995458474898582, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [02:56<29:36,  3.73s/it]
  9%|▊         | 45/520 [03:00<29:47,  3.76s/it]
                                                
{'loss': 1.3845, 'grad_norm': 0.0017364075336150917, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [03:00<29:47,  3.76s/it]
  9%|▉         | 46/520 [03:03<29:54,  3.79s/it]
                                                
{'loss': 1.4804, 'grad_norm': 0.0015617504031762099, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:03<29:54,  3.79s/it]
  9%|▉         | 47/520 [03:07<29:59,  3.81s/it]
                                                
{'loss': 1.3706, 'grad_norm': 0.0016526642871128418, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:07<29:59,  3.81s/it]
  9%|▉         | 48/520 [03:11<30:01,  3.82s/it]
                                                
{'loss': 1.3493, 'grad_norm': 0.0018253457245425758, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:11<30:01,  3.82s/it]
  9%|▉         | 49/520 [03:15<29:51,  3.80s/it]
                                                
{'loss': 1.3916, 'grad_norm': 0.0015700630531243001, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:15<29:51,  3.80s/it]
 10%|▉         | 50/520 [03:18<29:23,  3.75s/it]
                                                
{'loss': 1.3821, 'grad_norm': 0.0015760288308658184, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:18<29:23,  3.75s/it]
 10%|▉         | 51/520 [03:22<29:05,  3.72s/it]
                                                
{'loss': 1.3141, 'grad_norm': 0.0017854742201390656, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:22<29:05,  3.72s/it]
 10%|█         | 52/520 [03:26<28:57,  3.71s/it]
                                                
{'loss': 1.4418, 'grad_norm': 0.0019050674785142593, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:26<28:57,  3.71s/it]
 10%|█         | 53/520 [03:30<28:52,  3.71s/it]
                                                
{'loss': 1.4285, 'grad_norm': 0.001711051599779895, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:30<28:52,  3.71s/it]
 10%|█         | 54/520 [03:33<28:38,  3.69s/it]
                                                
{'loss': 1.3464, 'grad_norm': 0.0016461725836387242, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:33<28:38,  3.69s/it]
 11%|█         | 55/520 [03:37<28:34,  3.69s/it]
                                                
{'loss': 1.3179, 'grad_norm': 0.0018141877166014519, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [03:37<28:34,  3.69s/it]
 11%|█         | 56/520 [03:40<28:28,  3.68s/it]
                                                
{'loss': 1.4455, 'grad_norm': 0.0016685254964032897, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [03:41<28:28,  3.68s/it]
 11%|█         | 57/520 [03:44<28:26,  3.69s/it]
                                                
{'loss': 1.3088, 'grad_norm': 0.001950762411247369, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [03:44<28:26,  3.69s/it]
 11%|█         | 58/520 [03:48<28:22,  3.69s/it]
                                                
{'loss': 1.4723, 'grad_norm': 0.0016949302684881775, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [03:48<28:22,  3.69s/it]
 11%|█▏        | 59/520 [03:52<28:20,  3.69s/it]
                                                
{'loss': 1.2888, 'grad_norm': 0.0015019494857582102, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [03:52<28:20,  3.69s/it]
 12%|█▏        | 60/520 [03:55<28:27,  3.71s/it]
                                                
{'loss': 1.3801, 'grad_norm': 0.0016618672325071223, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [03:55<28:27,  3.71s/it]
 12%|█▏        | 61/520 [03:59<28:21,  3.71s/it]
                                                
{'loss': 1.377, 'grad_norm': 0.0018311931934312429, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [03:59<28:21,  3.71s/it]
 12%|█▏        | 62/520 [04:03<28:10,  3.69s/it]
                                                
{'loss': 1.3568, 'grad_norm': 0.0017285655655872232, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [04:03<28:10,  3.69s/it]
 12%|█▏        | 63/520 [04:06<28:06,  3.69s/it]
                                                
{'loss': 1.3468, 'grad_norm': 0.0015469362171265262, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:06<28:06,  3.69s/it]
 12%|█▏        | 64/520 [04:10<28:28,  3.75s/it]
                                                
{'loss': 1.376, 'grad_norm': 0.0017625184151040048, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:10<28:28,  3.75s/it]
 12%|█▎        | 65/520 [04:14<28:39,  3.78s/it]
                                                
{'loss': 1.3821, 'grad_norm': 0.002044890088573096, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:14<28:39,  3.78s/it]
 13%|█▎        | 66/520 [04:18<29:09,  3.85s/it]
                                                
{'loss': 1.334, 'grad_norm': 0.001429293357168382, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:18<29:09,  3.85s/it]
 13%|█▎        | 67/520 [04:22<29:04,  3.85s/it]
                                                
{'loss': 1.2395, 'grad_norm': 0.0016591724802696556, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:22<29:04,  3.85s/it]
 13%|█▎        | 68/520 [04:26<28:45,  3.82s/it]
                                                
{'loss': 1.3089, 'grad_norm': 0.0018955458866495866, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:26<28:45,  3.82s/it]
 13%|█▎        | 69/520 [04:29<28:23,  3.78s/it]
                                                
{'loss': 1.2863, 'grad_norm': 0.0016604109907293403, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:29<28:23,  3.78s/it]
 13%|█▎        | 70/520 [04:33<28:18,  3.77s/it]
                                                
{'loss': 1.318, 'grad_norm': 0.0017012189915838198, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [04:33<28:18,  3.77s/it]
 14%|█▎        | 71/520 [04:37<28:16,  3.78s/it]
                                                
{'loss': 1.2617, 'grad_norm': 0.0015217956180879645, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [04:37<28:16,  3.78s/it]
 14%|█▍        | 72/520 [04:41<27:59,  3.75s/it]
                                                
{'loss': 1.4111, 'grad_norm': 0.001889806636088258, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [04:41<27:59,  3.75s/it]
 14%|█▍        | 73/520 [04:44<27:45,  3.73s/it]
                                                
{'loss': 1.2374, 'grad_norm': 0.001560930150801324, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [04:44<27:45,  3.73s/it]
 14%|█▍        | 74/520 [04:48<27:38,  3.72s/it]
                                                
{'loss': 1.3433, 'grad_norm': 0.0017657722876304406, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [04:48<27:38,  3.72s/it]
 14%|█▍        | 75/520 [04:52<27:49,  3.75s/it]
                                                
{'loss': 1.2546, 'grad_norm': 0.0015072133407260007, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [04:52<27:49,  3.75s/it]
 15%|█▍        | 76/520 [04:56<28:07,  3.80s/it]
                                                
{'loss': 1.4221, 'grad_norm': 0.0014149481646556223, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [04:56<28:07,  3.80s/it]
 15%|█▍        | 77/520 [05:00<28:18,  3.83s/it]
                                                
{'loss': 1.1787, 'grad_norm': 0.0016501297436572666, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [05:00<28:18,  3.83s/it]
 15%|█▌        | 78/520 [05:04<28:17,  3.84s/it]
                                                
{'loss': 1.2956, 'grad_norm': 0.001608588895868481, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:04<28:17,  3.84s/it]
 15%|█▌        | 79/520 [05:07<28:16,  3.85s/it]
                                                
{'loss': 1.2843, 'grad_norm': 0.0015484869762293233, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:07<28:16,  3.85s/it]
 15%|█▌        | 80/520 [05:11<28:16,  3.86s/it]
                                                
{'loss': 1.4096, 'grad_norm': 0.0015688827681863437, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:11<28:16,  3.86s/it]
 16%|█▌        | 81/520 [05:15<28:10,  3.85s/it]
                                                
{'loss': 1.4215, 'grad_norm': 0.0021302573143341893, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:15<28:10,  3.85s/it]
 16%|█▌        | 82/520 [05:19<28:04,  3.85s/it]
                                                
{'loss': 1.348, 'grad_norm': 0.0014582086162089533, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:19<28:04,  3.85s/it]
 16%|█▌        | 83/520 [05:23<27:58,  3.84s/it]
                                                
{'loss': 1.3736, 'grad_norm': 0.001775652588849699, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:23<27:58,  3.84s/it]
 16%|█▌        | 84/520 [05:27<27:56,  3.85s/it]
                                                
{'loss': 1.3805, 'grad_norm': 0.001646084359355721, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:27<27:56,  3.85s/it]
 16%|█▋        | 85/520 [05:30<27:53,  3.85s/it]
                                                
{'loss': 1.4118, 'grad_norm': 0.001554876214368193, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:30<27:53,  3.85s/it]
 17%|█▋        | 86/520 [05:34<27:35,  3.81s/it]
                                                
{'loss': 1.4086, 'grad_norm': 0.0015445352868994286, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:34<27:35,  3.81s/it]
 17%|█▋        | 87/520 [05:38<27:25,  3.80s/it]
                                                
{'loss': 1.3564, 'grad_norm': 0.0015159804938173814, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [05:38<27:25,  3.80s/it]
 17%|█▋        | 88/520 [05:42<27:13,  3.78s/it]
                                                
{'loss': 1.3057, 'grad_norm': 0.0013373901471830586, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [05:42<27:13,  3.78s/it]
 17%|█▋        | 89/520 [05:45<27:00,  3.76s/it]
                                                
{'loss': 1.3716, 'grad_norm': 0.0016861604007217563, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [05:45<27:00,  3.76s/it]
 17%|█▋        | 90/520 [05:49<26:47,  3.74s/it]
                                                
{'loss': 1.3011, 'grad_norm': 0.001475114507589954, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [05:49<26:47,  3.74s/it]
 18%|█▊        | 91/520 [05:53<26:36,  3.72s/it]
                                                
{'loss': 1.3764, 'grad_norm': 0.001433317563651212, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [05:53<26:36,  3.72s/it]
 18%|█▊        | 92/520 [05:57<26:31,  3.72s/it]
                                                
{'loss': 1.3037, 'grad_norm': 0.0015201888245824702, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [05:57<26:31,  3.72s/it]
 18%|█▊        | 93/520 [06:00<26:28,  3.72s/it]
                                                
{'loss': 1.3311, 'grad_norm': 0.0017888800682353395, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [06:00<26:28,  3.72s/it]
 18%|█▊        | 94/520 [06:04<26:26,  3.72s/it]
                                                
{'loss': 1.4085, 'grad_norm': 0.001554429557135333, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [06:04<26:26,  3.72s/it]
 18%|█▊        | 95/520 [06:08<26:26,  3.73s/it]
                                                
{'loss': 1.3041, 'grad_norm': 0.002107433516458913, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:08<26:26,  3.73s/it]
 18%|█▊        | 96/520 [06:11<26:22,  3.73s/it]
                                                
{'loss': 1.3173, 'grad_norm': 0.0013438219983261683, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:11<26:22,  3.73s/it]
 19%|█▊        | 97/520 [06:15<26:22,  3.74s/it]
                                                
{'loss': 1.2825, 'grad_norm': 0.0019528961861824032, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:15<26:22,  3.74s/it]
 19%|█▉        | 98/520 [06:19<26:14,  3.73s/it]
                                                
{'loss': 1.2873, 'grad_norm': 0.001281455479529964, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:19<26:14,  3.73s/it]
 19%|█▉        | 99/520 [06:23<26:07,  3.72s/it]
                                                
{'loss': 1.3029, 'grad_norm': 0.0017054842691373315, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:23<26:07,  3.72s/it]
 19%|█▉        | 100/520 [06:26<26:07,  3.73s/it]
                                                 
{'loss': 1.296, 'grad_norm': 0.0015487035223328966, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:26<26:07,  3.73s/it]
 19%|█▉        | 101/520 [06:30<26:04,  3.73s/it]
                                                 
{'loss': 1.3007, 'grad_norm': 0.0015477712384694732, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:30<26:04,  3.73s/it]
 20%|█▉        | 102/520 [06:34<25:58,  3.73s/it]
                                                 
{'loss': 1.3089, 'grad_norm': 0.001595719226407262, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [06:34<25:58,  3.73s/it]
 20%|█▉        | 103/520 [06:38<25:57,  3.74s/it]
                                                 
{'loss': 1.235, 'grad_norm': 0.0013486300344068309, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [06:38<25:57,  3.74s/it]
 20%|██        | 104/520 [06:41<25:46,  3.72s/it]
                                                 
{'loss': 1.3109, 'grad_norm': 0.0014786664909579577, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [06:41<25:46,  3.72s/it]
 20%|██        | 105/520 [06:45<25:42,  3.72s/it]
                                                 
{'loss': 1.2986, 'grad_norm': 0.0014116301031966748, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [06:45<25:42,  3.72s/it]
 20%|██        | 106/520 [06:49<25:34,  3.71s/it]
                                                 
{'loss': 1.3153, 'grad_norm': 0.001520246627928881, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [06:49<25:34,  3.71s/it]
 21%|██        | 107/520 [06:52<25:28,  3.70s/it]
                                                 
{'loss': 1.2916, 'grad_norm': 0.0014712001525528937, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [06:52<25:28,  3.70s/it]
 21%|██        | 108/520 [06:56<25:27,  3.71s/it]
                                                 
{'loss': 1.2643, 'grad_norm': 0.0015880136111143662, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [06:56<25:27,  3.71s/it]
 21%|██        | 109/520 [07:00<25:20,  3.70s/it]
                                                 
{'loss': 1.2727, 'grad_norm': 0.0013177127230921337, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [07:00<25:20,  3.70s/it]
 21%|██        | 110/520 [07:03<25:13,  3.69s/it]
                                                 
{'loss': 1.4457, 'grad_norm': 0.0016988353235256573, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [07:03<25:13,  3.69s/it]
 21%|██▏       | 111/520 [07:07<25:12,  3.70s/it]
                                                 
{'loss': 1.4473, 'grad_norm': 0.001620360280644039, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:07<25:12,  3.70s/it]
 22%|██▏       | 112/520 [07:11<25:25,  3.74s/it]
                                                 
{'loss': 1.3304, 'grad_norm': 0.0014098899954486262, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:11<25:25,  3.74s/it]
 22%|██▏       | 113/520 [07:15<25:34,  3.77s/it]
                                                 
{'loss': 1.2224, 'grad_norm': 0.0015083336353761167, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:15<25:34,  3.77s/it]
 22%|██▏       | 114/520 [07:19<25:37,  3.79s/it]
                                                 
{'loss': 1.3145, 'grad_norm': 0.0013969642331601287, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:19<25:37,  3.79s/it]
 22%|██▏       | 115/520 [07:22<25:38,  3.80s/it]
                                                 
{'loss': 1.4138, 'grad_norm': 0.001380028849921102, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:22<25:38,  3.80s/it]
 22%|██▏       | 116/520 [07:26<25:37,  3.81s/it]
                                                 
{'loss': 1.4285, 'grad_norm': 0.0015264923247073493, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:26<25:37,  3.81s/it]
 22%|██▎       | 117/520 [07:30<25:34,  3.81s/it]
                                                 
{'loss': 1.3954, 'grad_norm': 0.0015307353579888238, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:30<25:34,  3.81s/it]
 23%|██▎       | 118/520 [07:34<25:28,  3.80s/it]
                                                 
{'loss': 1.3007, 'grad_norm': 0.0013356920225467738, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [07:34<25:28,  3.80s/it]
 23%|██▎       | 119/520 [07:38<25:26,  3.81s/it]
                                                 
{'loss': 1.2584, 'grad_norm': 0.001563427789976061, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [07:38<25:26,  3.81s/it]
 23%|██▎       | 120/520 [07:42<25:24,  3.81s/it]
                                                 
{'loss': 1.2714, 'grad_norm': 0.001840968975486302, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [07:42<25:24,  3.81s/it]
 23%|██▎       | 121/520 [07:45<25:18,  3.80s/it]
                                                 
{'loss': 1.3284, 'grad_norm': 0.001546092982875519, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [07:45<25:18,  3.80s/it]
 23%|██▎       | 122/520 [07:49<25:12,  3.80s/it]
                                                 
{'loss': 1.2339, 'grad_norm': 0.0013785928797355968, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [07:49<25:12,  3.80s/it]
 24%|██▎       | 123/520 [07:53<24:55,  3.77s/it]
                                                 
{'loss': 1.3588, 'grad_norm': 0.0014369877532283632, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [07:53<24:55,  3.77s/it]
 24%|██▍       | 124/520 [07:56<24:42,  3.74s/it]
                                                 
{'loss': 1.2964, 'grad_norm': 0.0015209327192432208, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [07:56<24:42,  3.74s/it]
 24%|██▍       | 125/520 [08:00<24:34,  3.73s/it]
                                                 
{'loss': 1.2892, 'grad_norm': 0.0014788220063786512, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [08:00<24:34,  3.73s/it]
 24%|██▍       | 126/520 [08:05<25:41,  3.91s/it]
                                                 
{'loss': 1.283, 'grad_norm': 0.001270711383212928, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [08:05<25:41,  3.91s/it]
 24%|██▍       | 127/520 [08:08<25:33,  3.90s/it]
                                                 
{'loss': 1.269, 'grad_norm': 0.001604769234620614, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:08<25:33,  3.90s/it]
 25%|██▍       | 128/520 [08:12<25:22,  3.88s/it]
                                                 
{'loss': 1.3282, 'grad_norm': 0.0016404134976549252, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:12<25:22,  3.88s/it]
 25%|██▍       | 129/520 [08:16<25:11,  3.87s/it]
                                                 
{'loss': 1.2691, 'grad_norm': 0.001348770927982782, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:16<25:11,  3.87s/it]
 25%|██▌       | 130/520 [08:20<25:02,  3.85s/it]
                                                 
{'loss': 1.3057, 'grad_norm': 0.0012467682087527042, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:20<25:02,  3.85s/it]
 25%|██▌       | 131/520 [08:24<24:58,  3.85s/it]
                                                 
{'loss': 1.2586, 'grad_norm': 0.0012708820280895618, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:24<24:58,  3.85s/it]
 25%|██▌       | 132/520 [08:28<24:53,  3.85s/it]
                                                 
{'loss': 1.3581, 'grad_norm': 0.001459000263829508, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:28<24:53,  3.85s/it]
 26%|██▌       | 133/520 [08:31<24:45,  3.84s/it]
                                                 
{'loss': 1.2724, 'grad_norm': 0.001535388468376067, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:31<24:45,  3.84s/it]
 26%|██▌       | 134/520 [08:35<24:42,  3.84s/it]
                                                 
{'loss': 1.3542, 'grad_norm': 0.0014355694803484776, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [08:35<24:42,  3.84s/it]
 26%|██▌       | 135/520 [08:39<24:36,  3.84s/it]
                                                 
{'loss': 1.4032, 'grad_norm': 0.0014368168350245864, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [08:39<24:36,  3.84s/it]
 26%|██▌       | 136/520 [08:43<24:34,  3.84s/it]
                                                 
{'loss': 1.3435, 'grad_norm': 0.0014387344894945182, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [08:43<24:34,  3.84s/it]
 26%|██▋       | 137/520 [08:47<24:33,  3.85s/it]
                                                 
{'loss': 1.2537, 'grad_norm': 0.0016020562639928629, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [08:47<24:33,  3.85s/it]
 27%|██▋       | 138/520 [08:51<24:26,  3.84s/it]
                                                 
{'loss': 1.2725, 'grad_norm': 0.0014141956978789626, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [08:51<24:26,  3.84s/it]
 27%|██▋       | 139/520 [08:54<24:22,  3.84s/it]
                                                 
{'loss': 1.1632, 'grad_norm': 0.001328296473416864, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [08:54<24:22,  3.84s/it]
 27%|██▋       | 140/520 [08:58<24:03,  3.80s/it]
                                                 
{'loss': 1.3061, 'grad_norm': 0.001374401774563347, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [08:58<24:03,  3.80s/it]
 27%|██▋       | 141/520 [09:02<23:46,  3.76s/it]
                                                 
{'loss': 1.3808, 'grad_norm': 0.001406745217590043, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [09:02<23:46,  3.76s/it]
 27%|██▋       | 142/520 [09:06<23:39,  3.76s/it]
                                                 
{'loss': 1.3137, 'grad_norm': 0.001243039835452269, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [09:06<23:39,  3.76s/it]
 28%|██▊       | 143/520 [09:09<23:30,  3.74s/it]
                                                 
{'loss': 1.3014, 'grad_norm': 0.0016177560127282877, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:09<23:30,  3.74s/it]
 28%|██▊       | 144/520 [09:13<23:24,  3.74s/it]
                                                 
{'loss': 1.2607, 'grad_norm': 0.0014739059878440038, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:13<23:24,  3.74s/it]
 28%|██▊       | 145/520 [09:17<23:23,  3.74s/it]
                                                 
{'loss': 1.1998, 'grad_norm': 0.0012819585626219398, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:17<23:23,  3.74s/it]
 28%|██▊       | 146/520 [09:20<23:18,  3.74s/it]
                                                 
{'loss': 1.38, 'grad_norm': 0.0013944246926185949, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:20<23:18,  3.74s/it]
 28%|██▊       | 147/520 [09:24<23:09,  3.72s/it]
                                                 
{'loss': 1.2471, 'grad_norm': 0.0015140120798766213, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:24<23:09,  3.72s/it]
 28%|██▊       | 148/520 [09:28<23:03,  3.72s/it]
                                                 
{'loss': 1.2735, 'grad_norm': 0.001391050128863681, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:28<23:03,  3.72s/it]
 29%|██▊       | 149/520 [09:32<22:59,  3.72s/it]
                                                 
{'loss': 1.2153, 'grad_norm': 0.0013518371166393035, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:32<22:59,  3.72s/it]
 29%|██▉       | 150/520 [09:35<22:51,  3.71s/it]
                                                 
{'loss': 1.4473, 'grad_norm': 0.0014770045995029287, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [09:35<22:51,  3.71s/it]
 29%|██▉       | 151/520 [09:39<22:44,  3.70s/it]
                                                 
{'loss': 1.2638, 'grad_norm': 0.001355142581779714, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [09:39<22:44,  3.70s/it]
 29%|██▉       | 152/520 [09:43<22:38,  3.69s/it]
                                                 
{'loss': 1.232, 'grad_norm': 0.0014363133210204475, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [09:43<22:38,  3.69s/it]
 29%|██▉       | 153/520 [09:46<22:35,  3.69s/it]
                                                 
{'loss': 1.269, 'grad_norm': 0.0013509668526242, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [09:46<22:35,  3.69s/it]
 30%|██▉       | 154/520 [09:50<22:29,  3.69s/it]
                                                 
{'loss': 1.3499, 'grad_norm': 0.0013820186923128833, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [09:50<22:29,  3.69s/it]
 30%|██▉       | 155/520 [09:54<22:23,  3.68s/it]
                                                 
{'loss': 1.2609, 'grad_norm': 0.00136624072626432, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [09:54<22:23,  3.68s/it]
 30%|███       | 156/520 [09:57<22:25,  3.70s/it]
                                                 
{'loss': 1.2939, 'grad_norm': 0.0018319985957306237, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [09:57<22:25,  3.70s/it]
 30%|███       | 157/520 [10:01<22:21,  3.69s/it]
                                                 
{'loss': 1.371, 'grad_norm': 0.0013912644364244497, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [10:01<22:21,  3.69s/it]
 30%|███       | 158/520 [10:05<22:14,  3.69s/it]
                                                 
{'loss': 1.271, 'grad_norm': 0.0015565671184198963, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [10:05<22:14,  3.69s/it]
 31%|███       | 159/520 [10:09<22:29,  3.74s/it]
                                                 
{'loss': 1.3135, 'grad_norm': 0.0013483986358297534, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [10:09<22:29,  3.74s/it]
 31%|███       | 160/520 [10:12<22:39,  3.78s/it]
                                                 
{'loss': 1.3221, 'grad_norm': 0.0014216910252711063, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:12<22:39,  3.78s/it]
 31%|███       | 161/520 [10:16<22:41,  3.79s/it]
                                                 
{'loss': 1.3053, 'grad_norm': 0.0014038104515213788, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:16<22:41,  3.79s/it]
 31%|███       | 162/520 [10:20<22:39,  3.80s/it]
                                                 
{'loss': 1.3017, 'grad_norm': 0.001399734793908826, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:20<22:39,  3.80s/it]
 31%|███▏      | 163/520 [10:24<22:35,  3.80s/it]
                                                 
{'loss': 1.1967, 'grad_norm': 0.0018328817549332493, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:24<22:35,  3.80s/it]
 32%|███▏      | 164/520 [10:28<22:35,  3.81s/it]
                                                 
{'loss': 1.1614, 'grad_norm': 0.0014066587776350424, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:28<22:35,  3.81s/it]
 32%|███▏      | 165/520 [10:32<22:36,  3.82s/it]
                                                 
{'loss': 1.3001, 'grad_norm': 0.0012762625262767555, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:32<22:36,  3.82s/it]
 32%|███▏      | 166/520 [10:35<22:26,  3.80s/it]
                                                 
{'loss': 1.2881, 'grad_norm': 0.0015121630751522347, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:35<22:26,  3.80s/it]
 32%|███▏      | 167/520 [10:39<22:14,  3.78s/it]
                                                 
{'loss': 1.2799, 'grad_norm': 0.0015666459656232118, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [10:39<22:14,  3.78s/it]
 32%|███▏      | 168/520 [10:43<22:01,  3.76s/it]
                                                 
{'loss': 1.2197, 'grad_norm': 0.001416548888548296, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [10:43<22:01,  3.76s/it]
 32%|███▎      | 169/520 [10:47<21:54,  3.75s/it]
                                                 
{'loss': 1.2962, 'grad_norm': 0.0013763230135793152, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [10:47<21:54,  3.75s/it]
 33%|███▎      | 170/520 [10:50<21:47,  3.74s/it]
                                                 
{'loss': 1.2517, 'grad_norm': 0.0012731934444229391, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [10:50<21:47,  3.74s/it]
 33%|███▎      | 171/520 [10:54<21:42,  3.73s/it]
                                                 
{'loss': 1.2237, 'grad_norm': 0.0014062265747263086, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [10:54<21:42,  3.73s/it]
 33%|███▎      | 172/520 [10:58<21:36,  3.73s/it]
                                                 
{'loss': 1.3, 'grad_norm': 0.0013680929611645768, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [10:58<21:36,  3.73s/it]
 33%|███▎      | 173/520 [11:01<21:34,  3.73s/it]
                                                 
{'loss': 1.243, 'grad_norm': 0.0013852395042065946, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [11:01<21:34,  3.73s/it]
 33%|███▎      | 174/520 [11:05<21:28,  3.72s/it]
                                                 
{'loss': 1.2917, 'grad_norm': 0.0014794329333746377, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [11:05<21:28,  3.72s/it]
 34%|███▎      | 175/520 [11:09<21:30,  3.74s/it]
                                                 
{'loss': 1.2065, 'grad_norm': 0.0013585394440828995, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [11:09<21:30,  3.74s/it]
 34%|███▍      | 176/520 [11:13<21:23,  3.73s/it]
                                                 
{'loss': 1.3362, 'grad_norm': 0.0013500635166657535, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:13<21:23,  3.73s/it]
 34%|███▍      | 177/520 [11:16<21:19,  3.73s/it]
                                                 
{'loss': 1.207, 'grad_norm': 0.0014415830536350573, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:16<21:19,  3.73s/it]
 34%|███▍      | 178/520 [11:20<21:24,  3.75s/it]
                                                 
{'loss': 1.2721, 'grad_norm': 0.0015144633876567534, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:20<21:24,  3.75s/it]
 34%|███▍      | 179/520 [11:24<21:38,  3.81s/it]
                                                 
{'loss': 1.3519, 'grad_norm': 0.001290156636250817, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:24<21:38,  3.81s/it]
 35%|███▍      | 180/520 [11:28<21:49,  3.85s/it]
                                                 
{'loss': 1.2753, 'grad_norm': 0.0014636173427695074, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:28<21:49,  3.85s/it]
 35%|███▍      | 181/520 [11:32<21:45,  3.85s/it]
                                                 
{'loss': 1.2407, 'grad_norm': 0.0012446812340469447, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:32<21:45,  3.85s/it]
 35%|███▌      | 182/520 [11:36<21:42,  3.85s/it]
                                                 
{'loss': 1.2535, 'grad_norm': 0.0013533175278431942, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:36<21:42,  3.85s/it]
 35%|███▌      | 183/520 [11:39<21:27,  3.82s/it]
                                                 
{'loss': 1.2833, 'grad_norm': 0.0013314885630242644, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [11:39<21:27,  3.82s/it]
 35%|███▌      | 184/520 [11:43<21:14,  3.79s/it]
                                                 
{'loss': 1.21, 'grad_norm': 0.0013632014209018373, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [11:43<21:14,  3.79s/it]
 36%|███▌      | 185/520 [11:47<21:00,  3.76s/it]
                                                 
{'loss': 1.3592, 'grad_norm': 0.001380788868076505, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [11:47<21:00,  3.76s/it]
 36%|███▌      | 186/520 [11:51<20:46,  3.73s/it]
                                                 
{'loss': 1.2336, 'grad_norm': 0.0013535261676143268, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [11:51<20:46,  3.73s/it]
 36%|███▌      | 187/520 [11:54<20:40,  3.72s/it]
                                                 
{'loss': 1.238, 'grad_norm': 0.0015622756582046558, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [11:54<20:40,  3.72s/it]
 36%|███▌      | 188/520 [11:58<20:43,  3.75s/it]
                                                 
{'loss': 1.3174, 'grad_norm': 0.001388342097746068, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [11:58<20:43,  3.75s/it]
 36%|███▋      | 189/520 [12:02<20:52,  3.78s/it]
                                                 
{'loss': 1.3286, 'grad_norm': 0.0012387230104637468, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [12:02<20:52,  3.78s/it]
 37%|███▋      | 190/520 [12:06<20:54,  3.80s/it]
                                                 
{'loss': 1.2495, 'grad_norm': 0.0014473433351330181, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [12:06<20:54,  3.80s/it]
 37%|███▋      | 191/520 [12:10<20:58,  3.82s/it]
                                                 
{'loss': 1.1988, 'grad_norm': 0.0011891458833533494, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [12:10<20:58,  3.82s/it]
 37%|███▋      | 192/520 [12:14<20:58,  3.84s/it]
                                                 
{'loss': 1.2855, 'grad_norm': 0.0013168479855403402, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:14<20:58,  3.84s/it]
 37%|███▋      | 193/520 [12:17<20:54,  3.84s/it]
                                                 
{'loss': 1.2522, 'grad_norm': 0.0015566041923614413, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:17<20:54,  3.84s/it]
 37%|███▋      | 194/520 [12:21<20:53,  3.84s/it]
                                                 
{'loss': 1.1482, 'grad_norm': 0.0013869861129276117, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:21<20:53,  3.84s/it]
 38%|███▊      | 195/520 [12:25<20:53,  3.86s/it]
                                                 
{'loss': 1.3039, 'grad_norm': 0.0013295320157306948, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:25<20:53,  3.86s/it]
 38%|███▊      | 196/520 [12:29<20:31,  3.80s/it]
                                                 
{'loss': 1.2774, 'grad_norm': 0.0015070948865408928, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:29<20:31,  3.80s/it]
 38%|███▊      | 197/520 [12:32<20:17,  3.77s/it]
                                                 
{'loss': 1.2299, 'grad_norm': 0.00131306053585719, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:32<20:17,  3.77s/it]
 38%|███▊      | 198/520 [12:36<20:05,  3.74s/it]
                                                 
{'loss': 1.2979, 'grad_norm': 0.0013975126410417813, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:36<20:05,  3.74s/it]
 38%|███▊      | 199/520 [12:40<20:07,  3.76s/it]
                                                 
{'loss': 1.2126, 'grad_norm': 0.0013514943148329067, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [12:40<20:07,  3.76s/it]
 38%|███▊      | 200/520 [12:44<19:56,  3.74s/it]
                                                 
{'loss': 1.1986, 'grad_norm': 0.001363241606732063, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [12:44<19:56,  3.74s/it]
 39%|███▊      | 201/520 [12:47<19:52,  3.74s/it]
                                                 
{'loss': 1.2243, 'grad_norm': 0.0011527701756615615, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [12:47<19:52,  3.74s/it]
 39%|███▉      | 202/520 [12:51<19:41,  3.72s/it]
                                                 
{'loss': 1.2164, 'grad_norm': 0.0013406017683508036, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [12:51<19:41,  3.72s/it]
 39%|███▉      | 203/520 [12:55<19:33,  3.70s/it]
                                                 
{'loss': 1.273, 'grad_norm': 0.001398432334317827, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [12:55<19:33,  3.70s/it]
 39%|███▉      | 204/520 [12:58<19:27,  3.70s/it]
                                                 
{'loss': 1.2826, 'grad_norm': 0.0013672953501422628, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [12:58<19:27,  3.70s/it]
 39%|███▉      | 205/520 [13:02<19:38,  3.74s/it]
                                                 
{'loss': 1.2293, 'grad_norm': 0.0013583819951896857, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [13:02<19:38,  3.74s/it]
 40%|███▉      | 206/520 [13:06<19:49,  3.79s/it]
                                                 
{'loss': 1.3195, 'grad_norm': 0.0013295748736889847, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [13:06<19:49,  3.79s/it]
 40%|███▉      | 207/520 [13:10<19:53,  3.81s/it]
                                                 
{'loss': 1.2025, 'grad_norm': 0.0013177726644727518, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [13:10<19:53,  3.81s/it]
 40%|████      | 208/520 [13:14<19:58,  3.84s/it]
                                                 
{'loss': 1.3097, 'grad_norm': 0.0015728958147375987, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [13:14<19:58,  3.84s/it]
 40%|████      | 209/520 [13:18<19:59,  3.86s/it]
                                                 
{'loss': 1.2245, 'grad_norm': 0.0013084471931639763, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:18<19:59,  3.86s/it]
 40%|████      | 210/520 [13:22<19:48,  3.84s/it]
                                                 
{'loss': 1.2983, 'grad_norm': 0.001399300161236161, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:22<19:48,  3.84s/it]
 41%|████      | 211/520 [13:25<19:40,  3.82s/it]
                                                 
{'loss': 1.2977, 'grad_norm': 0.0012857903599912447, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:25<19:40,  3.82s/it]
 41%|████      | 212/520 [13:29<19:23,  3.78s/it]
                                                 
{'loss': 1.2859, 'grad_norm': 0.0013212483903179953, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:29<19:23,  3.78s/it]
 41%|████      | 213/520 [13:33<19:12,  3.75s/it]
                                                 
{'loss': 1.2466, 'grad_norm': 0.001679252764952199, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:33<19:12,  3.75s/it]
 41%|████      | 214/520 [13:36<19:02,  3.73s/it]
                                                 
{'loss': 1.2276, 'grad_norm': 0.001369267911012063, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:36<19:02,  3.73s/it]
 41%|████▏     | 215/520 [13:40<18:54,  3.72s/it]
                                                 
{'loss': 1.161, 'grad_norm': 0.0012949715774155555, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [13:40<18:54,  3.72s/it]
 42%|████▏     | 216/520 [13:44<18:46,  3.71s/it]
                                                 
{'loss': 1.1521, 'grad_norm': 0.001271624479836165, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [13:44<18:46,  3.71s/it]
 42%|████▏     | 217/520 [13:48<18:44,  3.71s/it]
                                                 
{'loss': 1.2785, 'grad_norm': 0.0013588434615225626, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [13:48<18:44,  3.71s/it]
 42%|████▏     | 218/520 [13:51<18:37,  3.70s/it]
                                                 
{'loss': 1.2663, 'grad_norm': 0.0014241973132934026, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [13:51<18:37,  3.70s/it]
 42%|████▏     | 219/520 [13:55<18:33,  3.70s/it]
                                                 
{'loss': 1.2625, 'grad_norm': 0.0012208755347863354, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [13:55<18:33,  3.70s/it]
 42%|████▏     | 220/520 [13:59<18:29,  3.70s/it]
                                                 
{'loss': 1.2049, 'grad_norm': 0.0012695696957024005, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [13:59<18:29,  3.70s/it]
 42%|████▎     | 221/520 [14:02<18:25,  3.70s/it]
                                                 
{'loss': 1.2662, 'grad_norm': 0.0013264119944988407, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [14:02<18:25,  3.70s/it]
 43%|████▎     | 222/520 [14:06<18:20,  3.69s/it]
                                                 
{'loss': 1.2014, 'grad_norm': 0.0013783157326450082, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [14:06<18:20,  3.69s/it]
 43%|████▎     | 223/520 [14:10<18:16,  3.69s/it]
                                                 
{'loss': 1.2009, 'grad_norm': 0.0012492588691971845, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [14:10<18:16,  3.69s/it]
 43%|████▎     | 224/520 [14:13<18:15,  3.70s/it]
                                                 
{'loss': 1.2804, 'grad_norm': 0.0013891824433099732, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [14:13<18:15,  3.70s/it]
 43%|████▎     | 225/520 [14:17<18:10,  3.70s/it]
                                                 
{'loss': 1.2072, 'grad_norm': 0.001306086314778979, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:17<18:10,  3.70s/it]
 43%|████▎     | 226/520 [14:21<18:02,  3.68s/it]
                                                 
{'loss': 1.3049, 'grad_norm': 0.0012864165504995167, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:21<18:02,  3.68s/it]
 44%|████▎     | 227/520 [14:24<17:56,  3.67s/it]
                                                 
{'loss': 1.2902, 'grad_norm': 0.0012473875606403828, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:24<17:56,  3.67s/it]
 44%|████▍     | 228/520 [14:28<17:51,  3.67s/it]
                                                 
{'loss': 1.3161, 'grad_norm': 0.0013064061716557631, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:28<17:51,  3.67s/it]
 44%|████▍     | 229/520 [14:32<17:44,  3.66s/it]
                                                 
{'loss': 1.2617, 'grad_norm': 0.001195389184972959, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:32<17:44,  3.66s/it]
 44%|████▍     | 230/520 [14:35<17:47,  3.68s/it]
                                                 
{'loss': 1.1548, 'grad_norm': 0.0013269576424331294, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:35<17:47,  3.68s/it]
 44%|████▍     | 231/520 [14:39<17:41,  3.67s/it]
                                                 
{'loss': 1.2152, 'grad_norm': 0.0012278278380953782, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:39<17:41,  3.67s/it]
 45%|████▍     | 232/520 [14:43<17:36,  3.67s/it]
                                                 
{'loss': 1.3479, 'grad_norm': 0.001480359074947828, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [14:43<17:36,  3.67s/it]
 45%|████▍     | 233/520 [14:46<17:32,  3.67s/it]
                                                 
{'loss': 1.2399, 'grad_norm': 0.0014713714410087954, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [14:46<17:32,  3.67s/it]
 45%|████▌     | 234/520 [14:50<17:28,  3.67s/it]
                                                 
{'loss': 1.1658, 'grad_norm': 0.0014154621447305899, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [14:50<17:28,  3.67s/it]
 45%|████▌     | 235/520 [14:54<17:24,  3.66s/it]
                                                 
{'loss': 1.2208, 'grad_norm': 0.0014059970999919262, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [14:54<17:24,  3.66s/it]
 45%|████▌     | 236/520 [14:57<17:28,  3.69s/it]
                                                 
{'loss': 1.2992, 'grad_norm': 0.0012247149953407606, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [14:57<17:28,  3.69s/it]
 46%|████▌     | 237/520 [15:01<17:33,  3.72s/it]
                                                 
{'loss': 1.2966, 'grad_norm': 0.0012874975217619928, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [15:01<17:33,  3.72s/it]
 46%|████▌     | 238/520 [15:05<17:39,  3.76s/it]
                                                 
{'loss': 1.2294, 'grad_norm': 0.00138636927658776, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [15:05<17:39,  3.76s/it]
 46%|████▌     | 239/520 [15:09<17:39,  3.77s/it]
                                                 
{'loss': 1.3077, 'grad_norm': 0.0013572513631160985, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [15:09<17:39,  3.77s/it]
 46%|████▌     | 240/520 [15:13<17:39,  3.79s/it]
                                                 
{'loss': 1.112, 'grad_norm': 0.0012726445328067324, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:13<17:39,  3.79s/it]
 46%|████▋     | 241/520 [15:17<17:38,  3.79s/it]
                                                 
{'loss': 1.1975, 'grad_norm': 0.0013245476872121116, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:17<17:38,  3.79s/it]
 47%|████▋     | 242/520 [15:20<17:37,  3.80s/it]
                                                 
{'loss': 1.2101, 'grad_norm': 0.0012222334020432213, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:20<17:37,  3.80s/it]
 47%|████▋     | 243/520 [15:24<17:34,  3.81s/it]
                                                 
{'loss': 1.205, 'grad_norm': 0.0013239575429965689, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:24<17:34,  3.81s/it]
 47%|████▋     | 244/520 [15:28<17:30,  3.81s/it]
                                                 
{'loss': 1.3156, 'grad_norm': 0.0013562618081011109, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:28<17:30,  3.81s/it]
 47%|████▋     | 245/520 [15:32<17:26,  3.80s/it]
                                                 
{'loss': 1.1845, 'grad_norm': 0.0014473445267021608, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:32<17:26,  3.80s/it]
 47%|████▋     | 246/520 [15:36<17:24,  3.81s/it]
                                                 
{'loss': 1.3314, 'grad_norm': 0.0013515847004646545, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:36<17:24,  3.81s/it]
 48%|████▊     | 247/520 [15:39<17:21,  3.81s/it]
                                                 
{'loss': 1.3611, 'grad_norm': 0.0013476453106544923, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:39<17:21,  3.81s/it]
 48%|████▊     | 248/520 [15:43<17:13,  3.80s/it]
                                                 
{'loss': 1.1908, 'grad_norm': 0.001376443627957418, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [15:43<17:13,  3.80s/it]
 48%|████▊     | 249/520 [15:47<17:04,  3.78s/it]
                                                 
{'loss': 1.2829, 'grad_norm': 0.0012816586327013461, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [15:47<17:04,  3.78s/it]
 48%|████▊     | 250/520 [15:51<16:56,  3.76s/it]
                                                 
{'loss': 1.2217, 'grad_norm': 0.0014238560514600038, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [15:51<16:56,  3.76s/it]
 48%|████▊     | 251/520 [15:54<16:56,  3.78s/it]
                                                 
{'loss': 1.281, 'grad_norm': 0.0012529308925299562, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [15:54<16:56,  3.78s/it]
 48%|████▊     | 252/520 [15:58<16:59,  3.80s/it]
                                                 
{'loss': 1.2351, 'grad_norm': 0.0012432690908527981, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [15:58<16:59,  3.80s/it]
 49%|████▊     | 253/520 [16:02<17:03,  3.83s/it]
                                                 
{'loss': 1.2802, 'grad_norm': 0.0015017622574536094, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [16:02<17:03,  3.83s/it]
 49%|████▉     | 254/520 [16:06<17:02,  3.84s/it]
                                                 
{'loss': 1.2142, 'grad_norm': 0.0012203982658957459, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [16:06<17:02,  3.84s/it]
 49%|████▉     | 255/520 [16:10<17:01,  3.85s/it]
                                                 
{'loss': 1.2215, 'grad_norm': 0.0014721002562094133, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [16:10<17:01,  3.85s/it]
 49%|████▉     | 256/520 [16:14<16:57,  3.85s/it]
                                                 
{'loss': 1.2695, 'grad_norm': 0.0013812076254407844, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:14<16:57,  3.85s/it]
 49%|████▉     | 257/520 [16:18<16:55,  3.86s/it]
                                                 
{'loss': 1.2498, 'grad_norm': 0.0013397029864657408, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:18<16:55,  3.86s/it]
 50%|████▉     | 258/520 [16:21<16:46,  3.84s/it]
                                                 
{'loss': 1.2615, 'grad_norm': 0.001231902860368231, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:21<16:46,  3.84s/it]
 50%|████▉     | 259/520 [16:25<16:32,  3.80s/it]
                                                 
{'loss': 1.3382, 'grad_norm': 0.0015629410007929464, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:25<16:32,  3.80s/it]
 50%|█████     | 260/520 [16:29<16:27,  3.80s/it]
                                                 
{'loss': 1.2786, 'grad_norm': 0.0012794382400055102, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:29<16:27,  3.80s/it]
 50%|█████     | 261/520 [16:33<16:32,  3.83s/it]
                                                 
{'loss': 1.2254, 'grad_norm': 0.0013088622544783473, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:33<16:32,  3.83s/it]
 50%|█████     | 262/520 [16:37<16:34,  3.85s/it]
                                                 
{'loss': 1.1894, 'grad_norm': 0.0012807848815660633, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:37<16:34,  3.85s/it]
 51%|█████     | 263/520 [16:41<16:33,  3.87s/it]
                                                 
{'loss': 1.2451, 'grad_norm': 0.0012845749565636836, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:41<16:33,  3.87s/it]
 51%|█████     | 264/520 [16:45<16:33,  3.88s/it]
                                                 
{'loss': 1.2876, 'grad_norm': 0.0012974342365209553, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [16:45<16:33,  3.88s/it]
 51%|█████     | 265/520 [16:49<16:36,  3.91s/it]
                                                 
{'loss': 1.2005, 'grad_norm': 0.0014062695586780892, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [16:49<16:36,  3.91s/it]
 51%|█████     | 266/520 [16:52<16:32,  3.91s/it]
                                                 
{'loss': 1.0704, 'grad_norm': 0.001201798808344782, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [16:52<16:32,  3.91s/it]
 51%|█████▏    | 267/520 [16:56<16:29,  3.91s/it]
                                                 
{'loss': 1.2003, 'grad_norm': 0.0012984177529403744, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [16:56<16:29,  3.91s/it]
 52%|█████▏    | 268/520 [17:00<16:25,  3.91s/it]
                                                 
{'loss': 1.354, 'grad_norm': 0.0018393447976742874, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [17:00<16:25,  3.91s/it]
 52%|█████▏    | 269/520 [17:04<16:20,  3.91s/it]
                                                 
{'loss': 1.2986, 'grad_norm': 0.0013371393775111805, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [17:04<16:20,  3.91s/it]
 52%|█████▏    | 270/520 [17:08<16:18,  3.91s/it]
                                                 
{'loss': 1.169, 'grad_norm': 0.0012268609208475056, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [17:08<16:18,  3.91s/it]
 52%|█████▏    | 271/520 [17:12<16:13,  3.91s/it]
                                                 
{'loss': 1.2876, 'grad_norm': 0.0013317825915099433, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [17:12<16:13,  3.91s/it]
 52%|█████▏    | 272/520 [17:16<16:09,  3.91s/it]
                                                 
{'loss': 1.1913, 'grad_norm': 0.0013487697151285355, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:16<16:09,  3.91s/it]
 52%|█████▎    | 273/520 [17:20<16:02,  3.90s/it]
                                                 
{'loss': 1.3045, 'grad_norm': 0.0012922154055937816, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:20<16:02,  3.90s/it]
 53%|█████▎    | 274/520 [17:24<15:58,  3.90s/it]
                                                 
{'loss': 1.2651, 'grad_norm': 0.0014979995204542484, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:24<15:58,  3.90s/it]
 53%|█████▎    | 275/520 [17:28<15:52,  3.89s/it]
                                                 
{'loss': 1.2089, 'grad_norm': 0.0014628179952703463, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:28<15:52,  3.89s/it]
 53%|█████▎    | 276/520 [17:31<15:47,  3.88s/it]
                                                 
{'loss': 1.272, 'grad_norm': 0.0015138691419437338, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:31<15:47,  3.88s/it]
 53%|█████▎    | 277/520 [17:35<15:44,  3.89s/it]
                                                 
{'loss': 1.3059, 'grad_norm': 0.0012518871621282194, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:35<15:44,  3.89s/it]
 53%|█████▎    | 278/520 [17:39<15:40,  3.89s/it]
                                                 
{'loss': 1.1538, 'grad_norm': 0.0011663042195928107, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:39<15:40,  3.89s/it]
 54%|█████▎    | 279/520 [17:43<15:34,  3.88s/it]
                                                 
{'loss': 1.19, 'grad_norm': 0.0014232457534258448, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:43<15:34,  3.88s/it]
 54%|█████▍    | 280/520 [17:47<15:30,  3.88s/it]
                                                 
{'loss': 1.2022, 'grad_norm': 0.0014638647452385562, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [17:47<15:30,  3.88s/it]
 54%|█████▍    | 281/520 [17:51<15:24,  3.87s/it]
                                                 
{'loss': 1.3018, 'grad_norm': 0.0014050076154680272, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [17:51<15:24,  3.87s/it]
 54%|█████▍    | 282/520 [17:55<15:21,  3.87s/it]
                                                 
{'loss': 1.1709, 'grad_norm': 0.0012136587478393057, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [17:55<15:21,  3.87s/it]
 54%|█████▍    | 283/520 [17:59<15:21,  3.89s/it]
                                                 
{'loss': 1.3162, 'grad_norm': 0.0014112994663469823, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [17:59<15:21,  3.89s/it]
 55%|█████▍    | 284/520 [18:02<15:16,  3.88s/it]
                                                 
{'loss': 1.1899, 'grad_norm': 0.0014448901363593811, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [18:03<15:16,  3.88s/it]
 55%|█████▍    | 285/520 [18:06<15:12,  3.88s/it]
                                                 
{'loss': 1.1937, 'grad_norm': 0.0012850288670504487, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [18:06<15:12,  3.88s/it]
 55%|█████▌    | 286/520 [18:10<15:08,  3.88s/it]
                                                 
{'loss': 1.0775, 'grad_norm': 0.0014273038528227033, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [18:10<15:08,  3.88s/it]
 55%|█████▌    | 287/520 [18:14<15:02,  3.87s/it]
                                                 
{'loss': 1.3007, 'grad_norm': 0.0013406057496683974, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [18:14<15:02,  3.87s/it]
 55%|█████▌    | 288/520 [18:18<14:51,  3.84s/it]
                                                 
{'loss': 1.3375, 'grad_norm': 0.001284612086671252, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:18<14:51,  3.84s/it]
 56%|█████▌    | 289/520 [18:22<14:38,  3.80s/it]
                                                 
{'loss': 1.2089, 'grad_norm': 0.0012279415573327253, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:22<14:38,  3.80s/it]
 56%|█████▌    | 290/520 [18:25<14:27,  3.77s/it]
                                                 
{'loss': 1.1327, 'grad_norm': 0.0012485962791396136, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:25<14:27,  3.77s/it]
 56%|█████▌    | 291/520 [18:29<14:16,  3.74s/it]
                                                 
{'loss': 1.1795, 'grad_norm': 0.0012926791630339797, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:29<14:16,  3.74s/it]
 56%|█████▌    | 292/520 [18:33<14:09,  3.72s/it]
                                                 
{'loss': 1.2367, 'grad_norm': 0.0012711628324409581, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:33<14:09,  3.72s/it]
 56%|█████▋    | 293/520 [18:36<14:03,  3.72s/it]
                                                 
{'loss': 1.18, 'grad_norm': 0.0013839985592739412, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:36<14:03,  3.72s/it]
 57%|█████▋    | 294/520 [18:40<13:58,  3.71s/it]
                                                 
{'loss': 1.203, 'grad_norm': 0.0014860356684880684, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:40<13:58,  3.71s/it]
 57%|█████▋    | 295/520 [18:44<13:51,  3.70s/it]
                                                 
{'loss': 1.2334, 'grad_norm': 0.0013582510307069812, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:44<13:51,  3.70s/it]
 57%|█████▋    | 296/520 [18:47<13:47,  3.70s/it]
                                                 
{'loss': 1.1504, 'grad_norm': 0.0013697230071164485, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [18:47<13:47,  3.70s/it]
 57%|█████▋    | 297/520 [18:51<13:41,  3.69s/it]
                                                 
{'loss': 1.2804, 'grad_norm': 0.0013989110868597573, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [18:51<13:41,  3.69s/it]
 57%|█████▋    | 298/520 [18:55<13:37,  3.68s/it]
                                                 
{'loss': 1.2456, 'grad_norm': 0.001230322815018349, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [18:55<13:37,  3.68s/it]
 57%|█████▊    | 299/520 [18:58<13:32,  3.68s/it]
                                                 
{'loss': 1.2607, 'grad_norm': 0.0012100143276307384, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [18:58<13:32,  3.68s/it]
 58%|█████▊    | 300/520 [19:02<13:26,  3.67s/it]
                                                 
{'loss': 1.2918, 'grad_norm': 0.0012684246025927214, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [19:02<13:26,  3.67s/it]
 58%|█████▊    | 301/520 [19:06<13:28,  3.69s/it]
                                                 
{'loss': 1.2705, 'grad_norm': 0.001299823720217312, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [19:06<13:28,  3.69s/it]
 58%|█████▊    | 302/520 [19:09<13:24,  3.69s/it]
                                                 
{'loss': 1.2775, 'grad_norm': 0.0013125093788553838, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [19:09<13:24,  3.69s/it]
 58%|█████▊    | 303/520 [19:13<13:19,  3.68s/it]
                                                 
{'loss': 1.2024, 'grad_norm': 0.001486229117926504, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [19:13<13:19,  3.68s/it]
 58%|█████▊    | 304/520 [19:17<13:16,  3.69s/it]
                                                 
{'loss': 1.1764, 'grad_norm': 0.0013957137284220417, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:17<13:16,  3.69s/it]
 59%|█████▊    | 305/520 [19:21<13:13,  3.69s/it]
                                                 
{'loss': 1.3041, 'grad_norm': 0.0015473914313512028, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:21<13:13,  3.69s/it]
 59%|█████▉    | 306/520 [19:24<13:11,  3.70s/it]
                                                 
{'loss': 1.2503, 'grad_norm': 0.0012861220216650337, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:24<13:11,  3.70s/it]
 59%|█████▉    | 307/520 [19:28<13:10,  3.71s/it]
                                                 
{'loss': 1.1847, 'grad_norm': 0.0011954285018138347, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:28<13:10,  3.71s/it]
 59%|█████▉    | 308/520 [19:32<13:06,  3.71s/it]
                                                 
{'loss': 1.3044, 'grad_norm': 0.0014012287840518207, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:32<13:06,  3.71s/it]
 59%|█████▉    | 309/520 [19:36<13:25,  3.82s/it]
                                                 
{'loss': 1.1912, 'grad_norm': 0.0012405763477497886, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:36<13:25,  3.82s/it]
 60%|█████▉    | 310/520 [19:39<13:15,  3.79s/it]
                                                 
{'loss': 1.1721, 'grad_norm': 0.0013332343361108937, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:40<13:15,  3.79s/it]
 60%|█████▉    | 311/520 [19:43<13:07,  3.77s/it]
                                                 
{'loss': 1.1424, 'grad_norm': 0.0013284794983121945, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [19:43<13:07,  3.77s/it]
 60%|██████    | 312/520 [19:47<12:58,  3.74s/it]
                                                 
{'loss': 1.1337, 'grad_norm': 0.0014248184125004546, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [19:47<12:58,  3.74s/it]
 60%|██████    | 313/520 [19:51<12:53,  3.73s/it]
                                                 
{'loss': 1.123, 'grad_norm': 0.0012031555084594315, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [19:51<12:53,  3.73s/it]
 60%|██████    | 314/520 [19:55<13:15,  3.86s/it]
                                                 
{'loss': 1.1602, 'grad_norm': 0.001179364786070569, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [19:55<13:15,  3.86s/it]
 61%|██████    | 315/520 [19:58<13:02,  3.82s/it]
                                                 
{'loss': 1.2249, 'grad_norm': 0.001680433761483466, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [19:58<13:02,  3.82s/it]
 61%|██████    | 316/520 [20:03<13:18,  3.92s/it]
                                                 
{'loss': 1.1386, 'grad_norm': 0.0014895029674126354, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [20:03<13:18,  3.92s/it]
 61%|██████    | 317/520 [20:06<13:00,  3.85s/it]
                                                 
{'loss': 1.1579, 'grad_norm': 0.0011749842618943216, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [20:06<13:00,  3.85s/it]
 61%|██████    | 318/520 [20:10<12:51,  3.82s/it]
                                                 
{'loss': 1.2666, 'grad_norm': 0.0014436146316392695, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [20:10<12:51,  3.82s/it]
 61%|██████▏   | 319/520 [20:14<13:02,  3.89s/it]
                                                 
{'loss': 1.1395, 'grad_norm': 0.0012069017964304086, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [20:14<13:02,  3.89s/it]
 62%|██████▏   | 320/520 [20:18<12:46,  3.83s/it]
                                                 
{'loss': 1.0858, 'grad_norm': 0.0012957295740000031, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:18<12:46,  3.83s/it]
 62%|██████▏   | 321/520 [20:21<12:32,  3.78s/it]
                                                 
{'loss': 1.284, 'grad_norm': 0.0014031738141183844, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:21<12:32,  3.78s/it]
 62%|██████▏   | 322/520 [20:25<12:23,  3.76s/it]
                                                 
{'loss': 1.1229, 'grad_norm': 0.0012359843081290547, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:25<12:23,  3.76s/it]
 62%|██████▏   | 323/520 [20:29<12:14,  3.73s/it]
                                                 
{'loss': 1.1926, 'grad_norm': 0.0012496225051376304, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:29<12:14,  3.73s/it]
 62%|██████▏   | 324/520 [20:32<12:05,  3.70s/it]
                                                 
{'loss': 1.2193, 'grad_norm': 0.0012947469435586833, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:32<12:05,  3.70s/it]
 62%|██████▎   | 325/520 [20:36<12:01,  3.70s/it]
                                                 
{'loss': 1.228, 'grad_norm': 0.0013717190611794835, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:36<12:01,  3.70s/it]
 63%|██████▎   | 326/520 [20:40<11:57,  3.70s/it]
                                                 
{'loss': 1.2187, 'grad_norm': 0.001319115686372079, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:40<11:57,  3.70s/it]
 63%|██████▎   | 327/520 [20:44<11:53,  3.69s/it]
                                                 
{'loss': 1.2368, 'grad_norm': 0.0013904900846652016, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [20:44<11:53,  3.69s/it]
 63%|██████▎   | 328/520 [20:47<11:47,  3.69s/it]
                                                 
{'loss': 1.2659, 'grad_norm': 0.0013293899449133067, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [20:47<11:47,  3.69s/it]
 63%|██████▎   | 329/520 [20:51<11:44,  3.69s/it]
                                                 
{'loss': 1.1431, 'grad_norm': 0.0011357288275666735, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [20:51<11:44,  3.69s/it]
 63%|██████▎   | 330/520 [20:55<11:41,  3.69s/it]
                                                 
{'loss': 1.2121, 'grad_norm': 0.0011956254290104403, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [20:55<11:41,  3.69s/it]
 64%|██████▎   | 331/520 [20:58<11:37,  3.69s/it]
                                                 
{'loss': 1.1716, 'grad_norm': 0.0012298335464837285, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [20:58<11:37,  3.69s/it]
 64%|██████▍   | 332/520 [21:02<11:34,  3.69s/it]
                                                 
{'loss': 1.262, 'grad_norm': 0.0012258979457178505, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [21:02<11:34,  3.69s/it]
 64%|██████▍   | 333/520 [21:06<11:30,  3.69s/it]
                                                 
{'loss': 1.3146, 'grad_norm': 0.0013378216134409694, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [21:06<11:30,  3.69s/it]
 64%|██████▍   | 334/520 [21:09<11:27,  3.69s/it]
                                                 
{'loss': 1.2241, 'grad_norm': 0.0013509932990793372, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [21:09<11:27,  3.69s/it]
 64%|██████▍   | 335/520 [21:13<11:21,  3.68s/it]
                                                 
{'loss': 1.2239, 'grad_norm': 0.0012040361091208375, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [21:13<11:21,  3.68s/it]
 65%|██████▍   | 336/520 [21:17<11:17,  3.68s/it]
                                                 
{'loss': 1.1208, 'grad_norm': 0.0014019210108078698, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:17<11:17,  3.68s/it]
 65%|██████▍   | 337/520 [21:20<11:13,  3.68s/it]
                                                 
{'loss': 1.103, 'grad_norm': 0.0011891348791630205, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:20<11:13,  3.68s/it]
 65%|██████▌   | 338/520 [21:24<11:10,  3.68s/it]
                                                 
{'loss': 1.2238, 'grad_norm': 0.0012812588668386173, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:24<11:10,  3.68s/it]
 65%|██████▌   | 339/520 [21:28<11:06,  3.68s/it]
                                                 
{'loss': 1.1694, 'grad_norm': 0.0012544796246002343, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:28<11:06,  3.68s/it]
 65%|██████▌   | 340/520 [21:31<11:04,  3.69s/it]
                                                 
{'loss': 1.1624, 'grad_norm': 0.0012579808544879527, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:31<11:04,  3.69s/it]
 66%|██████▌   | 341/520 [21:35<11:01,  3.69s/it]
                                                 
{'loss': 1.1837, 'grad_norm': 0.0013427575539804281, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:35<11:01,  3.69s/it]
 66%|██████▌   | 342/520 [21:39<10:57,  3.69s/it]
                                                 
{'loss': 1.2295, 'grad_norm': 0.001534004458657942, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:39<10:57,  3.69s/it]
 66%|██████▌   | 343/520 [21:43<10:54,  3.70s/it]
                                                 
{'loss': 1.1815, 'grad_norm': 0.0012272143733915702, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [21:43<10:54,  3.70s/it]
 66%|██████▌   | 344/520 [21:46<10:52,  3.71s/it]
                                                 
{'loss': 1.1399, 'grad_norm': 0.0012351135266021073, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [21:46<10:52,  3.71s/it]
 66%|██████▋   | 345/520 [21:50<10:49,  3.71s/it]
                                                 
{'loss': 1.2537, 'grad_norm': 0.0013695178955028305, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [21:50<10:49,  3.71s/it]
 67%|██████▋   | 346/520 [21:54<10:45,  3.71s/it]
                                                 
{'loss': 1.2016, 'grad_norm': 0.0012212693106265014, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [21:54<10:45,  3.71s/it]
 67%|██████▋   | 347/520 [21:57<10:42,  3.72s/it]
                                                 
{'loss': 1.1547, 'grad_norm': 0.00117801381589837, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [21:57<10:42,  3.72s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [22:01<10:39,  3.72s/it]
                                                 
{'loss': 1.1133, 'grad_norm': 0.0014360135915564367, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [22:01<10:39,  3.72s/it]
 67%|██████▋   | 349/520 [22:05<10:34,  3.71s/it]
                                                 
{'loss': 1.1522, 'grad_norm': 0.0012994155782693195, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [22:05<10:34,  3.71s/it]
 67%|██████▋   | 350/520 [22:09<10:30,  3.71s/it]
                                                 
{'loss': 1.1967, 'grad_norm': 0.0013423763634473123, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [22:09<10:30,  3.71s/it]
 68%|██████▊   | 351/520 [22:12<10:27,  3.71s/it]
                                                 
{'loss': 1.1031, 'grad_norm': 0.0011838841406808284, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [22:12<10:27,  3.71s/it]
 68%|██████▊   | 352/520 [22:16<10:31,  3.76s/it]
                                                 
{'loss': 1.2277, 'grad_norm': 0.0012372392905159223, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [22:16<10:31,  3.76s/it]
 68%|██████▊   | 353/520 [22:20<10:33,  3.80s/it]
                                                 
{'loss': 1.1525, 'grad_norm': 0.0010948216996862154, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:20<10:33,  3.80s/it]
 68%|██████▊   | 354/520 [22:24<10:27,  3.78s/it]
                                                 
{'loss': 1.2728, 'grad_norm': 0.001191556922056347, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:24<10:27,  3.78s/it]
 68%|██████▊   | 355/520 [22:28<10:22,  3.77s/it]
                                                 
{'loss': 1.1654, 'grad_norm': 0.0012602211718475496, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:28<10:22,  3.77s/it]
 68%|██████▊   | 356/520 [22:31<10:15,  3.75s/it]
                                                 
{'loss': 1.165, 'grad_norm': 0.0012752932085975913, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:31<10:15,  3.75s/it]
 69%|██████▊   | 357/520 [22:35<10:09,  3.74s/it]
                                                 
{'loss': 1.1959, 'grad_norm': 0.0012542344260865024, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:35<10:09,  3.74s/it]
 69%|██████▉   | 358/520 [22:39<10:04,  3.73s/it]
                                                 
{'loss': 1.1285, 'grad_norm': 0.0012290880548291208, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:39<10:04,  3.73s/it]
 69%|██████▉   | 359/520 [22:42<09:59,  3.72s/it]
                                                 
{'loss': 1.2022, 'grad_norm': 0.0013310152956577503, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [22:42<09:59,  3.72s/it]
 69%|██████▉   | 360/520 [22:46<09:54,  3.71s/it]
                                                 
{'loss': 1.2134, 'grad_norm': 0.0012907094309813911, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [22:46<09:54,  3.71s/it]
 69%|██████▉   | 361/520 [22:50<09:49,  3.71s/it]
                                                 
{'loss': 1.2212, 'grad_norm': 0.0012909539879489007, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [22:50<09:49,  3.71s/it]
 70%|██████▉   | 362/520 [22:53<09:43,  3.69s/it]
                                                 
{'loss': 1.1817, 'grad_norm': 0.0013384163037301379, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [22:53<09:43,  3.69s/it]
 70%|██████▉   | 363/520 [22:57<09:37,  3.68s/it]
                                                 
{'loss': 1.2026, 'grad_norm': 0.0012470877245682435, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [22:57<09:37,  3.68s/it]
 70%|███████   | 364/520 [23:01<09:33,  3.68s/it]
                                                 
{'loss': 1.2441, 'grad_norm': 0.0012394433040807287, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [23:01<09:33,  3.68s/it]
 70%|███████   | 365/520 [23:04<09:29,  3.68s/it]
                                                 
{'loss': 1.2628, 'grad_norm': 0.0013525741300129998, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [23:04<09:29,  3.68s/it]
 70%|███████   | 366/520 [23:08<09:25,  3.67s/it]
                                                 
{'loss': 1.2188, 'grad_norm': 0.0013346351358589303, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [23:08<09:25,  3.67s/it]
 71%|███████   | 367/520 [23:12<09:22,  3.67s/it]
                                                 
{'loss': 1.2198, 'grad_norm': 0.0012639371286629613, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [23:12<09:22,  3.67s/it]
 71%|███████   | 368/520 [23:15<09:18,  3.67s/it]
                                                 
{'loss': 1.0723, 'grad_norm': 0.0013430954648986292, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [23:15<09:18,  3.67s/it]
 71%|███████   | 369/520 [23:19<09:16,  3.69s/it]
                                                 
{'loss': 1.1956, 'grad_norm': 0.0011021963128081725, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:19<09:16,  3.69s/it]
 71%|███████   | 370/520 [23:23<09:11,  3.68s/it]
                                                 
{'loss': 1.1335, 'grad_norm': 0.0011784721986398431, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:23<09:11,  3.68s/it]
 71%|███████▏  | 371/520 [23:26<09:07,  3.67s/it]
                                                 
{'loss': 1.1313, 'grad_norm': 0.0013012045077997823, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:26<09:07,  3.67s/it]
 72%|███████▏  | 372/520 [23:30<09:03,  3.67s/it]
                                                 
{'loss': 1.2722, 'grad_norm': 0.0011819225265918946, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:30<09:03,  3.67s/it]
 72%|███████▏  | 373/520 [23:34<08:58,  3.67s/it]
                                                 
{'loss': 1.162, 'grad_norm': 0.0013315136256614245, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:34<08:58,  3.67s/it]
 72%|███████▏  | 374/520 [23:37<08:54,  3.66s/it]
                                                 
{'loss': 1.2184, 'grad_norm': 0.001257670423882949, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:37<08:54,  3.66s/it]
 72%|███████▏  | 375/520 [23:41<08:50,  3.66s/it]
                                                 
{'loss': 1.1349, 'grad_norm': 0.0012613305186877893, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [23:41<08:50,  3.66s/it]
 72%|███████▏  | 376/520 [23:45<08:47,  3.66s/it]
                                                 
{'loss': 1.2491, 'grad_norm': 0.0012362190853027726, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [23:45<08:47,  3.66s/it]
 72%|███████▎  | 377/520 [23:48<08:42,  3.66s/it]
                                                 
{'loss': 1.1796, 'grad_norm': 0.0012596370387617064, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [23:48<08:42,  3.66s/it]
 73%|███████▎  | 378/520 [23:52<08:40,  3.66s/it]
                                                 
{'loss': 1.2398, 'grad_norm': 0.0012499650990169388, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [23:52<08:40,  3.66s/it]
 73%|███████▎  | 379/520 [23:56<08:37,  3.67s/it]
                                                 
{'loss': 1.2169, 'grad_norm': 0.0011828799329488915, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [23:56<08:37,  3.67s/it]
 73%|███████▎  | 380/520 [23:59<08:33,  3.67s/it]
                                                 
{'loss': 1.2469, 'grad_norm': 0.0012701085541015214, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [23:59<08:33,  3.67s/it]
 73%|███████▎  | 381/520 [24:03<08:29,  3.66s/it]
                                                 
{'loss': 1.2213, 'grad_norm': 0.0012754661346419693, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [24:03<08:29,  3.66s/it]
 73%|███████▎  | 382/520 [24:07<08:28,  3.68s/it]
                                                 
{'loss': 1.2108, 'grad_norm': 0.0012161811606481317, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [24:07<08:28,  3.68s/it]
 74%|███████▎  | 383/520 [24:10<08:23,  3.67s/it]
                                                 
{'loss': 1.0563, 'grad_norm': 0.001363916535397635, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [24:10<08:23,  3.67s/it]
 74%|███████▍  | 384/520 [24:14<08:19,  3.67s/it]
                                                 
{'loss': 1.2562, 'grad_norm': 0.0011873916059231002, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [24:14<08:19,  3.67s/it]
 74%|███████▍  | 385/520 [24:18<08:16,  3.68s/it]
                                                 
{'loss': 1.1947, 'grad_norm': 0.0011542567338081583, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:18<08:16,  3.68s/it]
 74%|███████▍  | 386/520 [24:22<08:14,  3.69s/it]
                                                 
{'loss': 1.1481, 'grad_norm': 0.0010801814958289022, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:22<08:14,  3.69s/it]
 74%|███████▍  | 387/520 [24:25<08:10,  3.69s/it]
                                                 
{'loss': 1.2713, 'grad_norm': 0.0012411430805209732, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:25<08:10,  3.69s/it]
 75%|███████▍  | 388/520 [24:29<08:06,  3.68s/it]
                                                 
{'loss': 1.1025, 'grad_norm': 0.0011982750676810259, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:29<08:06,  3.68s/it]
 75%|███████▍  | 389/520 [24:33<08:01,  3.68s/it]
                                                 
{'loss': 1.1533, 'grad_norm': 0.001382796346752328, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}
+
 75%|███████▍  | 389/520 [24:33<08:01,  3.68s/it]
 75%|███████▌  | 390/520 [24:36<07:58,  3.68s/it]
                                                 
{'loss': 1.2161, 'grad_norm': 0.0011719574390307047, 'learning_rate': 0.031074166151605298, 'epoch': 0.75}
+
 75%|███████▌  | 390/520 [24:36<07:58,  3.68s/it]
 75%|███████▌  | 391/520 [24:40<07:54,  3.68s/it]
                                                 
{'loss': 1.2918, 'grad_norm': 0.0012776277210315306, 'learning_rate': 0.030623894397145836, 'epoch': 0.75}
+
 75%|███████▌  | 391/520 [24:40<07:54,  3.68s/it]
 75%|███████▌  | 392/520 [24:44<07:49,  3.67s/it]
                                                 
{'loss': 1.1062, 'grad_norm': 0.0012466092634341214, 'learning_rate': 0.03017631819139273, 'epoch': 0.75}
+
 75%|███████▌  | 392/520 [24:44<07:49,  3.67s/it]
 76%|███████▌  | 393/520 [24:47<07:46,  3.67s/it]
                                                 
{'loss': 1.1099, 'grad_norm': 0.001086022200015021, 'learning_rate': 0.029731454924533086, 'epoch': 0.76}
+
 76%|███████▌  | 393/520 [24:47<07:46,  3.67s/it]
 76%|███████▌  | 394/520 [24:51<07:42,  3.67s/it]
                                                 
{'loss': 1.1671, 'grad_norm': 0.0013657687294806658, 'learning_rate': 0.029289321881345254, 'epoch': 0.76}
+
 76%|███████▌  | 394/520 [24:51<07:42,  3.67s/it]
 76%|███████▌  | 395/520 [24:55<07:38,  3.67s/it]
                                                 
{'loss': 1.1365, 'grad_norm': 0.0013485322745472896, 'learning_rate': 0.028849936240527008, 'epoch': 0.76}
+
 76%|███████▌  | 395/520 [24:55<07:38,  3.67s/it]
 76%|███████▌  | 396/520 [24:58<07:34,  3.67s/it]
                                                 
{'loss': 1.218, 'grad_norm': 0.0014100107938836409, 'learning_rate': 0.028413315074028157, 'epoch': 0.76}
+
 76%|███████▌  | 396/520 [24:58<07:34,  3.67s/it]
 76%|███████▋  | 397/520 [25:02<07:32,  3.68s/it]
                                                 
{'loss': 1.1943, 'grad_norm': 0.0011859711586483703, 'learning_rate': 0.027979475346387363, 'epoch': 0.76}
+
 76%|███████▋  | 397/520 [25:02<07:32,  3.68s/it]
 77%|███████▋  | 398/520 [25:06<07:28,  3.68s/it]
                                                 
{'loss': 1.1983, 'grad_norm': 0.0012996355590739065, 'learning_rate': 0.027548433914072735, 'epoch': 0.77}
+
 77%|███████▋  | 398/520 [25:06<07:28,  3.68s/it]
 77%|███████▋  | 399/520 [25:09<07:25,  3.68s/it]
                                                 
{'loss': 1.1542, 'grad_norm': 0.0012127022823063321, 'learning_rate': 0.027120207524827168, 'epoch': 0.77}
+
 77%|███████▋  | 399/520 [25:09<07:25,  3.68s/it]
 77%|███████▋  | 400/520 [25:13<07:22,  3.69s/it]
                                                 
{'loss': 1.1878, 'grad_norm': 0.0011386042970706598, 'learning_rate': 0.02669481281701739, 'epoch': 0.77}
+
 77%|███████▋  | 400/520 [25:13<07:22,  3.69s/it]
 77%|███████▋  | 401/520 [25:17<07:17,  3.68s/it]
                                                 
{'loss': 1.0339, 'grad_norm': 0.001468111954212263, 'learning_rate': 0.026272266318987603, 'epoch': 0.77}
+
 77%|███████▋  | 401/520 [25:17<07:17,  3.68s/it]
 77%|███████▋  | 402/520 [25:20<07:13,  3.67s/it]
                                                 
{'loss': 1.1519, 'grad_norm': 0.0012493983119957871, 'learning_rate': 0.02585258444841733, 'epoch': 0.77}
+
 77%|███████▋  | 402/520 [25:20<07:13,  3.67s/it]
 78%|███████▊  | 403/520 [25:24<07:12,  3.70s/it]
                                                 
{'loss': 1.1774, 'grad_norm': 0.001354580715614851, 'learning_rate': 0.025435783511683442, 'epoch': 0.78}
+
 78%|███████▊  | 403/520 [25:24<07:12,  3.70s/it]
 78%|███████▊  | 404/520 [25:28<07:13,  3.74s/it]
                                                 
{'loss': 1.0839, 'grad_norm': 0.0014879782694579482, 'learning_rate': 0.02502187970322657, 'epoch': 0.78}
+
 78%|███████▊  | 404/520 [25:28<07:13,  3.74s/it]
 78%|███████▊  | 405/520 [25:32<07:15,  3.78s/it]
                                                 
{'loss': 1.1629, 'grad_norm': 0.0012525797305866096, 'learning_rate': 0.02461088910492202, 'epoch': 0.78}
+
 78%|███████▊  | 405/520 [25:32<07:15,  3.78s/it]
 78%|███████▊  | 406/520 [25:36<07:13,  3.80s/it]
                                                 
{'loss': 1.0809, 'grad_norm': 0.0015251299578845506, 'learning_rate': 0.02420282768545469, 'epoch': 0.78}
+
 78%|███████▊  | 406/520 [25:36<07:13,  3.80s/it]
 78%|███████▊  | 407/520 [25:40<07:11,  3.82s/it]
                                                 
{'loss': 1.2622, 'grad_norm': 0.0012575441812063602, 'learning_rate': 0.02379771129969892, 'epoch': 0.78}
+
 78%|███████▊  | 407/520 [25:40<07:11,  3.82s/it]
 78%|███████▊  | 408/520 [25:43<07:08,  3.83s/it]
                                                 
{'loss': 1.1678, 'grad_norm': 0.001377527201757681, 'learning_rate': 0.023395555688102213, 'epoch': 0.78}
+
 78%|███████▊  | 408/520 [25:43<07:08,  3.83s/it]
 79%|███████▊  | 409/520 [25:47<07:05,  3.84s/it]
                                                 
{'loss': 1.2864, 'grad_norm': 0.0013451903077860068, 'learning_rate': 0.02299637647607372, 'epoch': 0.79}
+
 79%|███████▊  | 409/520 [25:47<07:05,  3.84s/it]
 79%|███████▉  | 410/520 [25:51<07:03,  3.85s/it]
                                                 
{'loss': 1.0177, 'grad_norm': 0.0012619785091852775, 'learning_rate': 0.022600189173377264, 'epoch': 0.79}
+
 79%|███████▉  | 410/520 [25:51<07:03,  3.85s/it]
 79%|███████▉  | 411/520 [25:55<07:00,  3.86s/it]
                                                 
{'loss': 1.2645, 'grad_norm': 0.001527573780960736, 'learning_rate': 0.022207009173528525, 'epoch': 0.79}
+
 79%|███████▉  | 411/520 [25:55<07:00,  3.86s/it]
 79%|███████▉  | 412/520 [25:59<06:57,  3.86s/it]
                                                 
{'loss': 1.1753, 'grad_norm': 0.0012577213425109413, 'learning_rate': 0.02181685175319702, 'epoch': 0.79}
+
 79%|███████▉  | 412/520 [25:59<06:57,  3.86s/it]
 79%|███████▉  | 413/520 [26:03<06:53,  3.87s/it]
                                                 
{'loss': 1.176, 'grad_norm': 0.0011482692406493756, 'learning_rate': 0.021429732071612653, 'epoch': 0.79}
+
 79%|███████▉  | 413/520 [26:03<06:53,  3.87s/it]
 80%|███████▉  | 414/520 [26:07<06:47,  3.85s/it]
                                                 
{'loss': 0.9888, 'grad_norm': 0.0010336007261086894, 'learning_rate': 0.02104566516997647, 'epoch': 0.8}
+
 80%|███████▉  | 414/520 [26:07<06:47,  3.85s/it]
 80%|███████▉  | 415/520 [26:10<06:38,  3.80s/it]
                                                 
{'loss': 1.1517, 'grad_norm': 0.0011666842748548928, 'learning_rate': 0.020664665970876496, 'epoch': 0.8}
+
 80%|███████▉  | 415/520 [26:10<06:38,  3.80s/it]
 80%|████████  | 416/520 [26:14<06:31,  3.76s/it]
                                                 
{'loss': 1.07, 'grad_norm': 0.001333175857301277, 'learning_rate': 0.020286749277707784, 'epoch': 0.8}
+
 80%|████████  | 416/520 [26:14<06:31,  3.76s/it]
 80%|████████  | 417/520 [26:18<06:25,  3.74s/it]
                                                 
{'loss': 1.2289, 'grad_norm': 0.0013537291677059619, 'learning_rate': 0.019911929774097215, 'epoch': 0.8}
+
 80%|████████  | 417/520 [26:18<06:25,  3.74s/it]
 80%|████████  | 418/520 [26:21<06:20,  3.73s/it]
                                                 
{'loss': 1.2185, 'grad_norm': 0.0012434001191357724, 'learning_rate': 0.019540222023333165, 'epoch': 0.8}
+
 80%|████████  | 418/520 [26:21<06:20,  3.73s/it]
 81%|████████  | 419/520 [26:25<06:15,  3.72s/it]
                                                 
{'loss': 1.2091, 'grad_norm': 0.0013918715821619834, 'learning_rate': 0.01917164046779948, 'epoch': 0.81}
+
 81%|████████  | 419/520 [26:25<06:15,  3.72s/it]
 81%|████████  | 420/520 [26:29<06:12,  3.72s/it]
                                                 
{'loss': 1.0999, 'grad_norm': 0.0013349412154291373, 'learning_rate': 0.018806199428414352, 'epoch': 0.81}
+
 81%|████████  | 420/520 [26:29<06:12,  3.72s/it]
 81%|████████  | 421/520 [26:32<06:09,  3.73s/it]
                                                 
{'loss': 1.0384, 'grad_norm': 0.0015797626553213917, 'learning_rate': 0.018443913104073985, 'epoch': 0.81}
+
 81%|████████  | 421/520 [26:32<06:09,  3.73s/it]
 81%|████████  | 422/520 [26:36<06:05,  3.73s/it]
                                                 
{'loss': 1.1558, 'grad_norm': 0.0013114022926348336, 'learning_rate': 0.01808479557110081, 'epoch': 0.81}
+
 81%|████████  | 422/520 [26:36<06:05,  3.73s/it]
 81%|████████▏ | 423/520 [26:40<06:01,  3.73s/it]
                                                 
{'loss': 1.136, 'grad_norm': 0.0014227077014815556, 'learning_rate': 0.017728860782696667, 'epoch': 0.81}
+
 81%|████████▏ | 423/520 [26:40<06:01,  3.73s/it]
 82%|████████▏ | 424/520 [26:44<05:58,  3.73s/it]
                                                 
{'loss': 1.2622, 'grad_norm': 0.0013622047801009215, 'learning_rate': 0.017376122568400532, 'epoch': 0.82}
+
 82%|████████▏ | 424/520 [26:44<05:58,  3.73s/it]
 82%|████████▏ | 425/520 [26:47<05:54,  3.73s/it]
                                                 
{'loss': 1.1518, 'grad_norm': 0.0012523471899006599, 'learning_rate': 0.017026594633551252, 'epoch': 0.82}
+
 82%|████████▏ | 425/520 [26:47<05:54,  3.73s/it]
 82%|████████▏ | 426/520 [26:51<05:50,  3.73s/it]
                                                 
{'loss': 1.1664, 'grad_norm': 0.0015569625478068527, 'learning_rate': 0.01668029055875512, 'epoch': 0.82}
+
 82%|████████▏ | 426/520 [26:51<05:50,  3.73s/it]
 82%|████████▏ | 427/520 [26:55<05:45,  3.72s/it]
                                                 
{'loss': 1.083, 'grad_norm': 0.0011972870604732863, 'learning_rate': 0.016337223799358026, 'epoch': 0.82}
+
 82%|████████▏ | 427/520 [26:55<05:45,  3.72s/it]
 82%|████████▏ | 428/520 [26:59<05:42,  3.72s/it]
                                                 
{'loss': 1.0636, 'grad_norm': 0.001280773019535236, 'learning_rate': 0.01599740768492286, 'epoch': 0.82}
+
 82%|████████▏ | 428/520 [26:59<05:42,  3.72s/it]
 82%|████████▎ | 429/520 [27:02<05:40,  3.75s/it]
                                                 
{'loss': 1.1585, 'grad_norm': 0.0012676213101164842, 'learning_rate': 0.015660855418711452, 'epoch': 0.82}
+
 82%|████████▎ | 429/520 [27:02<05:40,  3.75s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2076 > 2048). Running this sequence through the model will result in indexing errors
+
 83%|████████▎ | 430/520 [27:06<05:40,  3.78s/it]
                                                 
{'loss': 1.1629, 'grad_norm': 0.0011693476750943938, 'learning_rate': 0.015327580077171589, 'epoch': 0.83}
+
 83%|████████▎ | 430/520 [27:06<05:40,  3.78s/it]
 83%|████████▎ | 431/520 [27:10<05:35,  3.77s/it]
                                                 
{'loss': 1.1459, 'grad_norm': 0.0013210127971430158, 'learning_rate': 0.014997594609429088, 'epoch': 0.83}
+
 83%|████████▎ | 431/520 [27:10<05:35,  3.77s/it]
 83%|████████▎ | 432/520 [27:14<05:30,  3.75s/it]
                                                 
{'loss': 1.071, 'grad_norm': 0.0012995891826446122, 'learning_rate': 0.01467091183678444, 'epoch': 0.83}
+
 83%|████████▎ | 432/520 [27:14<05:30,  3.75s/it]
 83%|████████▎ | 433/520 [27:17<05:25,  3.74s/it]
                                                 
{'loss': 1.201, 'grad_norm': 0.001248860753872551, 'learning_rate': 0.014347544452214867, 'epoch': 0.83}
+
 83%|████████▎ | 433/520 [27:17<05:25,  3.74s/it]
 83%|████████▎ | 434/520 [27:21<05:20,  3.73s/it]
                                                 
{'loss': 0.955, 'grad_norm': 0.0012786377286556326, 'learning_rate': 0.014027505019880971, 'epoch': 0.83}
+
 83%|████████▎ | 434/520 [27:21<05:20,  3.73s/it]
 84%|████████▎ | 435/520 [27:25<05:15,  3.71s/it]
                                                 
{'loss': 1.2416, 'grad_norm': 0.0014105783318120488, 'learning_rate': 0.013710805974638696, 'epoch': 0.84}
+
 84%|████████▎ | 435/520 [27:25<05:15,  3.71s/it]
 84%|████████▍ | 436/520 [27:28<05:11,  3.71s/it]
                                                 
{'loss': 1.0416, 'grad_norm': 0.001330237717288208, 'learning_rate': 0.01339745962155613, 'epoch': 0.84}
+
 84%|████████▍ | 436/520 [27:28<05:11,  3.71s/it]
 84%|████████▍ | 437/520 [27:32<05:07,  3.71s/it]
                                                 
{'loss': 1.2607, 'grad_norm': 0.0012687509783257747, 'learning_rate': 0.01308747813543536, 'epoch': 0.84}
+
 84%|████████▍ | 437/520 [27:32<05:07,  3.71s/it]
 84%|████████▍ | 438/520 [27:36<05:03,  3.71s/it]
                                                 
{'loss': 1.0802, 'grad_norm': 0.0012059030615837105, 'learning_rate': 0.012780873560339467, 'epoch': 0.84}
+
 84%|████████▍ | 438/520 [27:36<05:03,  3.71s/it]
 84%|████████▍ | 439/520 [27:40<05:01,  3.72s/it]
                                                 
{'loss': 1.1301, 'grad_norm': 0.0010533291555289048, 'learning_rate': 0.012477657809124632, 'epoch': 0.84}
+
 84%|████████▍ | 439/520 [27:40<05:01,  3.72s/it]
 85%|████████▍ | 440/520 [27:43<04:58,  3.74s/it]
                                                 
{'loss': 1.1108, 'grad_norm': 0.001244535618688038, 'learning_rate': 0.012177842662977134, 'epoch': 0.85}
+
 85%|████████▍ | 440/520 [27:43<04:58,  3.74s/it]
 85%|████████▍ | 441/520 [27:47<04:55,  3.74s/it]
                                                 
{'loss': 1.1437, 'grad_norm': 0.001179753836725469, 'learning_rate': 0.01188143977095576, 'epoch': 0.85}
+
 85%|████████▍ | 441/520 [27:47<04:55,  3.74s/it]
 85%|████████▌ | 442/520 [27:51<04:49,  3.72s/it]
                                                 
{'loss': 1.178, 'grad_norm': 0.0013909196600555666, 'learning_rate': 0.011588460649539035, 'epoch': 0.85}
+
 85%|████████▌ | 442/520 [27:51<04:49,  3.72s/it]
 85%|████████▌ | 443/520 [27:55<04:46,  3.72s/it]
                                                 
{'loss': 1.1923, 'grad_norm': 0.0012421654630707076, 'learning_rate': 0.011298916682177829, 'epoch': 0.85}
+
 85%|████████▌ | 443/520 [27:55<04:46,  3.72s/it]
 85%|████████▌ | 444/520 [27:58<04:45,  3.76s/it]
                                                 
{'loss': 1.1554, 'grad_norm': 0.0011601938684385367, 'learning_rate': 0.011012819118853146, 'epoch': 0.85}
+
 85%|████████▌ | 444/520 [27:58<04:45,  3.76s/it]
 86%|████████▌ | 445/520 [28:02<04:42,  3.77s/it]
                                                 
{'loss': 1.0852, 'grad_norm': 0.0012295457027392638, 'learning_rate': 0.01073017907563887, 'epoch': 0.86}
+
 86%|████████▌ | 445/520 [28:02<04:42,  3.77s/it]
 86%|████████▌ | 446/520 [28:06<04:41,  3.80s/it]
                                                 
{'loss': 1.223, 'grad_norm': 0.0012186631876909104, 'learning_rate': 0.010451007534269908, 'epoch': 0.86}
+
 86%|████████▌ | 446/520 [28:06<04:41,  3.80s/it]
 86%|████████▌ | 447/520 [28:10<04:37,  3.80s/it]
                                                 
{'loss': 1.1665, 'grad_norm': 0.0012571932904021275, 'learning_rate': 0.010175315341715598, 'epoch': 0.86}
+
 86%|████████▌ | 447/520 [28:10<04:37,  3.80s/it]
 86%|████████▌ | 448/520 [28:14<04:34,  3.81s/it]
                                                 
{'loss': 1.1543, 'grad_norm': 0.001300304936698145, 'learning_rate': 0.009903113209758098, 'epoch': 0.86}
+
 86%|████████▌ | 448/520 [28:14<04:34,  3.81s/it]
 86%|████████▋ | 449/520 [28:18<04:31,  3.82s/it]
                                                 
{'loss': 1.179, 'grad_norm': 0.0012661574193465453, 'learning_rate': 0.009634411714576352, 'epoch': 0.86}
+
 86%|████████▋ | 449/520 [28:18<04:31,  3.82s/it]
 87%|████████▋ | 450/520 [28:21<04:28,  3.83s/it]
                                                 
{'loss': 1.1847, 'grad_norm': 0.0012672252236559346, 'learning_rate': 0.009369221296335007, 'epoch': 0.87}
+
 87%|████████▋ | 450/520 [28:21<04:28,  3.83s/it]
 87%|████████▋ | 451/520 [28:25<04:23,  3.82s/it]
                                                 
{'loss': 1.1796, 'grad_norm': 0.001254481378311575, 'learning_rate': 0.009107552258778906, 'epoch': 0.87}
+
 87%|████████▋ | 451/520 [28:25<04:23,  3.82s/it]
 87%|████████▋ | 452/520 [28:29<04:19,  3.82s/it]
                                                 
{'loss': 1.2198, 'grad_norm': 0.001154009003121641, 'learning_rate': 0.008849414768832687, 'epoch': 0.87}
+
 87%|████████▋ | 452/520 [28:29<04:19,  3.82s/it]
 87%|████████▋ | 453/520 [28:33<04:15,  3.82s/it]
                                                 
{'loss': 1.1935, 'grad_norm': 0.0012373262272623883, 'learning_rate': 0.008594818856205699, 'epoch': 0.87}
+
 87%|████████▋ | 453/520 [28:33<04:15,  3.82s/it]
 87%|████████▋ | 454/520 [28:37<04:13,  3.83s/it]
                                                 
{'loss': 1.0905, 'grad_norm': 0.0013072270521799546, 'learning_rate': 0.00834377441300238, 'epoch': 0.87}
+
 87%|████████▋ | 454/520 [28:37<04:13,  3.83s/it]
 88%|████████▊ | 455/520 [28:41<04:09,  3.84s/it]
                                                 
{'loss': 1.2298, 'grad_norm': 0.0012256900370289311, 'learning_rate': 0.008096291193337934, 'epoch': 0.88}
+
 88%|████████▊ | 455/520 [28:41<04:09,  3.84s/it]
 88%|████████▊ | 456/520 [28:44<04:05,  3.84s/it]
                                                 
{'loss': 1.1529, 'grad_norm': 0.0012692367755899093, 'learning_rate': 0.007852378812959226, 'epoch': 0.88}
+
 88%|████████▊ | 456/520 [28:44<04:05,  3.84s/it]
 88%|████████▊ | 457/520 [28:48<04:01,  3.83s/it]
                                                 
{'loss': 1.1072, 'grad_norm': 0.0010950307773607477, 'learning_rate': 0.007612046748871327, 'epoch': 0.88}
+
 88%|████████▊ | 457/520 [28:48<04:01,  3.83s/it]
 88%|████████▊ | 458/520 [28:52<03:55,  3.80s/it]
                                                 
{'loss': 1.2859, 'grad_norm': 0.0013415037214523032, 'learning_rate': 0.007375304338969136, 'epoch': 0.88}
+
 88%|████████▊ | 458/520 [28:52<03:55,  3.80s/it]
 88%|████████▊ | 459/520 [28:56<03:49,  3.77s/it]
                                                 
{'loss': 1.2146, 'grad_norm': 0.0014036231019542015, 'learning_rate': 0.007142160781674645, 'epoch': 0.88}
+
 88%|████████▊ | 459/520 [28:56<03:49,  3.77s/it]
 88%|████████▊ | 460/520 [28:59<03:45,  3.75s/it]
                                                 
{'loss': 1.1042, 'grad_norm': 0.0012706742862650146, 'learning_rate': 0.006912625135579587, 'epoch': 0.88}
+
 88%|████████▊ | 460/520 [28:59<03:45,  3.75s/it]
 89%|████████▊ | 461/520 [29:03<03:40,  3.74s/it]
                                                 
{'loss': 1.1891, 'grad_norm': 0.000979940445892709, 'learning_rate': 0.0066867063190933496, 'epoch': 0.89}
+
 89%|████████▊ | 461/520 [29:03<03:40,  3.74s/it]
 89%|████████▉ | 462/520 [29:07<03:35,  3.72s/it]
                                                 
{'loss': 1.2665, 'grad_norm': 0.0011898601148040587, 'learning_rate': 0.006464413110096601, 'epoch': 0.89}
+
 89%|████████▉ | 462/520 [29:07<03:35,  3.72s/it]
 89%|████████▉ | 463/520 [29:10<03:31,  3.71s/it]
                                                 
{'loss': 1.0601, 'grad_norm': 0.0012901654994532151, 'learning_rate': 0.006245754145600091, 'epoch': 0.89}
+
 89%|████████▉ | 463/520 [29:10<03:31,  3.71s/it]
 89%|████████▉ | 464/520 [29:14<03:27,  3.71s/it]
                                                 
{'loss': 1.1994, 'grad_norm': 0.0013660093336913882, 'learning_rate': 0.006030737921409169, 'epoch': 0.89}
+
 89%|████████▉ | 464/520 [29:14<03:27,  3.71s/it]
 89%|████████▉ | 465/520 [29:18<03:24,  3.72s/it]
                                                 
{'loss': 1.3031, 'grad_norm': 0.0013407717723476114, 'learning_rate': 0.005819372791793654, 'epoch': 0.89}
+
 89%|████████▉ | 465/520 [29:18<03:24,  3.72s/it]
 90%|████████▉ | 466/520 [29:22<03:20,  3.72s/it]
                                                 
{'loss': 1.1886, 'grad_norm': 0.0011693331651684944, 'learning_rate': 0.005611666969163243, 'epoch': 0.9}
+
 90%|████████▉ | 466/520 [29:22<03:20,  3.72s/it]
 90%|████████▉ | 467/520 [29:25<03:16,  3.72s/it]
                                                 
{'loss': 1.1571, 'grad_norm': 0.0011311445439010455, 'learning_rate': 0.005407628523748398, 'epoch': 0.9}
+
 90%|████████▉ | 467/520 [29:25<03:16,  3.72s/it]
 90%|█████████ | 468/520 [29:29<03:12,  3.70s/it]
                                                 
{'loss': 1.1629, 'grad_norm': 0.0014096597081296942, 'learning_rate': 0.00520726538328683, 'epoch': 0.9}
+
 90%|█████████ | 468/520 [29:29<03:12,  3.70s/it]
 90%|█████████ | 469/520 [29:33<03:10,  3.74s/it]
                                                 
{'loss': 1.2242, 'grad_norm': 0.0013777444651659056, 'learning_rate': 0.005010585332715401, 'epoch': 0.9}
+
 90%|█████████ | 469/520 [29:33<03:10,  3.74s/it]
 90%|█████████ | 470/520 [29:37<03:08,  3.78s/it]
                                                 
{'loss': 1.1034, 'grad_norm': 0.0011489831938634814, 'learning_rate': 0.004817596013867765, 'epoch': 0.9}
+
 90%|█████████ | 470/520 [29:37<03:08,  3.78s/it]
 91%|█████████ | 471/520 [29:40<03:05,  3.79s/it]
                                                 
{'loss': 1.1288, 'grad_norm': 0.0013021109931168457, 'learning_rate': 0.004628304925177318, 'epoch': 0.91}
+
 91%|█████████ | 471/520 [29:40<03:05,  3.79s/it]
 91%|█████████ | 472/520 [29:44<03:03,  3.83s/it]
                                                 
{'loss': 1.0918, 'grad_norm': 0.0013030235155294867, 'learning_rate': 0.004442719421385921, 'epoch': 0.91}
+
 91%|█████████ | 472/520 [29:44<03:03,  3.83s/it]
 91%|█████████ | 473/520 [29:48<02:59,  3.83s/it]
                                                 
{'loss': 1.1598, 'grad_norm': 0.0013275857682847774, 'learning_rate': 0.004260846713258193, 'epoch': 0.91}
+
 91%|█████████ | 473/520 [29:48<02:59,  3.83s/it]
 91%|█████████ | 474/520 [29:52<02:56,  3.83s/it]
                                                 
{'loss': 1.1925, 'grad_norm': 0.0011680565503060143, 'learning_rate': 0.004082693867301224, 'epoch': 0.91}
+
 91%|█████████ | 474/520 [29:52<02:56,  3.83s/it]
 91%|█████████▏| 475/520 [29:56<02:52,  3.84s/it]
                                                 
{'loss': 1.1125, 'grad_norm': 0.0012002870298428604, 'learning_rate': 0.003908267805490051, 'epoch': 0.91}
+
 91%|█████████▏| 475/520 [29:56<02:52,  3.84s/it]
 92%|█████████▏| 476/520 [30:00<02:48,  3.82s/it]
                                                 
{'loss': 1.1476, 'grad_norm': 0.0012873596939828782, 'learning_rate': 0.003737575304998797, 'epoch': 0.92}
+
 92%|█████████▏| 476/520 [30:00<02:48,  3.82s/it]
 92%|█████████▏| 477/520 [30:04<02:44,  3.83s/it]
                                                 
{'loss': 1.1382, 'grad_norm': 0.0013825130945376992, 'learning_rate': 0.003570622997937234, 'epoch': 0.92}
+
 92%|█████████▏| 477/520 [30:04<02:44,  3.83s/it]
 92%|█████████▏| 478/520 [30:07<02:40,  3.83s/it]
                                                 
{'loss': 1.0901, 'grad_norm': 0.0012857752777829293, 'learning_rate': 0.00340741737109318, 'epoch': 0.92}
+
 92%|█████████▏| 478/520 [30:07<02:40,  3.83s/it]
 92%|█████████▏| 479/520 [30:11<02:37,  3.84s/it]
                                                 
{'loss': 1.1562, 'grad_norm': 0.0013819713751605362, 'learning_rate': 0.003247964765680389, 'epoch': 0.92}
+
 92%|█████████▏| 479/520 [30:11<02:37,  3.84s/it]
 92%|█████████▏| 480/520 [30:15<02:33,  3.84s/it]
                                                 
{'loss': 1.1756, 'grad_norm': 0.0011252779912811351, 'learning_rate': 0.0030922713770922153, 'epoch': 0.92}
+
 92%|█████████▏| 480/520 [30:15<02:33,  3.84s/it]
 92%|█████████▎| 481/520 [30:19<02:28,  3.80s/it]
                                                 
{'loss': 1.1665, 'grad_norm': 0.0011580597826977245, 'learning_rate': 0.0029403432546609046, 'epoch': 0.93}
+
 92%|█████████▎| 481/520 [30:19<02:28,  3.80s/it]
 93%|█████████▎| 482/520 [30:22<02:22,  3.76s/it]
                                                 
{'loss': 1.1878, 'grad_norm': 0.0012445024517673809, 'learning_rate': 0.0027921863014225504, 'epoch': 0.93}
+
 93%|█████████▎| 482/520 [30:22<02:22,  3.76s/it]
 93%|█████████▎| 483/520 [30:26<02:18,  3.74s/it]
                                                 
{'loss': 1.1584, 'grad_norm': 0.0013726511429586667, 'learning_rate': 0.002647806273887665, 'epoch': 0.93}
+
 93%|█████████▎| 483/520 [30:26<02:18,  3.74s/it]
 93%|█████████▎| 484/520 [30:30<02:13,  3.72s/it]
                                                 
{'loss': 1.1641, 'grad_norm': 0.001298271202485989, 'learning_rate': 0.0025072087818176383, 'epoch': 0.93}
+
 93%|█████████▎| 484/520 [30:30<02:13,  3.72s/it]
 93%|█████████▎| 485/520 [30:33<02:09,  3.70s/it]
                                                 
{'loss': 1.1213, 'grad_norm': 0.0012358305517831665, 'learning_rate': 0.002370399288006664, 'epoch': 0.93}
+
 93%|█████████▎| 485/520 [30:33<02:09,  3.70s/it]
 93%|█████████▎| 486/520 [30:37<02:05,  3.69s/it]
                                                 
{'loss': 1.2415, 'grad_norm': 0.0013418179819873926, 'learning_rate': 0.0022373831080695463, 'epoch': 0.93}
+
 93%|█████████▎| 486/520 [30:37<02:05,  3.69s/it]
 94%|█████████▎| 487/520 [30:41<02:02,  3.70s/it]
                                                 
{'loss': 1.0915, 'grad_norm': 0.0012709075817366005, 'learning_rate': 0.0021081654102351635, 'epoch': 0.94}
+
 94%|█████████▎| 487/520 [30:41<02:02,  3.70s/it]
 94%|█████████▍| 488/520 [30:45<01:59,  3.73s/it]
                                                 
{'loss': 1.0393, 'grad_norm': 0.001334676752973535, 'learning_rate': 0.0019827512151456175, 'epoch': 0.94}
+
 94%|█████████▍| 488/520 [30:45<01:59,  3.73s/it]
 94%|█████████▍| 489/520 [30:48<01:56,  3.75s/it]
                                                 
{'loss': 1.1893, 'grad_norm': 0.0010550780085369635, 'learning_rate': 0.0018611453956612345, 'epoch': 0.94}
+
 94%|█████████▍| 489/520 [30:48<01:56,  3.75s/it]
 94%|█████████▍| 490/520 [30:52<01:53,  3.77s/it]
                                                 
{'loss': 1.1574, 'grad_norm': 0.001280727854279382, 'learning_rate': 0.0017433526766711727, 'epoch': 0.94}
+
 94%|█████████▍| 490/520 [30:52<01:53,  3.77s/it]
 94%|█████████▍| 491/520 [30:56<01:49,  3.78s/it]
                                                 
{'loss': 1.1232, 'grad_norm': 0.0013046965357525156, 'learning_rate': 0.0016293776349098677, 'epoch': 0.94}
+
 94%|█████████▍| 491/520 [30:56<01:49,  3.78s/it]
 95%|█████████▍| 492/520 [31:00<01:45,  3.77s/it]
                                                 
{'loss': 1.2363, 'grad_norm': 0.0013527797558919056, 'learning_rate': 0.0015192246987791981, 'epoch': 0.95}
+
 95%|█████████▍| 492/520 [31:00<01:45,  3.77s/it]
 95%|█████████▍| 493/520 [31:04<01:42,  3.78s/it]
                                                 
{'loss': 1.1893, 'grad_norm': 0.001267658955028913, 'learning_rate': 0.0014128981481764114, 'epoch': 0.95}
+
 95%|█████████▍| 493/520 [31:04<01:42,  3.78s/it]
 95%|█████████▌| 494/520 [31:07<01:39,  3.81s/it]
                                                 
{'loss': 1.1755, 'grad_norm': 0.0011619640073268953, 'learning_rate': 0.0013104021143278911, 'epoch': 0.95}
+
 95%|█████████▌| 494/520 [31:07<01:39,  3.81s/it]
 95%|█████████▌| 495/520 [31:11<01:35,  3.81s/it]
                                                 
{'loss': 1.1427, 'grad_norm': 0.0012546468384193486, 'learning_rate': 0.0012117405796285285, 'epoch': 0.95}
+
 95%|█████████▌| 495/520 [31:11<01:35,  3.81s/it]
 95%|█████████▌| 496/520 [31:15<01:31,  3.82s/it]
                                                 
{'loss': 1.061, 'grad_norm': 0.0012532978418817245, 'learning_rate': 0.0011169173774871477, 'epoch': 0.95}
+
 95%|█████████▌| 496/520 [31:15<01:31,  3.82s/it]
 96%|█████████▌| 497/520 [31:19<01:28,  3.83s/it]
                                                 
{'loss': 1.1194, 'grad_norm': 0.0011320743264580302, 'learning_rate': 0.0010259361921774012, 'epoch': 0.96}
+
 96%|█████████▌| 497/520 [31:19<01:28,  3.83s/it]
 96%|█████████▌| 498/520 [31:23<01:23,  3.79s/it]
                                                 
{'loss': 1.137, 'grad_norm': 0.0012944899977978838, 'learning_rate': 0.000938800558694719, 'epoch': 0.96}
+
 96%|█████████▌| 498/520 [31:23<01:23,  3.79s/it]
 96%|█████████▌| 499/520 [31:26<01:19,  3.76s/it]
                                                 
{'loss': 1.2564, 'grad_norm': 0.0012869464882710633, 'learning_rate': 0.0008555138626189618, 'epoch': 0.96}
+
 96%|█████████▌| 499/520 [31:26<01:19,  3.76s/it]
 96%|█████████▌| 500/520 [31:30<01:14,  3.73s/it]
                                                 
{'loss': 1.2575, 'grad_norm': 0.0014356866085431293, 'learning_rate': 0.0007760793399827937, 'epoch': 0.96}
+
 96%|█████████▌| 500/520 [31:30<01:14,  3.73s/it]
 96%|█████████▋| 501/520 [31:34<01:10,  3.72s/it]
                                                 
{'loss': 1.1659, 'grad_norm': 0.0013687827986543744, 'learning_rate': 0.000700500077146038, 'epoch': 0.96}
+
 96%|█████████▋| 501/520 [31:34<01:10,  3.72s/it]
 97%|█████████▋| 502/520 [31:37<01:06,  3.71s/it]
                                                 
{'loss': 1.1755, 'grad_norm': 0.0011620227066610213, 'learning_rate': 0.0006287790106757397, 'epoch': 0.97}
+
 97%|█████████▋| 502/520 [31:37<01:06,  3.71s/it]
 97%|█████████▋| 503/520 [31:41<01:02,  3.69s/it]
                                                 
{'loss': 1.1518, 'grad_norm': 0.0012504585464354504, 'learning_rate': 0.0005609189272320237, 'epoch': 0.97}
+
 97%|█████████▋| 503/520 [31:41<01:02,  3.69s/it]
 97%|█████████▋| 504/520 [31:45<00:59,  3.70s/it]
                                                 
{'loss': 1.1685, 'grad_norm': 0.0014093304561773583, 'learning_rate': 0.000496922463459859, 'epoch': 0.97}
+
 97%|█████████▋| 504/520 [31:45<00:59,  3.70s/it]
 97%|█████████▋| 505/520 [31:48<00:55,  3.69s/it]
                                                 
{'loss': 1.2027, 'grad_norm': 0.0013237220219954216, 'learning_rate': 0.0004367921058866187, 'epoch': 0.97}
+
 97%|█████████▋| 505/520 [31:48<00:55,  3.69s/it]
 97%|█████████▋| 506/520 [31:52<00:51,  3.69s/it]
                                                 
{'loss': 1.1266, 'grad_norm': 0.0012778719369473072, 'learning_rate': 0.0003805301908254455, 'epoch': 0.97}
+
 97%|█████████▋| 506/520 [31:52<00:51,  3.69s/it]
 98%|█████████▊| 507/520 [31:56<00:48,  3.70s/it]
                                                 
{'loss': 1.2971, 'grad_norm': 0.0011613706250504581, 'learning_rate': 0.0003281389042844918, 'epoch': 0.97}
+
 98%|█████████▊| 507/520 [31:56<00:48,  3.70s/it]
 98%|█████████▊| 508/520 [31:59<00:44,  3.69s/it]
                                                 
{'loss': 1.2456, 'grad_norm': 0.0012789093450849877, 'learning_rate': 0.00027962028188198705, 'epoch': 0.98}
+
 98%|█████████▊| 508/520 [31:59<00:44,  3.69s/it]
 98%|█████████▊| 509/520 [32:03<00:40,  3.68s/it]
                                                 
{'loss': 1.2183, 'grad_norm': 0.0012397618858042219, 'learning_rate': 0.00023497620876711256, 'epoch': 0.98}
+
 98%|█████████▊| 509/520 [32:03<00:40,  3.68s/it]
 98%|█████████▊| 510/520 [32:07<00:37,  3.70s/it]
                                                 
{'loss': 1.1634, 'grad_norm': 0.0012177766948724604, 'learning_rate': 0.00019420841954681523, 'epoch': 0.98}
+
 98%|█████████▊| 510/520 [32:07<00:37,  3.70s/it]
 98%|█████████▊| 511/520 [32:11<00:33,  3.69s/it]
                                                 
{'loss': 1.14, 'grad_norm': 0.0011850190533980284, 'learning_rate': 0.00015731849821833956, 'epoch': 0.98}
+
 98%|█████████▊| 511/520 [32:11<00:33,  3.69s/it]
 98%|█████████▊| 512/520 [32:14<00:29,  3.70s/it]
                                                 
{'loss': 1.0269, 'grad_norm': 0.0013681332197487976, 'learning_rate': 0.00012430787810776555, 'epoch': 0.98}
+
 98%|█████████▊| 512/520 [32:14<00:29,  3.70s/it]
 99%|█████████▊| 513/520 [32:18<00:25,  3.71s/it]
                                                 
{'loss': 1.2252, 'grad_norm': 0.0014464632707233192, 'learning_rate': 9.517784181422018e-05, 'epoch': 0.99}
+
 99%|█████████▊| 513/520 [32:18<00:25,  3.71s/it]
 99%|█████████▉| 514/520 [32:22<00:22,  3.70s/it]
                                                 
{'loss': 1.1932, 'grad_norm': 0.0011542237678141213, 'learning_rate': 6.992952116013917e-05, 'epoch': 0.99}
+
 99%|█████████▉| 514/520 [32:22<00:22,  3.70s/it]
 99%|█████████▉| 515/520 [32:25<00:18,  3.70s/it]
                                                 
{'loss': 1.2382, 'grad_norm': 0.0014629540678682076, 'learning_rate': 4.856389714723575e-05, 'epoch': 0.99}
+
 99%|█████████▉| 515/520 [32:25<00:18,  3.70s/it]
 99%|█████████▉| 516/520 [32:29<00:14,  3.70s/it]
                                                 
{'loss': 1.1436, 'grad_norm': 0.0012139199691062895, 'learning_rate': 3.108179991837545e-05, 'epoch': 0.99}
+
 99%|█████████▉| 516/520 [32:29<00:14,  3.70s/it]
 99%|█████████▉| 517/520 [32:33<00:11,  3.67s/it]
                                                 
{'loss': 1.1892, 'grad_norm': 0.0011839611447034516, 'learning_rate': 1.7483908725357544e-05, 'epoch': 0.99}
+
 99%|█████████▉| 517/520 [32:33<00:11,  3.67s/it]
100%|█████████▉| 518/520 [32:36<00:07,  3.66s/it]
                                                 
{'loss': 1.1594, 'grad_norm': 0.001292352071593433, 'learning_rate': 7.770751902513862e-06, 'epoch': 1.0}
+
100%|█████████▉| 518/520 [32:36<00:07,  3.66s/it]
100%|█████████▉| 519/520 [32:40<00:03,  3.64s/it]
                                                 
{'loss': 1.162, 'grad_norm': 0.001227152986519099, 'learning_rate': 1.9427068461808086e-06, 'epoch': 1.0}
+
100%|█████████▉| 519/520 [32:40<00:03,  3.64s/it]
100%|██████████| 520/520 [32:44<00:00,  3.90s/it]
                                                 
{'loss': 1.1649, 'grad_norm': 0.0012327041331646506, 'learning_rate': 0.0, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:44<00:00,  3.90s/it]
                                                 
{'train_runtime': 1964.9674, 'train_samples_per_second': 33.858, 'train_steps_per_second': 0.265, 'train_loss': 1.254533368922197, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:44<00:00,  3.90s/it]
100%|██████████| 520/520 [32:44<00:00,  3.78s/it]
+[2025-10-16 16:54:00,963] [INFO] [launch.py:348:main] Process 2523006 exits successfully.
+[2025-10-16 16:54:00,963] [INFO] [launch.py:348:main] Process 2523005 exits successfully.
+[2025-10-16 16:54:01,965] [INFO] [launch.py:348:main] Process 2523003 exits successfully.
+[2025-10-16 16:54:01,965] [INFO] [launch.py:348:main] Process 2523007 exits successfully.
+[2025-10-16 16:54:01,966] [INFO] [launch.py:348:main] Process 2523004 exits successfully.
+[2025-10-16 16:54:01,966] [INFO] [launch.py:348:main] Process 2523009 exits successfully.
+[2025-10-16 16:54:01,967] [INFO] [launch.py:348:main] Process 2523008 exits successfully.
+[2025-10-16 16:54:04,970] [INFO] [launch.py:348:main] Process 2523002 exits successfully.
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.5_2e-1_connector-5.0_2.5_2e-1_ablation_20251016_161943.log
+Timestamp: 2025-10-16 16:54:07
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation_20251016_165407.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation_20251016_165407.log
new file mode 100644
index 0000000000000000000000000000000000000000..33213f5196f47f33d976c542cde9caead3433f6c
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation_20251016_165407.log
@@ -0,0 +1,2312 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation_20251016_165407.log
+Timestamp: 2025-10-16 16:54:07
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 16:54:10,244] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:12,934] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 16:54:12,935] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 5.0 --temperature_attn_text 2.7 --temperature_mlp_text 2.7 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 5.0 --temperature_attn_vision 2.7 --temperature_mlp_vision 2.7 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 5.0 --temperature_connector 2.7 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 16:54:15,505] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:16,555] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 16:54:16,556] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 16:54:16,556] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 16:54:16,556] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 16:54:16,556] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 16:54:16,556] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 16:54:16,556] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 16:54:16,558] [INFO] [launch.py:253:main] process 2545596 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,560] [INFO] [launch.py:253:main] process 2545597 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,562] [INFO] [launch.py:253:main] process 2545598 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,564] [INFO] [launch.py:253:main] process 2545599 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,566] [INFO] [launch.py:253:main] process 2545600 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,567] [INFO] [launch.py:253:main] process 2545601 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,569] [INFO] [launch.py:253:main] process 2545602 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 16:54:16,571] [INFO] [launch.py:253:main] process 2545603 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.7', '--temperature_mlp_text', '2.7', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.7', '--temperature_mlp_vision', '2.7', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.7', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 16:54:23,204] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,440] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,548] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,552] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,581] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,581] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,601] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,601] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 16:54:23,603] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,603] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 16:54:23,839] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,945] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,953] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,981] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,981] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,995] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 16:54:23,995] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 2.7, 'temperature_mlp': 2.7, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 2.7, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 2.7,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 2.7,
+    "temperature_mlp": 2.7,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2545596:2545596 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545596:2545596 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545596:2545596 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545596:2545596 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545596:2545596 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2545596:2545596 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2545601:2545601 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545601:2545601 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545601:2545601 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545601:2545601 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545601:2545601 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545601:2545601 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2545602:2545602 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545602:2545602 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545602:2545602 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545602:2545602 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545602:2545602 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545602:2545602 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2545599:2545599 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545599:2545599 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545599:2545599 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545599:2545599 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545599:2545599 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545599:2545599 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2545597:2545597 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545597:2545597 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545597:2545597 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545600:2545600 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545600:2545600 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545600:2545600 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545597:2545597 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545597:2545597 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545597:2545597 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2545600:2545600 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545600:2545600 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545600:2545600 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545598:2545598 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545598:2545598 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545598:2545598 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545598:2545598 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545598:2545598 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545598:2545598 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2545603:2545603 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2545603:2545603 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545603:2545603 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545603:2545603 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2545603:2545603 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2545603:2545603 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO ncclCommInitRank comm 0x55ea018cc7e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO ncclCommInitRank comm 0x55ca1cd677c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO ncclCommInitRank comm 0x55f0f3b27770 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO ncclCommInitRank comm 0x55d9d8a78d00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO ncclCommInitRank comm 0x560afa2b0530 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO ncclCommInitRank comm 0x55a654dd5560 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO ncclCommInitRank comm 0x558cef9e88c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO ncclCommInitRank comm 0x55762cd4b510 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x85aa04431aba1802 - Init START
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO comm 0x55ea018cc7e0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO comm 0x560afa2b0530 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO comm 0x55ca1cd677c0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO comm 0x55f0f3b27770 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO comm 0x55d9d8a78d00 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO comm 0x55762cd4b510 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO comm 0x558cef9e88c0 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO comm 0x55a654dd5560 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545603:2547200 [7] NCCL INFO ncclCommInitRank comm 0x55f0f3b27770 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545601:2547178 [5] NCCL INFO ncclCommInitRank comm 0x55762cd4b510 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545602:2547195 [6] NCCL INFO ncclCommInitRank comm 0x55d9d8a78d00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545596:2547177 [0] NCCL INFO ncclCommInitRank comm 0x55ca1cd677c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2545598:2547199 [2] NCCL INFO ncclCommInitRank comm 0x55ea018cc7e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545599:2547196 [3] NCCL INFO ncclCommInitRank comm 0x55a654dd5560 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545597:2547197 [1] NCCL INFO ncclCommInitRank comm 0x560afa2b0530 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x85aa04431aba1802 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2545600:2547198 [4] NCCL INFO ncclCommInitRank comm 0x558cef9e88c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x85aa04431aba1802 - Init COMPLETE
+[2025-10-16 16:55:10,598] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 16:55:12,369] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=5.000000
+Pre-training init connector._connector.0.scores: Mean=5.000005
+Pre-training init connector._connector.2.scores: Mean=4.999970
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 16:55:30,184 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 16:55:30,192 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO ncclCommInitRank comm 0x7f794006af80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO ncclCommInitRank comm 0x7fe88c06ae00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO ncclCommInitRank comm 0x7f0ef006a8e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO ncclCommInitRank comm 0x7f0ed006b930 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO ncclCommInitRank comm 0x7f2be006b570 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO ncclCommInitRank comm 0x7f6f0806a8c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO ncclCommInitRank comm 0x7f11b806b1b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO ncclCommInitRank comm 0x7f593806adf0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xd3bbdc2810e0afe6 - Init START
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO comm 0x7f593806adf0 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO comm 0x7f0ed006b930 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO comm 0x7f11b806b1b0 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO comm 0x7f2be006b570 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO comm 0x7f0ef006a8e0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO comm 0x7f794006af80 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO comm 0x7fe88c06ae00 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO comm 0x7f6f0806a8c0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2545600:2552134 [4] NCCL INFO ncclCommInitRank comm 0x7f794006af80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545602:2552128 [6] NCCL INFO ncclCommInitRank comm 0x7f0ed006b930 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545598:2552129 [2] NCCL INFO ncclCommInitRank comm 0x7f0ef006a8e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545596:2552127 [0] NCCL INFO ncclCommInitRank comm 0x7f6f0806a8c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545603:2552131 [7] NCCL INFO ncclCommInitRank comm 0x7f593806adf0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545599:2552133 [3] NCCL INFO ncclCommInitRank comm 0x7f11b806b1b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545601:2552132 [5] NCCL INFO ncclCommInitRank comm 0x7f2be006b570 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2545597:2552130 [1] NCCL INFO ncclCommInitRank comm 0x7fe88c06ae00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xd3bbdc2810e0afe6 - Init COMPLETE
+
  0%|          | 1/520 [00:39<5:45:37, 39.96s/it]
                                                 
{'loss': 2.3769, 'grad_norm': 0.029798110377224307, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:39<5:45:37, 39.96s/it]
  0%|          | 2/520 [00:43<2:40:38, 18.61s/it]
                                                 
{'loss': 2.3049, 'grad_norm': 0.030016260541173925, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:43<2:40:38, 18.61s/it]
  1%|          | 3/520 [00:47<1:41:30, 11.78s/it]
                                                 
{'loss': 2.4879, 'grad_norm': 0.03427137544837902, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:47<1:41:30, 11.78s/it]
  1%|          | 4/520 [00:50<1:13:38,  8.56s/it]
                                                 
{'loss': 1.8305, 'grad_norm': 0.012803251273831088, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:50<1:13:38,  8.56s/it]
  1%|          | 5/520 [00:54<58:18,  6.79s/it]  
                                               
{'loss': 1.8054, 'grad_norm': 0.008779124598048698, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:54<58:18,  6.79s/it]
  1%|          | 6/520 [00:58<49:04,  5.73s/it]
                                               
{'loss': 1.5857, 'grad_norm': 0.0074798672904714, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:58<49:04,  5.73s/it]
  1%|▏         | 7/520 [01:01<43:12,  5.05s/it]
                                               
{'loss': 1.5839, 'grad_norm': 0.010080946335420723, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [01:01<43:12,  5.05s/it]
  2%|▏         | 8/520 [01:06<40:56,  4.80s/it]
                                               
{'loss': 1.5661, 'grad_norm': 0.006392728394634432, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [01:06<40:56,  4.80s/it]
  2%|▏         | 9/520 [01:09<37:46,  4.43s/it]
                                               
{'loss': 1.5974, 'grad_norm': 0.003681227043031144, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [01:09<37:46,  4.43s/it]
  2%|▏         | 10/520 [01:13<35:50,  4.22s/it]
                                                
{'loss': 1.4284, 'grad_norm': 0.0038261001574528847, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [01:13<35:50,  4.22s/it]
  2%|▏         | 11/520 [01:17<35:03,  4.13s/it]
                                                
{'loss': 1.4979, 'grad_norm': 0.004441614909413127, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [01:17<35:03,  4.13s/it]
  2%|▏         | 12/520 [01:21<34:10,  4.04s/it]
                                                
{'loss': 1.4258, 'grad_norm': 0.0035674939985428956, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [01:21<34:10,  4.04s/it][2025-10-16 16:57:00,667] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:25<34:55,  4.13s/it]
                                                
{'loss': 1.4356, 'grad_norm': 0.0026573559388901637, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:25<34:55,  4.13s/it]
  3%|▎         | 14/520 [01:29<33:43,  4.00s/it]
                                                
{'loss': 1.4823, 'grad_norm': 0.002998360553071215, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:29<33:43,  4.00s/it]
  3%|▎         | 15/520 [01:33<33:21,  3.96s/it]
                                                
{'loss': 1.4703, 'grad_norm': 0.0025435298848473804, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:33<33:21,  3.96s/it]
  3%|▎         | 16/520 [01:36<32:49,  3.91s/it]
                                                
{'loss': 1.4333, 'grad_norm': 0.002681622623899878, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:36<32:49,  3.91s/it]
  3%|▎         | 17/520 [01:40<32:34,  3.89s/it]
                                                
{'loss': 1.5195, 'grad_norm': 0.002507962204247305, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:40<32:34,  3.89s/it]
  3%|▎         | 18/520 [01:44<32:23,  3.87s/it]
                                                
{'loss': 1.3675, 'grad_norm': 0.002276943117249417, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:44<32:23,  3.87s/it]
  4%|▎         | 19/520 [01:48<32:11,  3.86s/it]
                                                
{'loss': 1.4233, 'grad_norm': 0.001983287793476691, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:48<32:11,  3.86s/it]
  4%|▍         | 20/520 [01:52<32:01,  3.84s/it]
                                                
{'loss': 1.3642, 'grad_norm': 0.0032461903162226977, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:52<32:01,  3.84s/it]
  4%|▍         | 21/520 [01:56<31:58,  3.85s/it]
                                                
{'loss': 1.4149, 'grad_norm': 0.0022935549548995706, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:56<31:58,  3.85s/it]
  4%|▍         | 22/520 [01:59<31:53,  3.84s/it]
                                                
{'loss': 1.5055, 'grad_norm': 0.0019536878043641607, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:59<31:53,  3.84s/it]
  4%|▍         | 23/520 [02:03<31:44,  3.83s/it]
                                                
{'loss': 1.4395, 'grad_norm': 0.0019179214214923954, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [02:03<31:44,  3.83s/it]
  5%|▍         | 24/520 [02:07<31:35,  3.82s/it]
                                                
{'loss': 1.3956, 'grad_norm': 0.0019235751588018797, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [02:07<31:35,  3.82s/it]
  5%|▍         | 25/520 [02:11<31:29,  3.82s/it]
                                                
{'loss': 1.4421, 'grad_norm': 0.0020939324686493565, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [02:11<31:29,  3.82s/it]
  5%|▌         | 26/520 [02:15<31:28,  3.82s/it]
                                                
{'loss': 1.4259, 'grad_norm': 0.001905632171700605, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [02:15<31:28,  3.82s/it]
  5%|▌         | 27/520 [02:19<31:24,  3.82s/it]
                                                
{'loss': 1.3336, 'grad_norm': 0.0020703887832360076, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [02:19<31:24,  3.82s/it]
  5%|▌         | 28/520 [02:22<31:19,  3.82s/it]
                                                
{'loss': 1.3385, 'grad_norm': 0.0019231214512069486, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [02:22<31:19,  3.82s/it]
  6%|▌         | 29/520 [02:26<30:55,  3.78s/it]
                                                
{'loss': 1.3692, 'grad_norm': 0.0019128986302675596, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [02:26<30:55,  3.78s/it]
  6%|▌         | 30/520 [02:30<30:35,  3.75s/it]
                                                
{'loss': 1.4706, 'grad_norm': 0.0019000638520114748, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:30<30:35,  3.75s/it]
  6%|▌         | 31/520 [02:33<30:22,  3.73s/it]
                                                
{'loss': 1.3422, 'grad_norm': 0.0016818345591014254, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:33<30:22,  3.73s/it]
  6%|▌         | 32/520 [02:37<30:16,  3.72s/it]
                                                
{'loss': 1.3348, 'grad_norm': 0.002032766477082332, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:37<30:16,  3.72s/it]
  6%|▋         | 33/520 [02:41<30:08,  3.71s/it]
                                                
{'loss': 1.3483, 'grad_norm': 0.0018187507936571524, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:41<30:08,  3.71s/it]
  7%|▋         | 34/520 [02:44<30:00,  3.71s/it]
                                                
{'loss': 1.3379, 'grad_norm': 0.0018902450097524482, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:44<30:00,  3.71s/it]
  7%|▋         | 35/520 [02:48<29:52,  3.70s/it]
                                                
{'loss': 1.3456, 'grad_norm': 0.002067908001275228, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:48<29:52,  3.70s/it]
  7%|▋         | 36/520 [02:52<29:50,  3.70s/it]
                                                
{'loss': 1.448, 'grad_norm': 0.0019264776285048523, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:52<29:50,  3.70s/it]
  7%|▋         | 37/520 [02:56<29:55,  3.72s/it]
                                                
{'loss': 1.4396, 'grad_norm': 0.0017788110999515919, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:56<29:55,  3.72s/it]
  7%|▋         | 38/520 [02:59<29:58,  3.73s/it]
                                                
{'loss': 1.5163, 'grad_norm': 0.0017628867655778704, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:59<29:58,  3.73s/it]
  8%|▊         | 39/520 [03:03<30:00,  3.74s/it]
                                                
{'loss': 1.3674, 'grad_norm': 0.0020402832961612044, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [03:03<30:00,  3.74s/it]
  8%|▊         | 40/520 [03:07<30:01,  3.75s/it]
                                                
{'loss': 1.3977, 'grad_norm': 0.0019062384234252083, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [03:07<30:01,  3.75s/it]
  8%|▊         | 41/520 [03:11<30:00,  3.76s/it]
                                                
{'loss': 1.3803, 'grad_norm': 0.0016852854630410577, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [03:11<30:00,  3.76s/it]
  8%|▊         | 42/520 [03:14<29:55,  3.76s/it]
                                                
{'loss': 1.3972, 'grad_norm': 0.0023966739155945214, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [03:14<29:55,  3.76s/it]
  8%|▊         | 43/520 [03:18<29:51,  3.76s/it]
                                                
{'loss': 1.3359, 'grad_norm': 0.0014960273475934929, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [03:18<29:51,  3.76s/it]
  8%|▊         | 44/520 [03:22<29:49,  3.76s/it]
                                                
{'loss': 1.4365, 'grad_norm': 0.0016846758406019859, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [03:22<29:49,  3.76s/it]
  9%|▊         | 45/520 [03:26<29:44,  3.76s/it]
                                                
{'loss': 1.3994, 'grad_norm': 0.0017660180512118746, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [03:26<29:44,  3.76s/it]
  9%|▉         | 46/520 [03:29<29:42,  3.76s/it]
                                                
{'loss': 1.4931, 'grad_norm': 0.0016715632677349066, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:29<29:42,  3.76s/it]
  9%|▉         | 47/520 [03:33<29:37,  3.76s/it]
                                                
{'loss': 1.3799, 'grad_norm': 0.0017883444502510906, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:33<29:37,  3.76s/it]
  9%|▉         | 48/520 [03:37<29:30,  3.75s/it]
                                                
{'loss': 1.3607, 'grad_norm': 0.001932312446528599, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:37<29:30,  3.75s/it]
  9%|▉         | 49/520 [03:41<29:26,  3.75s/it]
                                                
{'loss': 1.4023, 'grad_norm': 0.0016905356873504619, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:41<29:26,  3.75s/it]
 10%|▉         | 50/520 [03:44<29:25,  3.76s/it]
                                                
{'loss': 1.3888, 'grad_norm': 0.0016297660760595822, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:44<29:25,  3.76s/it]
 10%|▉         | 51/520 [03:48<29:18,  3.75s/it]
                                                
{'loss': 1.3242, 'grad_norm': 0.001793820908432587, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:48<29:18,  3.75s/it]
 10%|█         | 52/520 [03:52<29:16,  3.75s/it]
                                                
{'loss': 1.4523, 'grad_norm': 0.001974801651578422, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:52<29:16,  3.75s/it]
 10%|█         | 53/520 [03:56<29:09,  3.75s/it]
                                                
{'loss': 1.4369, 'grad_norm': 0.0018381467521449144, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:56<29:09,  3.75s/it]
 10%|█         | 54/520 [03:59<29:07,  3.75s/it]
                                                
{'loss': 1.3504, 'grad_norm': 0.0016957310763211992, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:59<29:07,  3.75s/it]
 11%|█         | 55/520 [04:03<29:12,  3.77s/it]
                                                
{'loss': 1.3243, 'grad_norm': 0.0018499403459966545, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [04:03<29:12,  3.77s/it]
 11%|█         | 56/520 [04:07<29:04,  3.76s/it]
                                                
{'loss': 1.454, 'grad_norm': 0.001757117519425879, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [04:07<29:04,  3.76s/it]
 11%|█         | 57/520 [04:11<29:04,  3.77s/it]
                                                
{'loss': 1.3232, 'grad_norm': 0.0021713355688120794, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [04:11<29:04,  3.77s/it]
 11%|█         | 58/520 [04:15<29:05,  3.78s/it]
                                                
{'loss': 1.4785, 'grad_norm': 0.0017648004461958508, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [04:15<29:05,  3.78s/it]
 11%|█▏        | 59/520 [04:18<29:00,  3.78s/it]
                                                
{'loss': 1.2956, 'grad_norm': 0.0015539019015725012, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [04:18<29:00,  3.78s/it]
 12%|█▏        | 60/520 [04:22<28:59,  3.78s/it]
                                                
{'loss': 1.3856, 'grad_norm': 0.0018135490009233302, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [04:22<28:59,  3.78s/it]
 12%|█▏        | 61/520 [04:26<28:53,  3.78s/it]
                                                
{'loss': 1.3875, 'grad_norm': 0.0019496754880535152, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [04:26<28:53,  3.78s/it]
 12%|█▏        | 62/520 [04:30<28:48,  3.77s/it]
                                                
{'loss': 1.3643, 'grad_norm': 0.0017453082465143524, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [04:30<28:48,  3.77s/it]
 12%|█▏        | 63/520 [04:33<28:45,  3.78s/it]
                                                
{'loss': 1.3548, 'grad_norm': 0.001627183917513781, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:34<28:45,  3.78s/it]
 12%|█▏        | 64/520 [04:37<28:41,  3.78s/it]
                                                
{'loss': 1.3846, 'grad_norm': 0.001880834895016523, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:37<28:41,  3.78s/it]
 12%|█▎        | 65/520 [04:41<28:39,  3.78s/it]
                                                
{'loss': 1.3901, 'grad_norm': 0.002173123036751027, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:41<28:39,  3.78s/it]
 13%|█▎        | 66/520 [04:45<28:38,  3.79s/it]
                                                
{'loss': 1.3421, 'grad_norm': 0.001546939052533465, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:45<28:38,  3.79s/it]
 13%|█▎        | 67/520 [04:49<28:34,  3.78s/it]
                                                
{'loss': 1.2475, 'grad_norm': 0.0018137908744409825, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:49<28:34,  3.78s/it]
 13%|█▎        | 68/520 [04:52<28:30,  3.78s/it]
                                                
{'loss': 1.3197, 'grad_norm': 0.002003416912491809, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:52<28:30,  3.78s/it]
 13%|█▎        | 69/520 [04:56<28:24,  3.78s/it]
                                                
{'loss': 1.2921, 'grad_norm': 0.001702850194407373, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:56<28:24,  3.78s/it]
 13%|█▎        | 70/520 [05:00<28:22,  3.78s/it]
                                                
{'loss': 1.3262, 'grad_norm': 0.0018190263449282887, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [05:00<28:22,  3.78s/it]
 14%|█▎        | 71/520 [05:04<28:18,  3.78s/it]
                                                
{'loss': 1.2651, 'grad_norm': 0.0016151612253392858, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [05:04<28:18,  3.78s/it]
 14%|█▍        | 72/520 [05:08<28:12,  3.78s/it]
                                                
{'loss': 1.4163, 'grad_norm': 0.0018823365867546337, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [05:08<28:12,  3.78s/it]
 14%|█▍        | 73/520 [05:11<28:10,  3.78s/it]
                                                
{'loss': 1.2454, 'grad_norm': 0.0016132601178282676, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [05:11<28:10,  3.78s/it]
 14%|█▍        | 74/520 [05:15<28:08,  3.79s/it]
                                                
{'loss': 1.3564, 'grad_norm': 0.0019816579664889153, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [05:15<28:08,  3.79s/it]
 14%|█▍        | 75/520 [05:19<27:59,  3.78s/it]
                                                
{'loss': 1.2608, 'grad_norm': 0.001560649730757898, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [05:19<27:59,  3.78s/it]
 15%|█▍        | 76/520 [05:23<27:44,  3.75s/it]
                                                
{'loss': 1.4323, 'grad_norm': 0.001481039864380433, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [05:23<27:44,  3.75s/it]
 15%|█▍        | 77/520 [05:26<27:33,  3.73s/it]
                                                
{'loss': 1.1855, 'grad_norm': 0.0017714855209727042, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [05:26<27:33,  3.73s/it]
 15%|█▌        | 78/520 [05:30<27:24,  3.72s/it]
                                                
{'loss': 1.303, 'grad_norm': 0.0017087144927351646, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:30<27:24,  3.72s/it]
 15%|█▌        | 79/520 [05:34<27:19,  3.72s/it]
                                                
{'loss': 1.2913, 'grad_norm': 0.001597924784834565, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:34<27:19,  3.72s/it]
 15%|█▌        | 80/520 [05:37<27:14,  3.71s/it]
                                                
{'loss': 1.4205, 'grad_norm': 0.0016098086706337294, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:37<27:14,  3.71s/it]
 16%|█▌        | 81/520 [05:41<27:10,  3.72s/it]
                                                
{'loss': 1.4308, 'grad_norm': 0.0020885265062736435, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:41<27:10,  3.72s/it]
 16%|█▌        | 82/520 [05:45<27:04,  3.71s/it]
                                                
{'loss': 1.3541, 'grad_norm': 0.0015535620002038743, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:45<27:04,  3.71s/it]
 16%|█▌        | 83/520 [05:48<26:55,  3.70s/it]
                                                
{'loss': 1.3819, 'grad_norm': 0.0017429777606384304, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:48<26:55,  3.70s/it]
 16%|█▌        | 84/520 [05:52<26:47,  3.69s/it]
                                                
{'loss': 1.389, 'grad_norm': 0.001697181560494442, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:52<26:47,  3.69s/it]
 16%|█▋        | 85/520 [05:56<26:41,  3.68s/it]
                                                
{'loss': 1.4169, 'grad_norm': 0.0016696137995506738, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:56<26:41,  3.68s/it]
 17%|█▋        | 86/520 [05:59<26:38,  3.68s/it]
                                                
{'loss': 1.4145, 'grad_norm': 0.0015619461543719692, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:59<26:38,  3.68s/it]
 17%|█▋        | 87/520 [06:03<26:39,  3.69s/it]
                                                
{'loss': 1.3709, 'grad_norm': 0.001756804105947966, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [06:03<26:39,  3.69s/it]
 17%|█▋        | 88/520 [06:07<26:35,  3.69s/it]
                                                
{'loss': 1.319, 'grad_norm': 0.0014528368976004036, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [06:07<26:35,  3.69s/it]
 17%|█▋        | 89/520 [06:11<26:48,  3.73s/it]
                                                
{'loss': 1.3785, 'grad_norm': 0.0017799828673876511, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [06:11<26:48,  3.73s/it]
 17%|█▋        | 90/520 [06:15<26:56,  3.76s/it]
                                                
{'loss': 1.3061, 'grad_norm': 0.001495787675660503, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [06:15<26:56,  3.76s/it]
 18%|█▊        | 91/520 [06:18<27:01,  3.78s/it]
                                                
{'loss': 1.3769, 'grad_norm': 0.0014533078767783986, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [06:18<27:01,  3.78s/it]
 18%|█▊        | 92/520 [06:22<27:06,  3.80s/it]
                                                
{'loss': 1.3115, 'grad_norm': 0.0016249498516385118, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [06:22<27:06,  3.80s/it]
 18%|█▊        | 93/520 [06:26<27:07,  3.81s/it]
                                                
{'loss': 1.3376, 'grad_norm': 0.0018883417511272248, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [06:26<27:07,  3.81s/it]
 18%|█▊        | 94/520 [06:30<27:09,  3.82s/it]
                                                
{'loss': 1.4147, 'grad_norm': 0.0016026772309582695, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [06:30<27:09,  3.82s/it]
 18%|█▊        | 95/520 [06:34<27:05,  3.82s/it]
                                                
{'loss': 1.3098, 'grad_norm': 0.0019500955317820205, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:34<27:05,  3.82s/it]
 18%|█▊        | 96/520 [06:38<26:57,  3.82s/it]
                                                
{'loss': 1.3259, 'grad_norm': 0.0013650718917831831, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:38<26:57,  3.82s/it]
 19%|█▊        | 97/520 [06:41<26:55,  3.82s/it]
                                                
{'loss': 1.2905, 'grad_norm': 0.0020613177119068577, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:41<26:55,  3.82s/it]
 19%|█▉        | 98/520 [06:45<26:52,  3.82s/it]
                                                
{'loss': 1.2932, 'grad_norm': 0.001350040591362297, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:45<26:52,  3.82s/it]
 19%|█▉        | 99/520 [06:49<26:50,  3.83s/it]
                                                
{'loss': 1.3041, 'grad_norm': 0.0016697960124967615, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:49<26:50,  3.83s/it]
 19%|█▉        | 100/520 [06:53<26:46,  3.83s/it]
                                                 
{'loss': 1.3105, 'grad_norm': 0.0016575759443162158, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:53<26:46,  3.83s/it]
 19%|█▉        | 101/520 [06:57<26:40,  3.82s/it]
                                                 
{'loss': 1.3085, 'grad_norm': 0.0016608386509603596, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:57<26:40,  3.82s/it]
 20%|█▉        | 102/520 [07:00<26:36,  3.82s/it]
                                                 
{'loss': 1.321, 'grad_norm': 0.0017188399110963433, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [07:00<26:36,  3.82s/it]
 20%|█▉        | 103/520 [07:04<26:30,  3.81s/it]
                                                 
{'loss': 1.2442, 'grad_norm': 0.0014053159935060496, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [07:04<26:30,  3.81s/it]
 20%|██        | 104/520 [07:08<26:28,  3.82s/it]
                                                 
{'loss': 1.3177, 'grad_norm': 0.001522277755344982, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [07:08<26:28,  3.82s/it]
 20%|██        | 105/520 [07:12<26:29,  3.83s/it]
                                                 
{'loss': 1.3059, 'grad_norm': 0.001440057425702297, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [07:12<26:29,  3.83s/it]
 20%|██        | 106/520 [07:16<26:24,  3.83s/it]
                                                 
{'loss': 1.3263, 'grad_norm': 0.0014962570378684627, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [07:16<26:24,  3.83s/it]
 21%|██        | 107/520 [07:20<26:23,  3.83s/it]
                                                 
{'loss': 1.3002, 'grad_norm': 0.0014752079895775446, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [07:20<26:23,  3.83s/it]
 21%|██        | 108/520 [07:23<26:18,  3.83s/it]
                                                 
{'loss': 1.2719, 'grad_norm': 0.0015777857562428105, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [07:23<26:18,  3.83s/it]
 21%|██        | 109/520 [07:27<26:13,  3.83s/it]
                                                 
{'loss': 1.2826, 'grad_norm': 0.0013341877682882801, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [07:27<26:13,  3.83s/it]
 21%|██        | 110/520 [07:31<26:07,  3.82s/it]
                                                 
{'loss': 1.4556, 'grad_norm': 0.0017278576061424131, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [07:31<26:07,  3.82s/it]
 21%|██▏       | 111/520 [07:35<26:05,  3.83s/it]
                                                 
{'loss': 1.4562, 'grad_norm': 0.001613950787541268, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:35<26:05,  3.83s/it]
 22%|██▏       | 112/520 [07:39<25:58,  3.82s/it]
                                                 
{'loss': 1.3332, 'grad_norm': 0.0014442073424097584, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:39<25:58,  3.82s/it]
 22%|██▏       | 113/520 [07:42<25:51,  3.81s/it]
                                                 
{'loss': 1.2277, 'grad_norm': 0.0016032218417205292, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:43<25:51,  3.81s/it]
 22%|██▏       | 114/520 [07:46<25:34,  3.78s/it]
                                                 
{'loss': 1.3212, 'grad_norm': 0.0015138614398172903, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:46<25:34,  3.78s/it]
 22%|██▏       | 115/520 [07:50<25:18,  3.75s/it]
                                                 
{'loss': 1.4212, 'grad_norm': 0.0014214200172357548, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:50<25:18,  3.75s/it]
 22%|██▏       | 116/520 [07:54<25:04,  3.73s/it]
                                                 
{'loss': 1.434, 'grad_norm': 0.0015626420057493118, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:54<25:04,  3.73s/it]
 22%|██▎       | 117/520 [07:57<24:52,  3.70s/it]
                                                 
{'loss': 1.4047, 'grad_norm': 0.0015674437270418156, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:57<24:52,  3.70s/it]
 23%|██▎       | 118/520 [08:01<24:47,  3.70s/it]
                                                 
{'loss': 1.3053, 'grad_norm': 0.001380473941803947, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [08:01<24:47,  3.70s/it]
 23%|██▎       | 119/520 [08:05<24:39,  3.69s/it]
                                                 
{'loss': 1.2651, 'grad_norm': 0.001620017102787244, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [08:05<24:39,  3.69s/it]
 23%|██▎       | 120/520 [08:08<24:31,  3.68s/it]
                                                 
{'loss': 1.2755, 'grad_norm': 0.0017275498477044822, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [08:08<24:31,  3.68s/it]
 23%|██▎       | 121/520 [08:12<24:23,  3.67s/it]
                                                 
{'loss': 1.3347, 'grad_norm': 0.0015430380851232887, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [08:12<24:23,  3.67s/it]
 23%|██▎       | 122/520 [08:16<24:21,  3.67s/it]
                                                 
{'loss': 1.2391, 'grad_norm': 0.0014006056489097442, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [08:16<24:21,  3.67s/it]
 24%|██▎       | 123/520 [08:19<24:21,  3.68s/it]
                                                 
{'loss': 1.3714, 'grad_norm': 0.0014785996948871035, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [08:19<24:21,  3.68s/it]
 24%|██▍       | 124/520 [08:23<24:15,  3.68s/it]
                                                 
{'loss': 1.3003, 'grad_norm': 0.0015107911793257604, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [08:23<24:15,  3.68s/it]
 24%|██▍       | 125/520 [08:27<24:21,  3.70s/it]
                                                 
{'loss': 1.2954, 'grad_norm': 0.0014924984268754975, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [08:27<24:21,  3.70s/it]
 24%|██▍       | 126/520 [08:31<25:35,  3.90s/it]
                                                 
{'loss': 1.2942, 'grad_norm': 0.0013690607727137336, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [08:31<25:35,  3.90s/it]
 24%|██▍       | 127/520 [08:35<25:09,  3.84s/it]
                                                 
{'loss': 1.2791, 'grad_norm': 0.0016923083150155334, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:35<25:09,  3.84s/it]
 25%|██▍       | 128/520 [08:38<24:48,  3.80s/it]
                                                 
{'loss': 1.3305, 'grad_norm': 0.0015736551715450503, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:38<24:48,  3.80s/it]
 25%|██▍       | 129/520 [08:42<24:37,  3.78s/it]
                                                 
{'loss': 1.2725, 'grad_norm': 0.001350291838702264, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:42<24:37,  3.78s/it]
 25%|██▌       | 130/520 [08:46<24:22,  3.75s/it]
                                                 
{'loss': 1.3129, 'grad_norm': 0.0012863256320060188, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:46<24:22,  3.75s/it]
 25%|██▌       | 131/520 [08:50<24:11,  3.73s/it]
                                                 
{'loss': 1.2643, 'grad_norm': 0.0013090769941210936, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:50<24:11,  3.73s/it]
 25%|██▌       | 132/520 [08:53<24:00,  3.71s/it]
                                                 
{'loss': 1.3639, 'grad_norm': 0.0015323011976756304, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:53<24:00,  3.71s/it]
 26%|██▌       | 133/520 [08:57<23:55,  3.71s/it]
                                                 
{'loss': 1.2746, 'grad_norm': 0.001571886393823312, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:57<23:55,  3.71s/it]
 26%|██▌       | 134/520 [09:01<23:48,  3.70s/it]
                                                 
{'loss': 1.3566, 'grad_norm': 0.0014776653785617733, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [09:01<23:48,  3.70s/it]
 26%|██▌       | 135/520 [09:04<23:45,  3.70s/it]
                                                 
{'loss': 1.4114, 'grad_norm': 0.0014943702785038412, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [09:04<23:45,  3.70s/it]
 26%|██▌       | 136/520 [09:08<23:38,  3.69s/it]
                                                 
{'loss': 1.3443, 'grad_norm': 0.001468255611958697, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [09:08<23:38,  3.69s/it]
 26%|██▋       | 137/520 [09:12<23:34,  3.69s/it]
                                                 
{'loss': 1.257, 'grad_norm': 0.001574236586967528, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [09:12<23:34,  3.69s/it]
 27%|██▋       | 138/520 [09:15<23:29,  3.69s/it]
                                                 
{'loss': 1.2787, 'grad_norm': 0.001466580277644756, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [09:15<23:29,  3.69s/it]
 27%|██▋       | 139/520 [09:19<23:33,  3.71s/it]
                                                 
{'loss': 1.1696, 'grad_norm': 0.0012930109673682975, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [09:19<23:33,  3.71s/it]
 27%|██▋       | 140/520 [09:23<23:26,  3.70s/it]
                                                 
{'loss': 1.3128, 'grad_norm': 0.0013591479632124627, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [09:23<23:26,  3.70s/it]
 27%|██▋       | 141/520 [09:26<23:23,  3.70s/it]
                                                 
{'loss': 1.3878, 'grad_norm': 0.0014545922624418074, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [09:26<23:23,  3.70s/it]
 27%|██▋       | 142/520 [09:30<23:25,  3.72s/it]
                                                 
{'loss': 1.3211, 'grad_norm': 0.0012755363110248503, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [09:30<23:25,  3.72s/it]
 28%|██▊       | 143/520 [09:34<23:16,  3.70s/it]
                                                 
{'loss': 1.3052, 'grad_norm': 0.0015710654027775227, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:34<23:16,  3.70s/it]
 28%|██▊       | 144/520 [09:38<23:11,  3.70s/it]
                                                 
{'loss': 1.2668, 'grad_norm': 0.0014942023504785815, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:38<23:11,  3.70s/it]
 28%|██▊       | 145/520 [09:41<23:09,  3.70s/it]
                                                 
{'loss': 1.2034, 'grad_norm': 0.0012735729456733718, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:41<23:09,  3.70s/it]
 28%|██▊       | 146/520 [09:45<23:01,  3.69s/it]
                                                 
{'loss': 1.3908, 'grad_norm': 0.001442489930070495, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:45<23:01,  3.69s/it]
 28%|██▊       | 147/520 [09:49<22:55,  3.69s/it]
                                                 
{'loss': 1.2461, 'grad_norm': 0.0014168310863357629, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:49<22:55,  3.69s/it]
 28%|██▊       | 148/520 [09:52<22:46,  3.67s/it]
                                                 
{'loss': 1.2778, 'grad_norm': 0.0014852330861966176, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:52<22:46,  3.67s/it]
 29%|██▊       | 149/520 [09:56<22:44,  3.68s/it]
                                                 
{'loss': 1.221, 'grad_norm': 0.0013830452131298981, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:56<22:44,  3.68s/it]
 29%|██▉       | 150/520 [10:00<22:38,  3.67s/it]
                                                 
{'loss': 1.4544, 'grad_norm': 0.001464683126445641, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [10:00<22:38,  3.67s/it]
 29%|██▉       | 151/520 [10:03<22:34,  3.67s/it]
                                                 
{'loss': 1.2658, 'grad_norm': 0.0014119450992981942, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [10:03<22:34,  3.67s/it]
 29%|██▉       | 152/520 [10:07<22:30,  3.67s/it]
                                                 
{'loss': 1.234, 'grad_norm': 0.001459856935298659, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [10:07<22:30,  3.67s/it]
 29%|██▉       | 153/520 [10:11<22:26,  3.67s/it]
                                                 
{'loss': 1.2704, 'grad_norm': 0.0013757881049726861, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [10:11<22:26,  3.67s/it]
 30%|██▉       | 154/520 [10:14<22:23,  3.67s/it]
                                                 
{'loss': 1.3537, 'grad_norm': 0.0014126563397368849, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [10:14<22:23,  3.67s/it]
 30%|██▉       | 155/520 [10:18<22:21,  3.68s/it]
                                                 
{'loss': 1.2689, 'grad_norm': 0.0014193442793529376, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [10:18<22:21,  3.68s/it]
 30%|███       | 156/520 [10:22<22:19,  3.68s/it]
                                                 
{'loss': 1.3001, 'grad_norm': 0.0015392289075397728, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [10:22<22:19,  3.68s/it]
 30%|███       | 157/520 [10:25<22:18,  3.69s/it]
                                                 
{'loss': 1.3783, 'grad_norm': 0.0014048842640407323, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [10:25<22:18,  3.69s/it]
 30%|███       | 158/520 [10:29<22:13,  3.68s/it]
                                                 
{'loss': 1.2772, 'grad_norm': 0.0016775982855674163, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [10:29<22:13,  3.68s/it]
 31%|███       | 159/520 [10:33<22:10,  3.69s/it]
                                                 
{'loss': 1.3139, 'grad_norm': 0.0013949596236183595, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [10:33<22:10,  3.69s/it]
 31%|███       | 160/520 [10:37<22:12,  3.70s/it]
                                                 
{'loss': 1.3257, 'grad_norm': 0.001462077432031761, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:37<22:12,  3.70s/it]
 31%|███       | 161/520 [10:40<22:13,  3.72s/it]
                                                 
{'loss': 1.31, 'grad_norm': 0.0014070177629720122, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:40<22:13,  3.72s/it]
 31%|███       | 162/520 [10:44<22:12,  3.72s/it]
                                                 
{'loss': 1.3133, 'grad_norm': 0.001435879423023296, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:44<22:12,  3.72s/it]
 31%|███▏      | 163/520 [10:48<22:05,  3.71s/it]
                                                 
{'loss': 1.2012, 'grad_norm': 0.0018181923206297226, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:48<22:05,  3.71s/it]
 32%|███▏      | 164/520 [10:51<22:02,  3.71s/it]
                                                 
{'loss': 1.1598, 'grad_norm': 0.0014139269163898208, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:51<22:02,  3.71s/it]
 32%|███▏      | 165/520 [10:55<21:58,  3.72s/it]
                                                 
{'loss': 1.3041, 'grad_norm': 0.0013110953418363992, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:55<21:58,  3.72s/it]
 32%|███▏      | 166/520 [10:59<21:51,  3.70s/it]
                                                 
{'loss': 1.2919, 'grad_norm': 0.001530367099095747, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:59<21:51,  3.70s/it]
 32%|███▏      | 167/520 [11:03<21:50,  3.71s/it]
                                                 
{'loss': 1.2817, 'grad_norm': 0.0015415794532928147, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [11:03<21:50,  3.71s/it]
 32%|███▏      | 168/520 [11:06<21:46,  3.71s/it]
                                                 
{'loss': 1.2194, 'grad_norm': 0.0014053666101231247, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [11:06<21:46,  3.71s/it]
 32%|███▎      | 169/520 [11:10<21:44,  3.72s/it]
                                                 
{'loss': 1.3008, 'grad_norm': 0.0014372668936921003, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [11:10<21:44,  3.72s/it]
 33%|███▎      | 170/520 [11:14<21:39,  3.71s/it]
                                                 
{'loss': 1.2637, 'grad_norm': 0.001330860222653734, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [11:14<21:39,  3.71s/it]
 33%|███▎      | 171/520 [11:17<21:35,  3.71s/it]
                                                 
{'loss': 1.2285, 'grad_norm': 0.0014834774410400816, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [11:17<21:35,  3.71s/it]
 33%|███▎      | 172/520 [11:21<21:31,  3.71s/it]
                                                 
{'loss': 1.2998, 'grad_norm': 0.0013554397201232698, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [11:21<21:31,  3.71s/it]
 33%|███▎      | 173/520 [11:25<21:28,  3.71s/it]
                                                 
{'loss': 1.2451, 'grad_norm': 0.001372918826771801, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [11:25<21:28,  3.71s/it]
 33%|███▎      | 174/520 [11:28<21:21,  3.70s/it]
                                                 
{'loss': 1.2991, 'grad_norm': 0.0014918235002407945, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [11:28<21:21,  3.70s/it]
 34%|███▎      | 175/520 [11:32<21:19,  3.71s/it]
                                                 
{'loss': 1.2092, 'grad_norm': 0.0013375165773006341, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [11:32<21:19,  3.71s/it]
 34%|███▍      | 176/520 [11:36<21:17,  3.71s/it]
                                                 
{'loss': 1.3426, 'grad_norm': 0.0013688885294825, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:36<21:17,  3.71s/it]
 34%|███▍      | 177/520 [11:40<21:14,  3.72s/it]
                                                 
{'loss': 1.2106, 'grad_norm': 0.0015592689967376, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:40<21:14,  3.72s/it]
 34%|███▍      | 178/520 [11:43<21:09,  3.71s/it]
                                                 
{'loss': 1.2747, 'grad_norm': 0.0014956025604134795, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:43<21:09,  3.71s/it]
 34%|███▍      | 179/520 [11:47<21:03,  3.71s/it]
                                                 
{'loss': 1.3571, 'grad_norm': 0.0013149765777752988, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:47<21:03,  3.71s/it]
 35%|███▍      | 180/520 [11:51<20:58,  3.70s/it]
                                                 
{'loss': 1.2789, 'grad_norm': 0.0014678044747205535, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:51<20:58,  3.70s/it]
 35%|███▍      | 181/520 [11:54<20:53,  3.70s/it]
                                                 
{'loss': 1.247, 'grad_norm': 0.001279448569916971, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:54<20:53,  3.70s/it]
 35%|███▌      | 182/520 [11:58<20:46,  3.69s/it]
                                                 
{'loss': 1.2625, 'grad_norm': 0.0014441986321643836, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:58<20:46,  3.69s/it]
 35%|███▌      | 183/520 [12:02<20:43,  3.69s/it]
                                                 
{'loss': 1.2881, 'grad_norm': 0.0013768103249452354, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [12:02<20:43,  3.69s/it]
 35%|███▌      | 184/520 [12:05<20:40,  3.69s/it]
                                                 
{'loss': 1.2129, 'grad_norm': 0.0014606453583232962, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [12:05<20:40,  3.69s/it]
 36%|███▌      | 185/520 [12:09<20:37,  3.69s/it]
                                                 
{'loss': 1.3665, 'grad_norm': 0.0013781681296412699, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [12:09<20:37,  3.69s/it]
 36%|███▌      | 186/520 [12:13<20:34,  3.70s/it]
                                                 
{'loss': 1.238, 'grad_norm': 0.0014008938072967226, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [12:13<20:34,  3.70s/it]
 36%|███▌      | 187/520 [12:17<20:32,  3.70s/it]
                                                 
{'loss': 1.2399, 'grad_norm': 0.0016101962421320364, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [12:17<20:32,  3.70s/it]
 36%|███▌      | 188/520 [12:20<20:28,  3.70s/it]
                                                 
{'loss': 1.3241, 'grad_norm': 0.0014420324313513, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [12:20<20:28,  3.70s/it]
 36%|███▋      | 189/520 [12:24<20:22,  3.69s/it]
                                                 
{'loss': 1.3321, 'grad_norm': 0.001247746802738541, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [12:24<20:22,  3.69s/it]
 37%|███▋      | 190/520 [12:28<20:17,  3.69s/it]
                                                 
{'loss': 1.2537, 'grad_norm': 0.0014927267394360965, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [12:28<20:17,  3.69s/it]
 37%|███▋      | 191/520 [12:31<20:14,  3.69s/it]
                                                 
{'loss': 1.2019, 'grad_norm': 0.0012517887009042604, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [12:31<20:14,  3.69s/it]
 37%|███▋      | 192/520 [12:35<20:11,  3.70s/it]
                                                 
{'loss': 1.2886, 'grad_norm': 0.0013222988539699717, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:35<20:11,  3.70s/it]
 37%|███▋      | 193/520 [12:39<20:07,  3.69s/it]
                                                 
{'loss': 1.2531, 'grad_norm': 0.0016316972081632667, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:39<20:07,  3.69s/it]
 37%|███▋      | 194/520 [12:42<20:03,  3.69s/it]
                                                 
{'loss': 1.1549, 'grad_norm': 0.0014311266215449238, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:42<20:03,  3.69s/it]
 38%|███▊      | 195/520 [12:46<19:58,  3.69s/it]
                                                 
{'loss': 1.3076, 'grad_norm': 0.0014102997990034667, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:46<19:58,  3.69s/it]
 38%|███▊      | 196/520 [12:50<19:51,  3.68s/it]
                                                 
{'loss': 1.2823, 'grad_norm': 0.0015065927560354113, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:50<19:51,  3.68s/it]
 38%|███▊      | 197/520 [12:53<19:49,  3.68s/it]
                                                 
{'loss': 1.2335, 'grad_norm': 0.0013290320774925145, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:53<19:49,  3.68s/it]
 38%|███▊      | 198/520 [12:57<19:46,  3.69s/it]
                                                 
{'loss': 1.3003, 'grad_norm': 0.001454416183633743, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:57<19:46,  3.69s/it]
 38%|███▊      | 199/520 [13:01<19:48,  3.70s/it]
                                                 
{'loss': 1.2202, 'grad_norm': 0.001396735262531841, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [13:01<19:48,  3.70s/it]
 38%|███▊      | 200/520 [13:05<19:43,  3.70s/it]
                                                 
{'loss': 1.2046, 'grad_norm': 0.001434586062619354, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [13:05<19:43,  3.70s/it]
 39%|███▊      | 201/520 [13:08<19:39,  3.70s/it]
                                                 
{'loss': 1.2336, 'grad_norm': 0.0011881494329047683, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [13:08<19:39,  3.70s/it]
 39%|███▉      | 202/520 [13:12<19:34,  3.69s/it]
                                                 
{'loss': 1.2194, 'grad_norm': 0.0013836204899725201, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [13:12<19:34,  3.69s/it]
 39%|███▉      | 203/520 [13:16<19:28,  3.69s/it]
                                                 
{'loss': 1.2795, 'grad_norm': 0.0014507992598395874, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [13:16<19:28,  3.69s/it]
 39%|███▉      | 204/520 [13:19<19:22,  3.68s/it]
                                                 
{'loss': 1.2883, 'grad_norm': 0.001420270484786974, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [13:19<19:22,  3.68s/it]
 39%|███▉      | 205/520 [13:23<19:20,  3.69s/it]
                                                 
{'loss': 1.2378, 'grad_norm': 0.001335656461996073, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [13:23<19:20,  3.69s/it]
 40%|███▉      | 206/520 [13:27<19:15,  3.68s/it]
                                                 
{'loss': 1.3272, 'grad_norm': 0.0013482803357300801, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [13:27<19:15,  3.68s/it]
 40%|███▉      | 207/520 [13:30<19:17,  3.70s/it]
                                                 
{'loss': 1.2145, 'grad_norm': 0.0014106883900654528, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [13:30<19:17,  3.70s/it]
 40%|████      | 208/520 [13:34<19:11,  3.69s/it]
                                                 
{'loss': 1.3122, 'grad_norm': 0.0014939485258707041, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [13:34<19:11,  3.69s/it]
 40%|████      | 209/520 [13:38<19:08,  3.69s/it]
                                                 
{'loss': 1.2262, 'grad_norm': 0.0012997233149038558, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:38<19:08,  3.69s/it]
 40%|████      | 210/520 [13:41<19:06,  3.70s/it]
                                                 
{'loss': 1.3023, 'grad_norm': 0.0014173373983993088, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:41<19:06,  3.70s/it]
 41%|████      | 211/520 [13:45<19:00,  3.69s/it]
                                                 
{'loss': 1.3009, 'grad_norm': 0.0013193241890018719, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:45<19:00,  3.69s/it]
 41%|████      | 212/520 [13:49<18:53,  3.68s/it]
                                                 
{'loss': 1.29, 'grad_norm': 0.0013408793873886295, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:49<18:53,  3.68s/it]
 41%|████      | 213/520 [13:53<18:51,  3.69s/it]
                                                 
{'loss': 1.2482, 'grad_norm': 0.0015779355016224859, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:53<18:51,  3.69s/it]
 41%|████      | 214/520 [13:56<18:45,  3.68s/it]
                                                 
{'loss': 1.2351, 'grad_norm': 0.0014138446771352201, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:56<18:45,  3.68s/it]
 41%|████▏     | 215/520 [14:00<18:41,  3.68s/it]
                                                 
{'loss': 1.1663, 'grad_norm': 0.0013006286489445833, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [14:00<18:41,  3.68s/it]
 42%|████▏     | 216/520 [14:04<18:37,  3.68s/it]
                                                 
{'loss': 1.1559, 'grad_norm': 0.0013422952719435428, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [14:04<18:37,  3.68s/it]
 42%|████▏     | 217/520 [14:07<18:36,  3.68s/it]
                                                 
{'loss': 1.2793, 'grad_norm': 0.0013631346744358952, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [14:07<18:36,  3.68s/it]
 42%|████▏     | 218/520 [14:11<18:37,  3.70s/it]
                                                 
{'loss': 1.2671, 'grad_norm': 0.0014468120470549873, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [14:11<18:37,  3.70s/it]
 42%|████▏     | 219/520 [14:15<18:30,  3.69s/it]
                                                 
{'loss': 1.2642, 'grad_norm': 0.001216242720135587, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [14:15<18:30,  3.69s/it]
 42%|████▏     | 220/520 [14:18<18:25,  3.68s/it]
                                                 
{'loss': 1.2131, 'grad_norm': 0.0013120531857130958, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [14:18<18:25,  3.68s/it]
 42%|████▎     | 221/520 [14:22<18:21,  3.68s/it]
                                                 
{'loss': 1.2725, 'grad_norm': 0.0014196329455767106, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [14:22<18:21,  3.68s/it]
 43%|████▎     | 222/520 [14:26<18:19,  3.69s/it]
                                                 
{'loss': 1.2042, 'grad_norm': 0.0013465528889959018, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [14:26<18:19,  3.69s/it]
 43%|████▎     | 223/520 [14:29<18:13,  3.68s/it]
                                                 
{'loss': 1.2023, 'grad_norm': 0.0012848997092014508, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [14:29<18:13,  3.68s/it]
 43%|████▎     | 224/520 [14:33<18:12,  3.69s/it]
                                                 
{'loss': 1.2885, 'grad_norm': 0.0013980723559765531, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [14:33<18:12,  3.69s/it]
 43%|████▎     | 225/520 [14:37<18:09,  3.69s/it]
                                                 
{'loss': 1.2076, 'grad_norm': 0.0013370316224778848, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:37<18:09,  3.69s/it]
 43%|████▎     | 226/520 [14:40<18:05,  3.69s/it]
                                                 
{'loss': 1.3107, 'grad_norm': 0.0013204096905779636, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:40<18:05,  3.69s/it]
 44%|████▎     | 227/520 [14:44<17:58,  3.68s/it]
                                                 
{'loss': 1.2942, 'grad_norm': 0.0012453916870752257, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:44<17:58,  3.68s/it]
 44%|████▍     | 228/520 [14:48<17:54,  3.68s/it]
                                                 
{'loss': 1.3256, 'grad_norm': 0.0014473725614444846, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:48<17:54,  3.68s/it]
 44%|████▍     | 229/520 [14:52<17:56,  3.70s/it]
                                                 
{'loss': 1.2647, 'grad_norm': 0.0012245523678887508, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:52<17:56,  3.70s/it]
 44%|████▍     | 230/520 [14:55<17:50,  3.69s/it]
                                                 
{'loss': 1.157, 'grad_norm': 0.00130622252095614, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:55<17:50,  3.69s/it]
 44%|████▍     | 231/520 [14:59<17:43,  3.68s/it]
                                                 
{'loss': 1.2157, 'grad_norm': 0.0012406919561111095, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:59<17:43,  3.68s/it]
 45%|████▍     | 232/520 [15:03<17:39,  3.68s/it]
                                                 
{'loss': 1.3561, 'grad_norm': 0.0014781461122434227, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [15:03<17:39,  3.68s/it]
 45%|████▍     | 233/520 [15:06<17:33,  3.67s/it]
                                                 
{'loss': 1.251, 'grad_norm': 0.0015967084427205545, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [15:06<17:33,  3.67s/it]
 45%|████▌     | 234/520 [15:10<17:29,  3.67s/it]
                                                 
{'loss': 1.1703, 'grad_norm': 0.0014203468366004936, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [15:10<17:29,  3.67s/it]
 45%|████▌     | 235/520 [15:14<17:25,  3.67s/it]
                                                 
{'loss': 1.2222, 'grad_norm': 0.0014069156343524982, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [15:14<17:25,  3.67s/it]
 45%|████▌     | 236/520 [15:17<17:23,  3.67s/it]
                                                 
{'loss': 1.3039, 'grad_norm': 0.0012568099006555174, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [15:17<17:23,  3.67s/it]
 46%|████▌     | 237/520 [15:21<17:18,  3.67s/it]
                                                 
{'loss': 1.2995, 'grad_norm': 0.0013143733666693327, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [15:21<17:18,  3.67s/it]
 46%|████▌     | 238/520 [15:25<17:14,  3.67s/it]
                                                 
{'loss': 1.2297, 'grad_norm': 0.0014122096569183843, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [15:25<17:14,  3.67s/it]
 46%|████▌     | 239/520 [15:28<17:12,  3.68s/it]
                                                 
{'loss': 1.3121, 'grad_norm': 0.0013698797093346711, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [15:28<17:12,  3.68s/it]
 46%|████▌     | 240/520 [15:32<17:09,  3.68s/it]
                                                 
{'loss': 1.1123, 'grad_norm': 0.0012702522581107955, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:32<17:09,  3.68s/it]
 46%|████▋     | 241/520 [15:36<17:04,  3.67s/it]
                                                 
{'loss': 1.201, 'grad_norm': 0.0013295406511225013, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:36<17:04,  3.67s/it]
 47%|████▋     | 242/520 [15:39<17:01,  3.67s/it]
                                                 
{'loss': 1.2117, 'grad_norm': 0.0012627623317803618, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:39<17:01,  3.67s/it]
 47%|████▋     | 243/520 [15:43<17:03,  3.69s/it]
                                                 
{'loss': 1.208, 'grad_norm': 0.0013250436229441875, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:43<17:03,  3.69s/it]
 47%|████▋     | 244/520 [15:47<16:57,  3.69s/it]
                                                 
{'loss': 1.3192, 'grad_norm': 0.0013611643619099363, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:47<16:57,  3.69s/it]
 47%|████▋     | 245/520 [15:50<16:52,  3.68s/it]
                                                 
{'loss': 1.1851, 'grad_norm': 0.0014174687911008635, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:50<16:52,  3.68s/it]
 47%|████▋     | 246/520 [15:54<16:49,  3.68s/it]
                                                 
{'loss': 1.3363, 'grad_norm': 0.001361007097521925, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:54<16:49,  3.68s/it]
 48%|████▊     | 247/520 [15:58<16:42,  3.67s/it]
                                                 
{'loss': 1.3638, 'grad_norm': 0.0013657951575159232, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:58<16:42,  3.67s/it]
 48%|████▊     | 248/520 [16:01<16:39,  3.68s/it]
                                                 
{'loss': 1.1916, 'grad_norm': 0.0013710653738743831, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [16:01<16:39,  3.68s/it]
 48%|████▊     | 249/520 [16:05<16:34,  3.67s/it]
                                                 
{'loss': 1.2876, 'grad_norm': 0.0013312621452288103, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [16:05<16:34,  3.67s/it]
 48%|████▊     | 250/520 [16:09<16:31,  3.67s/it]
                                                 
{'loss': 1.2246, 'grad_norm': 0.0014333107624794315, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [16:09<16:31,  3.67s/it]
 48%|████▊     | 251/520 [16:12<16:28,  3.68s/it]
                                                 
{'loss': 1.2861, 'grad_norm': 0.0012748049214666702, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [16:12<16:28,  3.68s/it]
 48%|████▊     | 252/520 [16:16<16:30,  3.69s/it]
                                                 
{'loss': 1.2431, 'grad_norm': 0.0012796287764146432, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [16:16<16:30,  3.69s/it]
 49%|████▊     | 253/520 [16:20<16:27,  3.70s/it]
                                                 
{'loss': 1.2849, 'grad_norm': 0.0014620208830068541, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [16:20<16:27,  3.70s/it]
 49%|████▉     | 254/520 [16:24<16:38,  3.76s/it]
                                                 
{'loss': 1.2186, 'grad_norm': 0.0012565423896619, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [16:24<16:38,  3.76s/it]
 49%|████▉     | 255/520 [16:28<16:52,  3.82s/it]
                                                 
{'loss': 1.2219, 'grad_norm': 0.001457360665954072, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [16:28<16:52,  3.82s/it]
 49%|████▉     | 256/520 [16:32<16:58,  3.86s/it]
                                                 
{'loss': 1.2721, 'grad_norm': 0.0014038468948342681, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:32<16:58,  3.86s/it]
 49%|████▉     | 257/520 [16:36<16:58,  3.87s/it]
                                                 
{'loss': 1.2544, 'grad_norm': 0.0013715217660027574, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:36<16:58,  3.87s/it]
 50%|████▉     | 258/520 [16:39<16:58,  3.89s/it]
                                                 
{'loss': 1.2645, 'grad_norm': 0.0012053579131581142, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:39<16:58,  3.89s/it]
 50%|████▉     | 259/520 [16:43<16:56,  3.90s/it]
                                                 
{'loss': 1.3427, 'grad_norm': 0.0015636979079921096, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:43<16:56,  3.90s/it]
 50%|█████     | 260/520 [16:47<16:56,  3.91s/it]
                                                 
{'loss': 1.2849, 'grad_norm': 0.001304632221809503, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:47<16:56,  3.91s/it]
 50%|█████     | 261/520 [16:51<16:56,  3.93s/it]
                                                 
{'loss': 1.2325, 'grad_norm': 0.0013263123158421403, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:51<16:56,  3.93s/it]
 50%|█████     | 262/520 [16:55<16:51,  3.92s/it]
                                                 
{'loss': 1.1888, 'grad_norm': 0.0012990212239000955, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:55<16:51,  3.92s/it]
 51%|█████     | 263/520 [16:59<16:48,  3.93s/it]
                                                 
{'loss': 1.2476, 'grad_norm': 0.0013462048718175777, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:59<16:48,  3.93s/it]
 51%|█████     | 264/520 [17:03<16:47,  3.93s/it]
                                                 
{'loss': 1.2899, 'grad_norm': 0.001349844308787341, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [17:03<16:47,  3.93s/it]
 51%|█████     | 265/520 [17:07<16:42,  3.93s/it]
                                                 
{'loss': 1.2068, 'grad_norm': 0.0014870404691147902, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [17:07<16:42,  3.93s/it]
 51%|█████     | 266/520 [17:11<16:25,  3.88s/it]
                                                 
{'loss': 1.0731, 'grad_norm': 0.0012365281252980913, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [17:11<16:25,  3.88s/it]
 51%|█████▏    | 267/520 [17:14<16:09,  3.83s/it]
                                                 
{'loss': 1.2019, 'grad_norm': 0.00133436730645543, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [17:14<16:09,  3.83s/it]
 52%|█████▏    | 268/520 [17:18<15:54,  3.79s/it]
                                                 
{'loss': 1.3611, 'grad_norm': 0.0018840780705234798, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [17:18<15:54,  3.79s/it]
 52%|█████▏    | 269/520 [17:22<15:44,  3.76s/it]
                                                 
{'loss': 1.3038, 'grad_norm': 0.00136215315348964, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [17:22<15:44,  3.76s/it]
 52%|█████▏    | 270/520 [17:26<15:35,  3.74s/it]
                                                 
{'loss': 1.1741, 'grad_norm': 0.0012301750976536291, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [17:26<15:35,  3.74s/it]
 52%|█████▏    | 271/520 [17:29<15:25,  3.72s/it]
                                                 
{'loss': 1.2901, 'grad_norm': 0.0013463355764944877, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [17:29<15:25,  3.72s/it]
 52%|█████▏    | 272/520 [17:33<15:20,  3.71s/it]
                                                 
{'loss': 1.1953, 'grad_norm': 0.0013091634858904124, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:33<15:20,  3.71s/it]
 52%|█████▎    | 273/520 [17:37<15:13,  3.70s/it]
                                                 
{'loss': 1.3127, 'grad_norm': 0.0013313030380501792, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:37<15:13,  3.70s/it]
 53%|█████▎    | 274/520 [17:40<15:06,  3.69s/it]
                                                 
{'loss': 1.2648, 'grad_norm': 0.001464754509604515, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:40<15:06,  3.69s/it]
 53%|█████▎    | 275/520 [17:44<15:02,  3.68s/it]
                                                 
{'loss': 1.2124, 'grad_norm': 0.0014753443588136003, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:44<15:02,  3.68s/it]
 53%|█████▎    | 276/520 [17:48<14:59,  3.69s/it]
                                                 
{'loss': 1.2775, 'grad_norm': 0.0016408603995751875, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:48<14:59,  3.69s/it]
 53%|█████▎    | 277/520 [17:51<14:55,  3.69s/it]
                                                 
{'loss': 1.3111, 'grad_norm': 0.0012394961130134305, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:51<14:55,  3.69s/it]
 53%|█████▎    | 278/520 [17:55<14:51,  3.68s/it]
                                                 
{'loss': 1.1569, 'grad_norm': 0.001188395039913709, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:55<14:51,  3.68s/it]
 54%|█████▎    | 279/520 [17:59<14:50,  3.69s/it]
                                                 
{'loss': 1.1963, 'grad_norm': 0.0015705460330058124, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:59<14:50,  3.69s/it]
 54%|█████▍    | 280/520 [18:02<14:49,  3.71s/it]
                                                 
{'loss': 1.2064, 'grad_norm': 0.0014859918568985823, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [18:02<14:49,  3.71s/it]
 54%|█████▍    | 281/520 [18:06<14:44,  3.70s/it]
                                                 
{'loss': 1.3051, 'grad_norm': 0.0014459139326717674, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [18:06<14:44,  3.70s/it]
 54%|█████▍    | 282/520 [18:10<14:40,  3.70s/it]
                                                 
{'loss': 1.1729, 'grad_norm': 0.0012289721455118832, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [18:10<14:40,  3.70s/it]
 54%|█████▍    | 283/520 [18:13<14:37,  3.70s/it]
                                                 
{'loss': 1.3214, 'grad_norm': 0.001465303959517262, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [18:13<14:37,  3.70s/it]
 55%|█████▍    | 284/520 [18:17<14:33,  3.70s/it]
                                                 
{'loss': 1.1923, 'grad_norm': 0.0013752900725109765, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [18:17<14:33,  3.70s/it]
 55%|█████▍    | 285/520 [18:21<14:29,  3.70s/it]
                                                 
{'loss': 1.1949, 'grad_norm': 0.0013207955792789354, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [18:21<14:29,  3.70s/it]
 55%|█████▌    | 286/520 [18:25<14:23,  3.69s/it]
                                                 
{'loss': 1.0783, 'grad_norm': 0.0015102003342448204, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [18:25<14:23,  3.69s/it]
 55%|█████▌    | 287/520 [18:28<14:21,  3.70s/it]
                                                 
{'loss': 1.3027, 'grad_norm': 0.0013472097262300906, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [18:28<14:21,  3.70s/it]
 55%|█████▌    | 288/520 [18:32<14:19,  3.70s/it]
                                                 
{'loss': 1.3397, 'grad_norm': 0.001331311241643678, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:32<14:19,  3.70s/it]
 56%|█████▌    | 289/520 [18:36<14:25,  3.75s/it]
                                                 
{'loss': 1.2116, 'grad_norm': 0.001226607551116687, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:36<14:25,  3.75s/it]
 56%|█████▌    | 290/520 [18:40<14:30,  3.79s/it]
                                                 
{'loss': 1.1342, 'grad_norm': 0.0012747287109011716, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:40<14:30,  3.79s/it]
 56%|█████▌    | 291/520 [18:44<14:35,  3.82s/it]
                                                 
{'loss': 1.1834, 'grad_norm': 0.0013355578992054165, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:44<14:35,  3.82s/it]
 56%|█████▌    | 292/520 [18:47<14:34,  3.84s/it]
                                                 
{'loss': 1.2388, 'grad_norm': 0.0012974506216163952, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:47<14:34,  3.84s/it]
 56%|█████▋    | 293/520 [18:51<14:35,  3.85s/it]
                                                 
{'loss': 1.1811, 'grad_norm': 0.0013851275899409055, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:51<14:35,  3.85s/it]
 57%|█████▋    | 294/520 [18:55<14:32,  3.86s/it]
                                                 
{'loss': 1.2079, 'grad_norm': 0.0014793969525456405, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:55<14:32,  3.86s/it]
 57%|█████▋    | 295/520 [18:59<14:32,  3.88s/it]
                                                 
{'loss': 1.2386, 'grad_norm': 0.0013333755079689038, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:59<14:32,  3.88s/it]
 57%|█████▋    | 296/520 [19:03<14:27,  3.87s/it]
                                                 
{'loss': 1.1537, 'grad_norm': 0.0014212432894004237, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [19:03<14:27,  3.87s/it]
 57%|█████▋    | 297/520 [19:07<14:24,  3.88s/it]
                                                 
{'loss': 1.2822, 'grad_norm': 0.0014760708397199662, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [19:07<14:24,  3.88s/it]
 57%|█████▋    | 298/520 [19:11<14:18,  3.87s/it]
                                                 
{'loss': 1.2474, 'grad_norm': 0.0012657998388807992, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [19:11<14:18,  3.87s/it]
 57%|█████▊    | 299/520 [19:15<14:15,  3.87s/it]
                                                 
{'loss': 1.2677, 'grad_norm': 0.0012469463338156909, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [19:15<14:15,  3.87s/it]
 58%|█████▊    | 300/520 [19:19<14:11,  3.87s/it]
                                                 
{'loss': 1.2974, 'grad_norm': 0.0013191963668587846, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [19:19<14:11,  3.87s/it]
 58%|█████▊    | 301/520 [19:22<14:08,  3.87s/it]
                                                 
{'loss': 1.274, 'grad_norm': 0.0013140580568227105, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [19:22<14:08,  3.87s/it]
 58%|█████▊    | 302/520 [19:26<14:03,  3.87s/it]
                                                 
{'loss': 1.282, 'grad_norm': 0.0013304208153875022, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [19:26<14:03,  3.87s/it]
 58%|█████▊    | 303/520 [19:30<14:00,  3.87s/it]
                                                 
{'loss': 1.2039, 'grad_norm': 0.001514971844218521, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [19:30<14:00,  3.87s/it]
 58%|█████▊    | 304/520 [19:34<14:19,  3.98s/it]
                                                 
{'loss': 1.1841, 'grad_norm': 0.0014946542819492808, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:34<14:19,  3.98s/it]
 59%|█████▊    | 305/520 [19:38<14:05,  3.93s/it]
                                                 
{'loss': 1.31, 'grad_norm': 0.0015030122603261799, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:38<14:05,  3.93s/it]
 59%|█████▉    | 306/520 [19:42<13:54,  3.90s/it]
                                                 
{'loss': 1.254, 'grad_norm': 0.0013318291532786737, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:42<13:54,  3.90s/it]
 59%|█████▉    | 307/520 [19:46<13:44,  3.87s/it]
                                                 
{'loss': 1.1868, 'grad_norm': 0.0012408112547066608, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:46<13:44,  3.87s/it]
 59%|█████▉    | 308/520 [19:50<13:35,  3.85s/it]
                                                 
{'loss': 1.3064, 'grad_norm': 0.0014498512241624168, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:50<13:35,  3.85s/it]
 59%|█████▉    | 309/520 [19:53<13:30,  3.84s/it]
                                                 
{'loss': 1.1945, 'grad_norm': 0.001256313747008845, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:53<13:30,  3.84s/it]
 60%|█████▉    | 310/520 [19:57<13:27,  3.84s/it]
                                                 
{'loss': 1.175, 'grad_norm': 0.0013250609674751272, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:57<13:27,  3.84s/it]
 60%|█████▉    | 311/520 [20:01<13:22,  3.84s/it]
                                                 
{'loss': 1.1418, 'grad_norm': 0.0013261013245104895, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [20:01<13:22,  3.84s/it]
 60%|██████    | 312/520 [20:05<13:15,  3.82s/it]
                                                 
{'loss': 1.1356, 'grad_norm': 0.0014688250246866696, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [20:05<13:15,  3.82s/it]
 60%|██████    | 313/520 [20:09<13:03,  3.79s/it]
                                                 
{'loss': 1.1238, 'grad_norm': 0.0011864133034198432, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [20:09<13:03,  3.79s/it]
 60%|██████    | 314/520 [20:13<13:24,  3.90s/it]
                                                 
{'loss': 1.1641, 'grad_norm': 0.0012067646376146288, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [20:13<13:24,  3.90s/it]
 61%|██████    | 315/520 [20:16<13:05,  3.83s/it]
                                                 
{'loss': 1.2308, 'grad_norm': 0.0017227650588142616, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [20:16<13:05,  3.83s/it]
 61%|██████    | 316/520 [20:21<13:19,  3.92s/it]
                                                 
{'loss': 1.1413, 'grad_norm': 0.0015094619673083585, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [20:21<13:19,  3.92s/it]
 61%|██████    | 317/520 [20:24<12:57,  3.83s/it]
                                                 
{'loss': 1.1584, 'grad_norm': 0.001221626093329796, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [20:24<12:57,  3.83s/it]
 61%|██████    | 318/520 [20:28<12:43,  3.78s/it]
                                                 
{'loss': 1.2692, 'grad_norm': 0.001465001268880618, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [20:28<12:43,  3.78s/it]
 61%|██████▏   | 319/520 [20:32<12:55,  3.86s/it]
                                                 
{'loss': 1.1422, 'grad_norm': 0.0012444279840545505, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [20:32<12:55,  3.86s/it]
 62%|██████▏   | 320/520 [20:36<12:38,  3.79s/it]
                                                 
{'loss': 1.0909, 'grad_norm': 0.0013346024395228277, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:36<12:38,  3.79s/it]
 62%|██████▏   | 321/520 [20:39<12:29,  3.76s/it]
                                                 
{'loss': 1.2859, 'grad_norm': 0.0014449193216112015, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:39<12:29,  3.76s/it]
 62%|██████▏   | 322/520 [20:43<12:22,  3.75s/it]
                                                 
{'loss': 1.1262, 'grad_norm': 0.0012481415158091549, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:43<12:22,  3.75s/it]
 62%|██████▏   | 323/520 [20:47<12:15,  3.73s/it]
                                                 
{'loss': 1.1973, 'grad_norm': 0.0012961793757169326, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:47<12:15,  3.73s/it]
 62%|██████▏   | 324/520 [20:50<12:07,  3.71s/it]
                                                 
{'loss': 1.2232, 'grad_norm': 0.001332404485008512, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:50<12:07,  3.71s/it]
 62%|██████▎   | 325/520 [20:54<12:03,  3.71s/it]
                                                 
{'loss': 1.2301, 'grad_norm': 0.001342707147251463, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:54<12:03,  3.71s/it]
 63%|██████▎   | 326/520 [20:58<11:58,  3.70s/it]
                                                 
{'loss': 1.2215, 'grad_norm': 0.0013797642901817767, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:58<11:58,  3.70s/it]
 63%|██████▎   | 327/520 [21:01<11:54,  3.70s/it]
                                                 
{'loss': 1.2438, 'grad_norm': 0.001435663873784387, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [21:01<11:54,  3.70s/it]
 63%|██████▎   | 328/520 [21:05<11:49,  3.70s/it]
                                                 
{'loss': 1.2686, 'grad_norm': 0.0013646211831662791, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [21:05<11:49,  3.70s/it]
 63%|██████▎   | 329/520 [21:09<11:47,  3.70s/it]
                                                 
{'loss': 1.1451, 'grad_norm': 0.0011562425090150936, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [21:09<11:47,  3.70s/it]
 63%|██████▎   | 330/520 [21:12<11:42,  3.70s/it]
                                                 
{'loss': 1.2135, 'grad_norm': 0.0012189736994147595, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [21:12<11:42,  3.70s/it]
 64%|██████▎   | 331/520 [21:16<11:37,  3.69s/it]
                                                 
{'loss': 1.1749, 'grad_norm': 0.001262597362790727, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [21:16<11:37,  3.69s/it]
 64%|██████▍   | 332/520 [21:20<11:43,  3.74s/it]
                                                 
{'loss': 1.2684, 'grad_norm': 0.0012631428963529932, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [21:20<11:43,  3.74s/it]
 64%|██████▍   | 333/520 [21:24<11:37,  3.73s/it]
                                                 
{'loss': 1.3184, 'grad_norm': 0.0013644078517652523, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [21:24<11:37,  3.73s/it]
 64%|██████▍   | 334/520 [21:27<11:31,  3.72s/it]
                                                 
{'loss': 1.2249, 'grad_norm': 0.0013782732670792738, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [21:27<11:31,  3.72s/it]
 64%|██████▍   | 335/520 [21:31<11:25,  3.71s/it]
                                                 
{'loss': 1.2256, 'grad_norm': 0.0012260754833876223, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [21:31<11:25,  3.71s/it]
 65%|██████▍   | 336/520 [21:35<11:20,  3.70s/it]
                                                 
{'loss': 1.1228, 'grad_norm': 0.0014603292305307774, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:35<11:20,  3.70s/it]
 65%|██████▍   | 337/520 [21:38<11:15,  3.69s/it]
                                                 
{'loss': 1.1059, 'grad_norm': 0.001269390541725017, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:38<11:15,  3.69s/it]
 65%|██████▌   | 338/520 [21:42<11:12,  3.69s/it]
                                                 
{'loss': 1.2256, 'grad_norm': 0.0013125759899973834, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:42<11:12,  3.69s/it]
 65%|██████▌   | 339/520 [21:46<11:07,  3.69s/it]
                                                 
{'loss': 1.1732, 'grad_norm': 0.0012878443756152483, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:46<11:07,  3.69s/it]
 65%|██████▌   | 340/520 [21:49<11:03,  3.69s/it]
                                                 
{'loss': 1.1651, 'grad_norm': 0.001279072005288677, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:49<11:03,  3.69s/it]
 66%|██████▌   | 341/520 [21:53<11:00,  3.69s/it]
                                                 
{'loss': 1.1863, 'grad_norm': 0.0013632962091374961, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:53<11:00,  3.69s/it]
 66%|██████▌   | 342/520 [21:57<10:55,  3.68s/it]
                                                 
{'loss': 1.2341, 'grad_norm': 0.0015417530885308158, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:57<10:55,  3.68s/it]
 66%|██████▌   | 343/520 [22:01<10:53,  3.69s/it]
                                                 
{'loss': 1.1863, 'grad_norm': 0.0012184390389758445, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [22:01<10:53,  3.69s/it]
 66%|██████▌   | 344/520 [22:04<10:50,  3.69s/it]
                                                 
{'loss': 1.1413, 'grad_norm': 0.0012568302665183632, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [22:04<10:50,  3.69s/it]
 66%|██████▋   | 345/520 [22:08<10:46,  3.69s/it]
                                                 
{'loss': 1.2532, 'grad_norm': 0.0014048998909958414, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [22:08<10:46,  3.69s/it]
 67%|██████▋   | 346/520 [22:12<10:41,  3.69s/it]
                                                 
{'loss': 1.2042, 'grad_norm': 0.0012365319168710274, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [22:12<10:41,  3.69s/it]
 67%|██████▋   | 347/520 [22:15<10:37,  3.68s/it]
                                                 
{'loss': 1.1589, 'grad_norm': 0.0012262526972204595, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [22:15<10:37,  3.68s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [22:19<10:38,  3.71s/it]
                                                 
{'loss': 1.1137, 'grad_norm': 0.001482747903096459, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [22:19<10:38,  3.71s/it]
 67%|██████▋   | 349/520 [22:23<10:38,  3.73s/it]
                                                 
{'loss': 1.154, 'grad_norm': 0.0012964699105130527, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [22:23<10:38,  3.73s/it]
 67%|██████▋   | 350/520 [22:27<10:34,  3.73s/it]
                                                 
{'loss': 1.1986, 'grad_norm': 0.001411436459793438, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [22:27<10:34,  3.73s/it]
 68%|██████▊   | 351/520 [22:30<10:38,  3.78s/it]
                                                 
{'loss': 1.1028, 'grad_norm': 0.001213359600336043, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [22:30<10:38,  3.78s/it]
 68%|██████▊   | 352/520 [22:34<10:39,  3.81s/it]
                                                 
{'loss': 1.2303, 'grad_norm': 0.0012994110955172318, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [22:34<10:39,  3.81s/it]
 68%|██████▊   | 353/520 [22:38<10:42,  3.85s/it]
                                                 
{'loss': 1.1579, 'grad_norm': 0.0011146371888949014, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:38<10:42,  3.85s/it]
 68%|██████▊   | 354/520 [22:42<10:33,  3.82s/it]
                                                 
{'loss': 1.2757, 'grad_norm': 0.0011989059639271218, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:42<10:33,  3.82s/it]
 68%|██████▊   | 355/520 [22:46<10:24,  3.78s/it]
                                                 
{'loss': 1.1686, 'grad_norm': 0.0013419957499433883, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:46<10:24,  3.78s/it]
 68%|██████▊   | 356/520 [22:50<10:27,  3.83s/it]
                                                 
{'loss': 1.1676, 'grad_norm': 0.0013424497918719228, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:50<10:27,  3.83s/it]
 69%|██████▊   | 357/520 [22:54<10:26,  3.85s/it]
                                                 
{'loss': 1.1993, 'grad_norm': 0.0012642272284040252, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:54<10:26,  3.85s/it]
 69%|██████▉   | 358/520 [22:57<10:24,  3.86s/it]
                                                 
{'loss': 1.1333, 'grad_norm': 0.0012532225380013995, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:57<10:24,  3.86s/it]
 69%|██████▉   | 359/520 [23:01<10:22,  3.87s/it]
                                                 
{'loss': 1.2058, 'grad_norm': 0.00137622979061854, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [23:01<10:22,  3.87s/it]
 69%|██████▉   | 360/520 [23:05<10:12,  3.83s/it]
                                                 
{'loss': 1.217, 'grad_norm': 0.0013276362756085578, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [23:05<10:12,  3.83s/it]
 69%|██████▉   | 361/520 [23:09<10:03,  3.79s/it]
                                                 
{'loss': 1.2261, 'grad_norm': 0.0012247016686520956, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [23:09<10:03,  3.79s/it]
 70%|██████▉   | 362/520 [23:12<09:54,  3.76s/it]
                                                 
{'loss': 1.186, 'grad_norm': 0.0013855252967218234, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [23:12<09:54,  3.76s/it]
 70%|██████▉   | 363/520 [23:16<09:45,  3.73s/it]
                                                 
{'loss': 1.2054, 'grad_norm': 0.0012729472511586251, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [23:16<09:45,  3.73s/it]
 70%|███████   | 364/520 [23:20<09:40,  3.72s/it]
                                                 
{'loss': 1.2491, 'grad_norm': 0.0012759079823110105, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [23:20<09:40,  3.72s/it]
 70%|███████   | 365/520 [23:23<09:33,  3.70s/it]
                                                 
{'loss': 1.2634, 'grad_norm': 0.0013836007956427052, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [23:23<09:33,  3.70s/it]
 70%|███████   | 366/520 [23:27<09:27,  3.69s/it]
                                                 
{'loss': 1.2208, 'grad_norm': 0.001363186296602676, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [23:27<09:27,  3.69s/it]
 71%|███████   | 367/520 [23:31<09:23,  3.68s/it]
                                                 
{'loss': 1.2212, 'grad_norm': 0.0012977803251371127, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [23:31<09:23,  3.68s/it]
 71%|███████   | 368/520 [23:34<09:19,  3.68s/it]
                                                 
{'loss': 1.0752, 'grad_norm': 0.0014725467068658334, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [23:35<09:19,  3.68s/it]
 71%|███████   | 369/520 [23:38<09:15,  3.68s/it]
                                                 
{'loss': 1.2002, 'grad_norm': 0.0011395450214429347, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:38<09:15,  3.68s/it]
 71%|███████   | 370/520 [23:42<09:12,  3.68s/it]
                                                 
{'loss': 1.1365, 'grad_norm': 0.0013374029519111926, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:42<09:12,  3.68s/it]
 71%|███████▏  | 371/520 [23:45<09:06,  3.67s/it]
                                                 
{'loss': 1.1324, 'grad_norm': 0.0013274930088434471, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:45<09:06,  3.67s/it]
 72%|███████▏  | 372/520 [23:49<09:03,  3.67s/it]
                                                 
{'loss': 1.2765, 'grad_norm': 0.001174337723397528, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:49<09:03,  3.67s/it]
 72%|███████▏  | 373/520 [23:53<09:00,  3.67s/it]
                                                 
{'loss': 1.1673, 'grad_norm': 0.0013903637649770814, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:53<09:00,  3.67s/it]
 72%|███████▏  | 374/520 [23:57<08:56,  3.67s/it]
                                                 
{'loss': 1.2205, 'grad_norm': 0.0012916245405580234, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:57<08:56,  3.67s/it]
 72%|███████▏  | 375/520 [24:00<08:53,  3.68s/it]
                                                 
{'loss': 1.1394, 'grad_norm': 0.0013204629470047204, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [24:00<08:53,  3.68s/it]
 72%|███████▏  | 376/520 [24:04<08:50,  3.68s/it]
                                                 
{'loss': 1.2537, 'grad_norm': 0.0012954946468706552, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [24:04<08:50,  3.68s/it]
 72%|███████▎  | 377/520 [24:08<08:45,  3.68s/it]
                                                 
{'loss': 1.1818, 'grad_norm': 0.0012803289834102496, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [24:08<08:45,  3.68s/it]
 73%|███████▎  | 378/520 [24:11<08:42,  3.68s/it]
                                                 
{'loss': 1.241, 'grad_norm': 0.0012846817901528507, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [24:11<08:42,  3.68s/it]
 73%|███████▎  | 379/520 [24:15<08:38,  3.68s/it]
                                                 
{'loss': 1.2191, 'grad_norm': 0.0012088902373238928, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [24:15<08:38,  3.68s/it]
 73%|███████▎  | 380/520 [24:19<08:34,  3.67s/it]
                                                 
{'loss': 1.2512, 'grad_norm': 0.0013068623230976706, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [24:19<08:34,  3.67s/it]
 73%|███████▎  | 381/520 [24:22<08:29,  3.66s/it]
                                                 
{'loss': 1.2222, 'grad_norm': 0.0012675922467241, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [24:22<08:29,  3.66s/it]
 73%|███████▎  | 382/520 [24:26<08:25,  3.66s/it]
                                                 
{'loss': 1.2155, 'grad_norm': 0.0012520251851478475, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [24:26<08:25,  3.66s/it]
 74%|███████▎  | 383/520 [24:30<08:23,  3.68s/it]
                                                 
{'loss': 1.0571, 'grad_norm': 0.001357684520878418, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [24:30<08:23,  3.68s/it]
 74%|███████▍  | 384/520 [24:33<08:19,  3.67s/it]
                                                 
{'loss': 1.2611, 'grad_norm': 0.0011560214760910693, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [24:33<08:19,  3.67s/it]
 74%|███████▍  | 385/520 [24:37<08:15,  3.67s/it]
                                                 
{'loss': 1.1969, 'grad_norm': 0.0011839835052746637, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:37<08:15,  3.67s/it]
 74%|███████▍  | 386/520 [24:41<08:11,  3.67s/it]
                                                 
{'loss': 1.1499, 'grad_norm': 0.001105547417297014, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:41<08:11,  3.67s/it]
 74%|███████▍  | 387/520 [24:44<08:09,  3.68s/it]
                                                 
{'loss': 1.2757, 'grad_norm': 0.0012348155056305303, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:44<08:09,  3.68s/it]
 75%|███████▍  | 388/520 [24:48<08:05,  3.68s/it]
                                                 
{'loss': 1.1037, 'grad_norm': 0.0011988756539280494, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:48<08:05,  3.68s/it]
 75%|███████▍  | 389/520 [24:52<08:05,  3.71s/it]
                                                 
{'loss': 1.1518, 'grad_norm': 0.0014134735135724298, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}
+
 75%|███████▍  | 389/520 [24:52<08:05,  3.71s/it]
 75%|███████▌  | 390/520 [24:56<08:07,  3.75s/it]
                                                 
{'loss': 1.2198, 'grad_norm': 0.0012211613812093806, 'learning_rate': 0.031074166151605298, 'epoch': 0.75}
+
 75%|███████▌  | 390/520 [24:56<08:07,  3.75s/it]
 75%|███████▌  | 391/520 [24:59<08:07,  3.78s/it]
                                                 
{'loss': 1.2949, 'grad_norm': 0.0013506367640879785, 'learning_rate': 0.030623894397145836, 'epoch': 0.75}
+
 75%|███████▌  | 391/520 [24:59<08:07,  3.78s/it]
 75%|███████▌  | 392/520 [25:03<08:07,  3.81s/it]
                                                 
{'loss': 1.1099, 'grad_norm': 0.0012372151753158559, 'learning_rate': 0.03017631819139273, 'epoch': 0.75}
+
 75%|███████▌  | 392/520 [25:03<08:07,  3.81s/it]
 76%|███████▌  | 393/520 [25:07<08:04,  3.81s/it]
                                                 
{'loss': 1.1166, 'grad_norm': 0.0011108408000762057, 'learning_rate': 0.029731454924533086, 'epoch': 0.76}
+
 76%|███████▌  | 393/520 [25:07<08:04,  3.81s/it]
 76%|███████▌  | 394/520 [25:11<08:00,  3.81s/it]
                                                 
{'loss': 1.1677, 'grad_norm': 0.0014169296400873865, 'learning_rate': 0.029289321881345254, 'epoch': 0.76}
+
 76%|███████▌  | 394/520 [25:11<08:00,  3.81s/it]
 76%|███████▌  | 395/520 [25:15<07:59,  3.84s/it]
                                                 
{'loss': 1.1371, 'grad_norm': 0.0013716556263643532, 'learning_rate': 0.028849936240527008, 'epoch': 0.76}
+
 76%|███████▌  | 395/520 [25:15<07:59,  3.84s/it]
 76%|███████▌  | 396/520 [25:19<07:54,  3.83s/it]
                                                 
{'loss': 1.22, 'grad_norm': 0.0013829008044011911, 'learning_rate': 0.028413315074028157, 'epoch': 0.76}
+
 76%|███████▌  | 396/520 [25:19<07:54,  3.83s/it]
 76%|███████▋  | 397/520 [25:23<07:53,  3.85s/it]
                                                 
{'loss': 1.1981, 'grad_norm': 0.001218717160782899, 'learning_rate': 0.027979475346387363, 'epoch': 0.76}
+
 76%|███████▋  | 397/520 [25:23<07:53,  3.85s/it]
 77%|███████▋  | 398/520 [25:26<07:49,  3.84s/it]
                                                 
{'loss': 1.1988, 'grad_norm': 0.0013138650139156987, 'learning_rate': 0.027548433914072735, 'epoch': 0.77}
+
 77%|███████▋  | 398/520 [25:26<07:49,  3.84s/it]
 77%|███████▋  | 399/520 [25:30<07:46,  3.86s/it]
                                                 
{'loss': 1.1557, 'grad_norm': 0.0012480449448206963, 'learning_rate': 0.027120207524827168, 'epoch': 0.77}
+
 77%|███████▋  | 399/520 [25:30<07:46,  3.86s/it]
 77%|███████▋  | 400/520 [25:34<07:42,  3.85s/it]
                                                 
{'loss': 1.1887, 'grad_norm': 0.0011936638824038162, 'learning_rate': 0.02669481281701739, 'epoch': 0.77}
+
 77%|███████▋  | 400/520 [25:34<07:42,  3.85s/it]
 77%|███████▋  | 401/520 [25:38<07:38,  3.86s/it]
                                                 
{'loss': 1.0335, 'grad_norm': 0.0014207909864911627, 'learning_rate': 0.026272266318987603, 'epoch': 0.77}
+
 77%|███████▋  | 401/520 [25:38<07:38,  3.86s/it]
 77%|███████▋  | 402/520 [25:42<07:35,  3.86s/it]
                                                 
{'loss': 1.1544, 'grad_norm': 0.0013103923232514634, 'learning_rate': 0.02585258444841733, 'epoch': 0.77}
+
 77%|███████▋  | 402/520 [25:42<07:35,  3.86s/it]
 78%|███████▊  | 403/520 [25:46<07:31,  3.86s/it]
                                                 
{'loss': 1.1794, 'grad_norm': 0.0014125545859681502, 'learning_rate': 0.025435783511683442, 'epoch': 0.78}
+
 78%|███████▊  | 403/520 [25:46<07:31,  3.86s/it]
 78%|███████▊  | 404/520 [25:50<07:26,  3.85s/it]
                                                 
{'loss': 1.0861, 'grad_norm': 0.0015393623682918177, 'learning_rate': 0.02502187970322657, 'epoch': 0.78}
+
 78%|███████▊  | 404/520 [25:50<07:26,  3.85s/it]
 78%|███████▊  | 405/520 [25:53<07:22,  3.85s/it]
                                                 
{'loss': 1.1657, 'grad_norm': 0.0012717751986771858, 'learning_rate': 0.02461088910492202, 'epoch': 0.78}
+
 78%|███████▊  | 405/520 [25:53<07:22,  3.85s/it]
 78%|███████▊  | 406/520 [25:57<07:17,  3.84s/it]
                                                 
{'loss': 1.0873, 'grad_norm': 0.0014520088390855912, 'learning_rate': 0.02420282768545469, 'epoch': 0.78}
+
 78%|███████▊  | 406/520 [25:57<07:17,  3.84s/it]
 78%|███████▊  | 407/520 [26:01<07:14,  3.85s/it]
                                                 
{'loss': 1.2643, 'grad_norm': 0.0013228035890277784, 'learning_rate': 0.02379771129969892, 'epoch': 0.78}
+
 78%|███████▊  | 407/520 [26:01<07:14,  3.85s/it]
 78%|███████▊  | 408/520 [26:05<07:10,  3.84s/it]
                                                 
{'loss': 1.1678, 'grad_norm': 0.0013887714261305035, 'learning_rate': 0.023395555688102213, 'epoch': 0.78}
+
 78%|███████▊  | 408/520 [26:05<07:10,  3.84s/it]
 79%|███████▊  | 409/520 [26:09<07:06,  3.84s/it]
                                                 
{'loss': 1.2862, 'grad_norm': 0.0013723601379723612, 'learning_rate': 0.02299637647607372, 'epoch': 0.79}
+
 79%|███████▊  | 409/520 [26:09<07:06,  3.84s/it]
 79%|███████▉  | 410/520 [26:13<07:02,  3.84s/it]
                                                 
{'loss': 1.0179, 'grad_norm': 0.0012591792437611313, 'learning_rate': 0.022600189173377264, 'epoch': 0.79}
+
 79%|███████▉  | 410/520 [26:13<07:02,  3.84s/it]
 79%|███████▉  | 411/520 [26:16<06:58,  3.84s/it]
                                                 
{'loss': 1.2665, 'grad_norm': 0.001537953828995777, 'learning_rate': 0.022207009173528525, 'epoch': 0.79}
+
 79%|███████▉  | 411/520 [26:16<06:58,  3.84s/it]
 79%|███████▉  | 412/520 [26:20<06:54,  3.84s/it]
                                                 
{'loss': 1.1775, 'grad_norm': 0.0012937842968747044, 'learning_rate': 0.02181685175319702, 'epoch': 0.79}
+
 79%|███████▉  | 412/520 [26:20<06:54,  3.84s/it]
 79%|███████▉  | 413/520 [26:24<06:50,  3.83s/it]
                                                 
{'loss': 1.1803, 'grad_norm': 0.0011774483431981927, 'learning_rate': 0.021429732071612653, 'epoch': 0.79}
+
 79%|███████▉  | 413/520 [26:24<06:50,  3.83s/it]
 80%|███████▉  | 414/520 [26:28<06:46,  3.83s/it]
                                                 
{'loss': 0.993, 'grad_norm': 0.0010743405913557812, 'learning_rate': 0.02104566516997647, 'epoch': 0.8}
+
 80%|███████▉  | 414/520 [26:28<06:46,  3.83s/it]
 80%|███████▉  | 415/520 [26:32<06:41,  3.83s/it]
                                                 
{'loss': 1.1523, 'grad_norm': 0.0011963979391571662, 'learning_rate': 0.020664665970876496, 'epoch': 0.8}
+
 80%|███████▉  | 415/520 [26:32<06:41,  3.83s/it]
 80%|████████  | 416/520 [26:36<06:37,  3.82s/it]
                                                 
{'loss': 1.072, 'grad_norm': 0.001351982522529133, 'learning_rate': 0.020286749277707784, 'epoch': 0.8}
+
 80%|████████  | 416/520 [26:36<06:37,  3.82s/it]
 80%|████████  | 417/520 [26:39<06:32,  3.82s/it]
                                                 
{'loss': 1.2317, 'grad_norm': 0.0013586330007634396, 'learning_rate': 0.019911929774097215, 'epoch': 0.8}
+
 80%|████████  | 417/520 [26:39<06:32,  3.82s/it]
 80%|████████  | 418/520 [26:43<06:29,  3.82s/it]
                                                 
{'loss': 1.2193, 'grad_norm': 0.001265251617094496, 'learning_rate': 0.019540222023333165, 'epoch': 0.8}
+
 80%|████████  | 418/520 [26:43<06:29,  3.82s/it]
 81%|████████  | 419/520 [26:47<06:25,  3.81s/it]
                                                 
{'loss': 1.2112, 'grad_norm': 0.001419776478607778, 'learning_rate': 0.01917164046779948, 'epoch': 0.81}
+
 81%|████████  | 419/520 [26:47<06:25,  3.81s/it]
 81%|████████  | 420/520 [26:51<06:21,  3.82s/it]
                                                 
{'loss': 1.1018, 'grad_norm': 0.001374044810752546, 'learning_rate': 0.018806199428414352, 'epoch': 0.81}
+
 81%|████████  | 420/520 [26:51<06:21,  3.82s/it]
 81%|████████  | 421/520 [26:54<06:12,  3.77s/it]
                                                 
{'loss': 1.0387, 'grad_norm': 0.0015041404439923326, 'learning_rate': 0.018443913104073985, 'epoch': 0.81}
+
 81%|████████  | 421/520 [26:54<06:12,  3.77s/it]
 81%|████████  | 422/520 [26:58<06:06,  3.74s/it]
                                                 
{'loss': 1.1591, 'grad_norm': 0.0013193213443093027, 'learning_rate': 0.01808479557110081, 'epoch': 0.81}
+
 81%|████████  | 422/520 [26:58<06:06,  3.74s/it]
 81%|████████▏ | 423/520 [27:02<06:00,  3.72s/it]
                                                 
{'loss': 1.1378, 'grad_norm': 0.0014499111456476481, 'learning_rate': 0.017728860782696667, 'epoch': 0.81}
+
 81%|████████▏ | 423/520 [27:02<06:00,  3.72s/it]
 82%|████████▏ | 424/520 [27:05<05:56,  3.71s/it]
                                                 
{'loss': 1.2654, 'grad_norm': 0.001335034444326127, 'learning_rate': 0.017376122568400532, 'epoch': 0.82}
+
 82%|████████▏ | 424/520 [27:05<05:56,  3.71s/it]
 82%|████████▏ | 425/520 [27:09<05:51,  3.70s/it]
                                                 
{'loss': 1.1558, 'grad_norm': 0.0012584933953644613, 'learning_rate': 0.017026594633551252, 'epoch': 0.82}
+
 82%|████████▏ | 425/520 [27:09<05:51,  3.70s/it]
 82%|████████▏ | 426/520 [27:13<05:46,  3.69s/it]
                                                 
{'loss': 1.1697, 'grad_norm': 0.0015832389589113688, 'learning_rate': 0.01668029055875512, 'epoch': 0.82}
+
 82%|████████▏ | 426/520 [27:13<05:46,  3.69s/it]
 82%|████████▏ | 427/520 [27:16<05:41,  3.68s/it]
                                                 
{'loss': 1.0848, 'grad_norm': 0.0012288419075211741, 'learning_rate': 0.016337223799358026, 'epoch': 0.82}
+
 82%|████████▏ | 427/520 [27:16<05:41,  3.68s/it]
 82%|████████▏ | 428/520 [27:20<05:38,  3.68s/it]
                                                 
{'loss': 1.0638, 'grad_norm': 0.0013159969702194717, 'learning_rate': 0.01599740768492286, 'epoch': 0.82}
+
 82%|████████▏ | 428/520 [27:20<05:38,  3.68s/it]
 82%|████████▎ | 429/520 [27:24<05:34,  3.68s/it]
                                                 
{'loss': 1.1584, 'grad_norm': 0.0012690448433445901, 'learning_rate': 0.015660855418711452, 'epoch': 0.82}
+
 82%|████████▎ | 429/520 [27:24<05:34,  3.68s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2076 > 2048). Running this sequence through the model will result in indexing errors
+
 83%|████████▎ | 430/520 [27:28<05:31,  3.68s/it]
                                                 
{'loss': 1.1641, 'grad_norm': 0.001201065265607941, 'learning_rate': 0.015327580077171589, 'epoch': 0.83}
+
 83%|████████▎ | 430/520 [27:28<05:31,  3.68s/it]
 83%|████████▎ | 431/520 [27:31<05:27,  3.68s/it]
                                                 
{'loss': 1.1476, 'grad_norm': 0.0013297205886926968, 'learning_rate': 0.014997594609429088, 'epoch': 0.83}
+
 83%|████████▎ | 431/520 [27:31<05:27,  3.68s/it]
 83%|████████▎ | 432/520 [27:35<05:23,  3.68s/it]
                                                 
{'loss': 1.07, 'grad_norm': 0.0013060820118660909, 'learning_rate': 0.01467091183678444, 'epoch': 0.83}
+
 83%|████████▎ | 432/520 [27:35<05:23,  3.68s/it]
 83%|████████▎ | 433/520 [27:39<05:19,  3.67s/it]
                                                 
{'loss': 1.2015, 'grad_norm': 0.001245693366757975, 'learning_rate': 0.014347544452214867, 'epoch': 0.83}
+
 83%|████████▎ | 433/520 [27:39<05:19,  3.67s/it]
 83%|████████▎ | 434/520 [27:42<05:16,  3.68s/it]
                                                 
{'loss': 0.9555, 'grad_norm': 0.001276597896467006, 'learning_rate': 0.014027505019880971, 'epoch': 0.83}
+
 83%|████████▎ | 434/520 [27:42<05:16,  3.68s/it]
 84%|████████▎ | 435/520 [27:46<05:12,  3.67s/it]
                                                 
{'loss': 1.2433, 'grad_norm': 0.0014644644635312433, 'learning_rate': 0.013710805974638696, 'epoch': 0.84}
+
 84%|████████▎ | 435/520 [27:46<05:12,  3.67s/it]
 84%|████████▍ | 436/520 [27:50<05:09,  3.68s/it]
                                                 
{'loss': 1.0423, 'grad_norm': 0.0013654541145904178, 'learning_rate': 0.01339745962155613, 'epoch': 0.84}
+
 84%|████████▍ | 436/520 [27:50<05:09,  3.68s/it]
 84%|████████▍ | 437/520 [27:53<05:06,  3.69s/it]
                                                 
{'loss': 1.2622, 'grad_norm': 0.0013093141063179616, 'learning_rate': 0.01308747813543536, 'epoch': 0.84}
+
 84%|████████▍ | 437/520 [27:53<05:06,  3.69s/it]
 84%|████████▍ | 438/520 [27:57<05:02,  3.69s/it]
                                                 
{'loss': 1.0826, 'grad_norm': 0.001253752118054535, 'learning_rate': 0.012780873560339467, 'epoch': 0.84}
+
 84%|████████▍ | 438/520 [27:57<05:02,  3.69s/it]
 84%|████████▍ | 439/520 [28:01<04:59,  3.70s/it]
                                                 
{'loss': 1.1329, 'grad_norm': 0.0010849049580419568, 'learning_rate': 0.012477657809124632, 'epoch': 0.84}
+
 84%|████████▍ | 439/520 [28:01<04:59,  3.70s/it]
 85%|████████▍ | 440/520 [28:04<04:56,  3.71s/it]
                                                 
{'loss': 1.1133, 'grad_norm': 0.0012318563738940311, 'learning_rate': 0.012177842662977134, 'epoch': 0.85}
+
 85%|████████▍ | 440/520 [28:04<04:56,  3.71s/it]
 85%|████████▍ | 441/520 [28:08<04:53,  3.71s/it]
                                                 
{'loss': 1.1473, 'grad_norm': 0.0012534352128924407, 'learning_rate': 0.01188143977095576, 'epoch': 0.85}
+
 85%|████████▍ | 441/520 [28:08<04:53,  3.71s/it]
 85%|████████▌ | 442/520 [28:12<04:53,  3.76s/it]
                                                 
{'loss': 1.1812, 'grad_norm': 0.0014178067483432348, 'learning_rate': 0.011588460649539035, 'epoch': 0.85}
+
 85%|████████▌ | 442/520 [28:12<04:53,  3.76s/it]
 85%|████████▌ | 443/520 [28:16<04:52,  3.80s/it]
                                                 
{'loss': 1.1912, 'grad_norm': 0.0012754146428117991, 'learning_rate': 0.011298916682177829, 'epoch': 0.85}
+
 85%|████████▌ | 443/520 [28:16<04:52,  3.80s/it]
 85%|████████▌ | 444/520 [28:20<04:51,  3.84s/it]
                                                 
{'loss': 1.1556, 'grad_norm': 0.001168337967664458, 'learning_rate': 0.011012819118853146, 'epoch': 0.85}
+
 85%|████████▌ | 444/520 [28:20<04:51,  3.84s/it]
 86%|████████▌ | 445/520 [28:24<04:49,  3.86s/it]
                                                 
{'loss': 1.0858, 'grad_norm': 0.0012481600411826661, 'learning_rate': 0.01073017907563887, 'epoch': 0.86}
+
 86%|████████▌ | 445/520 [28:24<04:49,  3.86s/it]
 86%|████████▌ | 446/520 [28:28<04:45,  3.86s/it]
                                                 
{'loss': 1.2275, 'grad_norm': 0.0012419153897566247, 'learning_rate': 0.010451007534269908, 'epoch': 0.86}
+
 86%|████████▌ | 446/520 [28:28<04:45,  3.86s/it]
 86%|████████▌ | 447/520 [28:32<04:43,  3.89s/it]
                                                 
{'loss': 1.1653, 'grad_norm': 0.0012741356631163195, 'learning_rate': 0.010175315341715598, 'epoch': 0.86}
+
 86%|████████▌ | 447/520 [28:32<04:43,  3.89s/it]
 86%|████████▌ | 448/520 [28:35<04:39,  3.88s/it]
                                                 
{'loss': 1.1559, 'grad_norm': 0.0013057154008172622, 'learning_rate': 0.009903113209758098, 'epoch': 0.86}
+
 86%|████████▌ | 448/520 [28:35<04:39,  3.88s/it]
 86%|████████▋ | 449/520 [28:39<04:34,  3.87s/it]
                                                 
{'loss': 1.1821, 'grad_norm': 0.0013329682893053712, 'learning_rate': 0.009634411714576352, 'epoch': 0.86}
+
 86%|████████▋ | 449/520 [28:39<04:34,  3.87s/it]
 87%|████████▋ | 450/520 [28:43<04:27,  3.82s/it]
                                                 
{'loss': 1.1862, 'grad_norm': 0.001312861267604802, 'learning_rate': 0.009369221296335007, 'epoch': 0.87}
+
 87%|████████▋ | 450/520 [28:43<04:27,  3.82s/it]
 87%|████████▋ | 451/520 [28:47<04:21,  3.78s/it]
                                                 
{'loss': 1.1822, 'grad_norm': 0.0012953661771451484, 'learning_rate': 0.009107552258778906, 'epoch': 0.87}
+
 87%|████████▋ | 451/520 [28:47<04:21,  3.78s/it]
 87%|████████▋ | 452/520 [28:50<04:15,  3.76s/it]
                                                 
{'loss': 1.2233, 'grad_norm': 0.0011790104985273653, 'learning_rate': 0.008849414768832687, 'epoch': 0.87}
+
 87%|████████▋ | 452/520 [28:50<04:15,  3.76s/it]
 87%|████████▋ | 453/520 [28:54<04:11,  3.75s/it]
                                                 
{'loss': 1.1993, 'grad_norm': 0.0012452157221091453, 'learning_rate': 0.008594818856205699, 'epoch': 0.87}
+
 87%|████████▋ | 453/520 [28:54<04:11,  3.75s/it]
 87%|████████▋ | 454/520 [28:58<04:06,  3.73s/it]
                                                 
{'loss': 1.0928, 'grad_norm': 0.0013159883686187641, 'learning_rate': 0.00834377441300238, 'epoch': 0.87}
+
 87%|████████▋ | 454/520 [28:58<04:06,  3.73s/it]
 88%|████████▊ | 455/520 [29:01<04:01,  3.72s/it]
                                                 
{'loss': 1.2301, 'grad_norm': 0.0012547861458684548, 'learning_rate': 0.008096291193337934, 'epoch': 0.88}
+
 88%|████████▊ | 455/520 [29:01<04:01,  3.72s/it]
 88%|████████▊ | 456/520 [29:05<03:57,  3.70s/it]
                                                 
{'loss': 1.1532, 'grad_norm': 0.0012966410307555006, 'learning_rate': 0.007852378812959226, 'epoch': 0.88}
+
 88%|████████▊ | 456/520 [29:05<03:57,  3.70s/it]
 88%|████████▊ | 457/520 [29:09<03:55,  3.73s/it]
                                                 
{'loss': 1.1122, 'grad_norm': 0.001172456941986787, 'learning_rate': 0.007612046748871327, 'epoch': 0.88}
+
 88%|████████▊ | 457/520 [29:09<03:55,  3.73s/it]
 88%|████████▊ | 458/520 [29:13<03:54,  3.78s/it]
                                                 
{'loss': 1.2864, 'grad_norm': 0.0013551142587228462, 'learning_rate': 0.007375304338969136, 'epoch': 0.88}
+
 88%|████████▊ | 458/520 [29:13<03:54,  3.78s/it]
 88%|████████▊ | 459/520 [29:17<03:53,  3.82s/it]
                                                 
{'loss': 1.2173, 'grad_norm': 0.0013226275171857504, 'learning_rate': 0.007142160781674645, 'epoch': 0.88}
+
 88%|████████▊ | 459/520 [29:17<03:53,  3.82s/it]
 88%|████████▊ | 460/520 [29:21<03:50,  3.85s/it]
                                                 
{'loss': 1.1053, 'grad_norm': 0.0012917030839707125, 'learning_rate': 0.006912625135579587, 'epoch': 0.88}
+
 88%|████████▊ | 460/520 [29:21<03:50,  3.85s/it]
 89%|████████▊ | 461/520 [29:25<03:48,  3.87s/it]
                                                 
{'loss': 1.194, 'grad_norm': 0.0009977198623187483, 'learning_rate': 0.0066867063190933496, 'epoch': 0.89}
+
 89%|████████▊ | 461/520 [29:25<03:48,  3.87s/it]
 89%|████████▉ | 462/520 [29:29<03:45,  3.88s/it]
                                                 
{'loss': 1.2707, 'grad_norm': 0.001221475713671325, 'learning_rate': 0.006464413110096601, 'epoch': 0.89}
+
 89%|████████▉ | 462/520 [29:29<03:45,  3.88s/it]
 89%|████████▉ | 463/520 [29:32<03:41,  3.89s/it]
                                                 
{'loss': 1.064, 'grad_norm': 0.0013121227505549335, 'learning_rate': 0.006245754145600091, 'epoch': 0.89}
+
 89%|████████▉ | 463/520 [29:32<03:41,  3.89s/it]
 89%|████████▉ | 464/520 [29:36<03:37,  3.89s/it]
                                                 
{'loss': 1.2013, 'grad_norm': 0.0013553535587194033, 'learning_rate': 0.006030737921409169, 'epoch': 0.89}
+
 89%|████████▉ | 464/520 [29:36<03:37,  3.89s/it]
 89%|████████▉ | 465/520 [29:40<03:33,  3.89s/it]
                                                 
{'loss': 1.3083, 'grad_norm': 0.0013736858499822915, 'learning_rate': 0.005819372791793654, 'epoch': 0.89}
+
 89%|████████▉ | 465/520 [29:40<03:33,  3.89s/it]
 90%|████████▉ | 466/520 [29:44<03:30,  3.89s/it]
                                                 
{'loss': 1.1912, 'grad_norm': 0.0011804864495515398, 'learning_rate': 0.005611666969163243, 'epoch': 0.9}
+
 90%|████████▉ | 466/520 [29:44<03:30,  3.89s/it]
 90%|████████▉ | 467/520 [29:48<03:26,  3.90s/it]
                                                 
{'loss': 1.159, 'grad_norm': 0.0011507578633713333, 'learning_rate': 0.005407628523748398, 'epoch': 0.9}
+
 90%|████████▉ | 467/520 [29:48<03:26,  3.90s/it]
 90%|█████████ | 468/520 [29:52<03:22,  3.90s/it]
                                                 
{'loss': 1.1644, 'grad_norm': 0.0014152679839123748, 'learning_rate': 0.00520726538328683, 'epoch': 0.9}
+
 90%|█████████ | 468/520 [29:52<03:22,  3.90s/it]
 90%|█████████ | 469/520 [29:56<03:18,  3.90s/it]
                                                 
{'loss': 1.2227, 'grad_norm': 0.0013948121112100419, 'learning_rate': 0.005010585332715401, 'epoch': 0.9}
+
 90%|█████████ | 469/520 [29:56<03:18,  3.90s/it]
 90%|█████████ | 470/520 [30:00<03:14,  3.89s/it]
                                                 
{'loss': 1.1062, 'grad_norm': 0.001184683901851702, 'learning_rate': 0.004817596013867765, 'epoch': 0.9}
+
 90%|█████████ | 470/520 [30:00<03:14,  3.89s/it]
 91%|█████████ | 471/520 [30:03<03:07,  3.83s/it]
                                                 
{'loss': 1.1301, 'grad_norm': 0.0013284335000437492, 'learning_rate': 0.004628304925177318, 'epoch': 0.91}
+
 91%|█████████ | 471/520 [30:03<03:07,  3.83s/it]
 91%|█████████ | 472/520 [30:07<03:01,  3.79s/it]
                                                 
{'loss': 1.0946, 'grad_norm': 0.0013479504671442225, 'learning_rate': 0.004442719421385921, 'epoch': 0.91}
+
 91%|█████████ | 472/520 [30:07<03:01,  3.79s/it]
 91%|█████████ | 473/520 [30:11<02:56,  3.76s/it]
                                                 
{'loss': 1.1613, 'grad_norm': 0.001312597201168865, 'learning_rate': 0.004260846713258193, 'epoch': 0.91}
+
 91%|█████████ | 473/520 [30:11<02:56,  3.76s/it]
 91%|█████████ | 474/520 [30:14<02:52,  3.75s/it]
                                                 
{'loss': 1.1941, 'grad_norm': 0.00119065225869285, 'learning_rate': 0.004082693867301224, 'epoch': 0.91}
+
 91%|█████████ | 474/520 [30:14<02:52,  3.75s/it]
 91%|█████████▏| 475/520 [30:18<02:47,  3.73s/it]
                                                 
{'loss': 1.1134, 'grad_norm': 0.001186473757492844, 'learning_rate': 0.003908267805490051, 'epoch': 0.91}
+
 91%|█████████▏| 475/520 [30:18<02:47,  3.73s/it]
 92%|█████████▏| 476/520 [30:22<02:43,  3.72s/it]
                                                 
{'loss': 1.1483, 'grad_norm': 0.00131553476179598, 'learning_rate': 0.003737575304998797, 'epoch': 0.92}
+
 92%|█████████▏| 476/520 [30:22<02:43,  3.72s/it]
 92%|█████████▏| 477/520 [30:26<02:39,  3.71s/it]
                                                 
{'loss': 1.139, 'grad_norm': 0.001415132729576043, 'learning_rate': 0.003570622997937234, 'epoch': 0.92}
+
 92%|█████████▏| 477/520 [30:26<02:39,  3.71s/it]
 92%|█████████▏| 478/520 [30:29<02:35,  3.71s/it]
                                                 
{'loss': 1.091, 'grad_norm': 0.001285939680900139, 'learning_rate': 0.00340741737109318, 'epoch': 0.92}
+
 92%|█████████▏| 478/520 [30:29<02:35,  3.71s/it]
 92%|█████████▏| 479/520 [30:33<02:32,  3.71s/it]
                                                 
{'loss': 1.1582, 'grad_norm': 0.0013253635800008793, 'learning_rate': 0.003247964765680389, 'epoch': 0.92}
+
 92%|█████████▏| 479/520 [30:33<02:32,  3.71s/it]
 92%|█████████▏| 480/520 [30:37<02:29,  3.73s/it]
                                                 
{'loss': 1.1779, 'grad_norm': 0.0011513149376919036, 'learning_rate': 0.0030922713770922153, 'epoch': 0.92}
+
 92%|█████████▏| 480/520 [30:37<02:29,  3.73s/it]
 92%|█████████▎| 481/520 [30:41<02:26,  3.77s/it]
                                                 
{'loss': 1.1704, 'grad_norm': 0.0011561427702979127, 'learning_rate': 0.0029403432546609046, 'epoch': 0.93}
+
 92%|█████████▎| 481/520 [30:41<02:26,  3.77s/it]
 93%|█████████▎| 482/520 [30:44<02:24,  3.81s/it]
                                                 
{'loss': 1.1908, 'grad_norm': 0.0012446972033030045, 'learning_rate': 0.0027921863014225504, 'epoch': 0.93}
+
 93%|█████████▎| 482/520 [30:44<02:24,  3.81s/it]
 93%|█████████▎| 483/520 [30:48<02:21,  3.83s/it]
                                                 
{'loss': 1.1608, 'grad_norm': 0.0013771843735055495, 'learning_rate': 0.002647806273887665, 'epoch': 0.93}
+
 93%|█████████▎| 483/520 [30:48<02:21,  3.83s/it]
 93%|█████████▎| 484/520 [30:52<02:18,  3.85s/it]
                                                 
{'loss': 1.1683, 'grad_norm': 0.001341680202387324, 'learning_rate': 0.0025072087818176383, 'epoch': 0.93}
+
 93%|█████████▎| 484/520 [30:52<02:18,  3.85s/it]
 93%|█████████▎| 485/520 [30:56<02:15,  3.87s/it]
                                                 
{'loss': 1.1228, 'grad_norm': 0.0012546791463715664, 'learning_rate': 0.002370399288006664, 'epoch': 0.93}
+
 93%|█████████▎| 485/520 [30:56<02:15,  3.87s/it]
 93%|█████████▎| 486/520 [31:00<02:11,  3.88s/it]
                                                 
{'loss': 1.2438, 'grad_norm': 0.0013994940497722972, 'learning_rate': 0.0022373831080695463, 'epoch': 0.93}
+
 93%|█████████▎| 486/520 [31:00<02:11,  3.88s/it]
 94%|█████████▎| 487/520 [31:04<02:08,  3.88s/it]
                                                 
{'loss': 1.0941, 'grad_norm': 0.001243315956327689, 'learning_rate': 0.0021081654102351635, 'epoch': 0.94}
+
 94%|█████████▎| 487/520 [31:04<02:08,  3.88s/it]
 94%|█████████▍| 488/520 [31:08<02:04,  3.88s/it]
                                                 
{'loss': 1.039, 'grad_norm': 0.0013500720800946975, 'learning_rate': 0.0019827512151456175, 'epoch': 0.94}
+
 94%|█████████▍| 488/520 [31:08<02:04,  3.88s/it]
 94%|█████████▍| 489/520 [31:12<01:59,  3.85s/it]
                                                 
{'loss': 1.1911, 'grad_norm': 0.0010823833747567522, 'learning_rate': 0.0018611453956612345, 'epoch': 0.94}
+
 94%|█████████▍| 489/520 [31:12<01:59,  3.85s/it]
 94%|█████████▍| 490/520 [31:15<01:54,  3.80s/it]
                                                 
{'loss': 1.1584, 'grad_norm': 0.001303139845863755, 'learning_rate': 0.0017433526766711727, 'epoch': 0.94}
+
 94%|█████████▍| 490/520 [31:15<01:54,  3.80s/it]
 94%|█████████▍| 491/520 [31:19<01:49,  3.77s/it]
                                                 
{'loss': 1.1229, 'grad_norm': 0.0013301048784302458, 'learning_rate': 0.0016293776349098677, 'epoch': 0.94}
+
 94%|█████████▍| 491/520 [31:19<01:49,  3.77s/it]
 95%|█████████▍| 492/520 [31:23<01:44,  3.75s/it]
                                                 
{'loss': 1.2364, 'grad_norm': 0.0013606543943005354, 'learning_rate': 0.0015192246987791981, 'epoch': 0.95}
+
 95%|█████████▍| 492/520 [31:23<01:44,  3.75s/it]
 95%|█████████▍| 493/520 [31:26<01:40,  3.73s/it]
                                                 
{'loss': 1.1945, 'grad_norm': 0.0012771630682429772, 'learning_rate': 0.0014128981481764114, 'epoch': 0.95}
+
 95%|█████████▍| 493/520 [31:26<01:40,  3.73s/it]
 95%|█████████▌| 494/520 [31:30<01:36,  3.72s/it]
                                                 
{'loss': 1.1766, 'grad_norm': 0.0011766634792869444, 'learning_rate': 0.0013104021143278911, 'epoch': 0.95}
+
 95%|█████████▌| 494/520 [31:30<01:36,  3.72s/it]
 95%|█████████▌| 495/520 [31:34<01:32,  3.71s/it]
                                                 
{'loss': 1.1431, 'grad_norm': 0.0013152522468960562, 'learning_rate': 0.0012117405796285285, 'epoch': 0.95}
+
 95%|█████████▌| 495/520 [31:34<01:32,  3.71s/it]
 95%|█████████▌| 496/520 [31:38<01:28,  3.71s/it]
                                                 
{'loss': 1.0627, 'grad_norm': 0.0013961035060756012, 'learning_rate': 0.0011169173774871477, 'epoch': 0.95}
+
 95%|█████████▌| 496/520 [31:38<01:28,  3.71s/it]
 96%|█████████▌| 497/520 [31:41<01:25,  3.70s/it]
                                                 
{'loss': 1.1241, 'grad_norm': 0.00110716856320092, 'learning_rate': 0.0010259361921774012, 'epoch': 0.96}
+
 96%|█████████▌| 497/520 [31:41<01:25,  3.70s/it]
 96%|█████████▌| 498/520 [31:45<01:21,  3.70s/it]
                                                 
{'loss': 1.1382, 'grad_norm': 0.0012711270992862749, 'learning_rate': 0.000938800558694719, 'epoch': 0.96}
+
 96%|█████████▌| 498/520 [31:45<01:21,  3.70s/it]
 96%|█████████▌| 499/520 [31:49<01:17,  3.69s/it]
                                                 
{'loss': 1.2602, 'grad_norm': 0.001305456510205752, 'learning_rate': 0.0008555138626189618, 'epoch': 0.96}
+
 96%|█████████▌| 499/520 [31:49<01:17,  3.69s/it]
 96%|█████████▌| 500/520 [31:52<01:13,  3.69s/it]
                                                 
{'loss': 1.2594, 'grad_norm': 0.0014692822451304037, 'learning_rate': 0.0007760793399827937, 'epoch': 0.96}
+
 96%|█████████▌| 500/520 [31:52<01:13,  3.69s/it]
 96%|█████████▋| 501/520 [31:56<01:10,  3.69s/it]
                                                 
{'loss': 1.167, 'grad_norm': 0.001397443976955959, 'learning_rate': 0.000700500077146038, 'epoch': 0.96}
+
 96%|█████████▋| 501/520 [31:56<01:10,  3.69s/it]
 97%|█████████▋| 502/520 [32:00<01:06,  3.69s/it]
                                                 
{'loss': 1.1755, 'grad_norm': 0.0011848111038683187, 'learning_rate': 0.0006287790106757397, 'epoch': 0.97}
+
 97%|█████████▋| 502/520 [32:00<01:06,  3.69s/it]
 97%|█████████▋| 503/520 [32:03<01:02,  3.70s/it]
                                                 
{'loss': 1.1559, 'grad_norm': 0.0012772298592986136, 'learning_rate': 0.0005609189272320237, 'epoch': 0.97}
+
 97%|█████████▋| 503/520 [32:03<01:02,  3.70s/it]
 97%|█████████▋| 504/520 [32:07<00:59,  3.69s/it]
                                                 
{'loss': 1.1694, 'grad_norm': 0.0016201110094997747, 'learning_rate': 0.000496922463459859, 'epoch': 0.97}
+
 97%|█████████▋| 504/520 [32:07<00:59,  3.69s/it]
 97%|█████████▋| 505/520 [32:11<00:55,  3.69s/it]
                                                 
{'loss': 1.2044, 'grad_norm': 0.0013351028795320484, 'learning_rate': 0.0004367921058866187, 'epoch': 0.97}
+
 97%|█████████▋| 505/520 [32:11<00:55,  3.69s/it]
 97%|█████████▋| 506/520 [32:14<00:51,  3.69s/it]
                                                 
{'loss': 1.1295, 'grad_norm': 0.001324626831806818, 'learning_rate': 0.0003805301908254455, 'epoch': 0.97}
+
 97%|█████████▋| 506/520 [32:14<00:51,  3.69s/it]
 98%|█████████▊| 507/520 [32:18<00:47,  3.69s/it]
                                                 
{'loss': 1.3, 'grad_norm': 0.0011780318081826913, 'learning_rate': 0.0003281389042844918, 'epoch': 0.97}
+
 98%|█████████▊| 507/520 [32:18<00:47,  3.69s/it]
 98%|█████████▊| 508/520 [32:22<00:44,  3.69s/it]
                                                 
{'loss': 1.2492, 'grad_norm': 0.0013142299532791825, 'learning_rate': 0.00027962028188198705, 'epoch': 0.98}
+
 98%|█████████▊| 508/520 [32:22<00:44,  3.69s/it]
 98%|█████████▊| 509/520 [32:26<00:40,  3.70s/it]
                                                 
{'loss': 1.2173, 'grad_norm': 0.0012516561667586337, 'learning_rate': 0.00023497620876711256, 'epoch': 0.98}
+
 98%|█████████▊| 509/520 [32:26<00:40,  3.70s/it]
 98%|█████████▊| 510/520 [32:29<00:37,  3.74s/it]
                                                 
{'loss': 1.1627, 'grad_norm': 0.0012623927491672738, 'learning_rate': 0.00019420841954681523, 'epoch': 0.98}
+
 98%|█████████▊| 510/520 [32:29<00:37,  3.74s/it]
 98%|█████████▊| 511/520 [32:33<00:33,  3.72s/it]
                                                 
{'loss': 1.1411, 'grad_norm': 0.0012112569612206232, 'learning_rate': 0.00015731849821833956, 'epoch': 0.98}
+
 98%|█████████▊| 511/520 [32:33<00:33,  3.72s/it]
 98%|█████████▊| 512/520 [32:37<00:29,  3.72s/it]
                                                 
{'loss': 1.0287, 'grad_norm': 0.001336950600038162, 'learning_rate': 0.00012430787810776555, 'epoch': 0.98}
+
 98%|█████████▊| 512/520 [32:37<00:29,  3.72s/it]
 99%|█████████▊| 513/520 [32:40<00:26,  3.71s/it]
                                                 
{'loss': 1.2252, 'grad_norm': 0.0013866211502383432, 'learning_rate': 9.517784181422018e-05, 'epoch': 0.99}
+
 99%|█████████▊| 513/520 [32:40<00:26,  3.71s/it]
 99%|█████████▉| 514/520 [32:44<00:22,  3.70s/it]
                                                 
{'loss': 1.1947, 'grad_norm': 0.0011685455642037766, 'learning_rate': 6.992952116013917e-05, 'epoch': 0.99}
+
 99%|█████████▉| 514/520 [32:44<00:22,  3.70s/it]
 99%|█████████▉| 515/520 [32:48<00:18,  3.70s/it]
                                                 
{'loss': 1.2414, 'grad_norm': 0.0015016485859420607, 'learning_rate': 4.856389714723575e-05, 'epoch': 0.99}
+
 99%|█████████▉| 515/520 [32:48<00:18,  3.70s/it]
 99%|█████████▉| 516/520 [32:51<00:14,  3.69s/it]
                                                 
{'loss': 1.1452, 'grad_norm': 0.0012294254131195012, 'learning_rate': 3.108179991837545e-05, 'epoch': 0.99}
+
 99%|█████████▉| 516/520 [32:51<00:14,  3.69s/it]
 99%|█████████▉| 517/520 [32:55<00:11,  3.67s/it]
                                                 
{'loss': 1.192, 'grad_norm': 0.0012192499707845866, 'learning_rate': 1.7483908725357544e-05, 'epoch': 0.99}
+
 99%|█████████▉| 517/520 [32:55<00:11,  3.67s/it]
100%|█████████▉| 518/520 [32:59<00:07,  3.64s/it]
                                                 
{'loss': 1.1624, 'grad_norm': 0.0014543934667867002, 'learning_rate': 7.770751902513862e-06, 'epoch': 1.0}
+
100%|█████████▉| 518/520 [32:59<00:07,  3.64s/it]
100%|█████████▉| 519/520 [33:02<00:03,  3.64s/it]
                                                 
{'loss': 1.1617, 'grad_norm': 0.0012370912847196828, 'learning_rate': 1.9427068461808086e-06, 'epoch': 1.0}
+
100%|█████████▉| 519/520 [33:02<00:03,  3.64s/it]
100%|██████████| 520/520 [33:07<00:00,  3.89s/it]
                                                 
{'loss': 1.1701, 'grad_norm': 0.0012292789258514707, 'learning_rate': 0.0, 'epoch': 1.0}
+
100%|██████████| 520/520 [33:07<00:00,  3.89s/it]
                                                 
{'train_runtime': 1987.2884, 'train_samples_per_second': 33.477, 'train_steps_per_second': 0.262, 'train_loss': 1.2597867743327067, 'epoch': 1.0}
+
100%|██████████| 520/520 [33:07<00:00,  3.89s/it]
100%|██████████| 520/520 [33:07<00:00,  3.82s/it]
+[2025-10-16 17:28:47,781] [INFO] [launch.py:348:main] Process 2545602 exits successfully.
+[2025-10-16 17:28:48,782] [INFO] [launch.py:348:main] Process 2545597 exits successfully.
+[2025-10-16 17:28:48,783] [INFO] [launch.py:348:main] Process 2545598 exits successfully.
+[2025-10-16 17:28:48,783] [INFO] [launch.py:348:main] Process 2545600 exits successfully.
+[2025-10-16 17:28:49,785] [INFO] [launch.py:348:main] Process 2545603 exits successfully.
+[2025-10-16 17:28:49,786] [INFO] [launch.py:348:main] Process 2545599 exits successfully.
+[2025-10-16 17:28:49,786] [INFO] [launch.py:348:main] Process 2545601 exits successfully.
+[2025-10-16 17:28:52,790] [INFO] [launch.py:348:main] Process 2545596 exits successfully.
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.7_2e-1_connector-5.0_2.7_2e-1_ablation_20251016_165407.log
+Timestamp: 2025-10-16 17:28:55
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation_20251016_172855.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation_20251016_172855.log
new file mode 100644
index 0000000000000000000000000000000000000000..38fdaad7dac932a21d2d2f7236aa9aff39471316
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation_20251016_172855.log
@@ -0,0 +1,2312 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation_20251016_172855.log
+Timestamp: 2025-10-16 17:28:55
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 17:28:57,994] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:01,241] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 17:29:01,242] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 5.0 --temperature_attn_text 2.9 --temperature_mlp_text 2.9 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 5.0 --temperature_attn_vision 2.9 --temperature_mlp_vision 2.9 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 5.0 --temperature_connector 2.9 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 17:29:03,785] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:04,849] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 17:29:04,849] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 17:29:04,849] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 17:29:04,849] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 17:29:04,849] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 17:29:04,849] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 17:29:04,849] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 17:29:04,852] [INFO] [launch.py:253:main] process 2567433 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,853] [INFO] [launch.py:253:main] process 2567434 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,855] [INFO] [launch.py:253:main] process 2567435 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,857] [INFO] [launch.py:253:main] process 2567436 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,859] [INFO] [launch.py:253:main] process 2567437 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,861] [INFO] [launch.py:253:main] process 2567438 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,863] [INFO] [launch.py:253:main] process 2567439 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 17:29:04,865] [INFO] [launch.py:253:main] process 2567440 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '5.0', '--temperature_attn_text', '2.9', '--temperature_mlp_text', '2.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '5.0', '--temperature_attn_vision', '2.9', '--temperature_mlp_vision', '2.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '5.0', '--temperature_connector', '2.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 17:29:11,911] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:11,913] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:11,969] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:11,981] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:12,015] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:12,016] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:12,017] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:12,048] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 17:29:12,416] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,416] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,416] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,416] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,430] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,430] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,447] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,458] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 17:29:12,458] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 2.9, 'temperature_mlp': 2.9, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 2.9, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 2.9,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 2.9,
+    "temperature_mlp": 2.9,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2567433:2567433 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567433:2567433 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567433:2567433 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567433:2567433 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567433:2567433 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2567433:2567433 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Using network Socket
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2567436:2567436 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567436:2567436 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567436:2567436 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567436:2567436 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567436:2567436 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567436:2567436 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2567437:2567437 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567437:2567437 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567437:2567437 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567437:2567437 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567437:2567437 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567437:2567437 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2567435:2567435 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567435:2567435 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567435:2567435 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567435:2567435 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567435:2567435 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567435:2567435 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2567434:2567434 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567434:2567434 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567434:2567434 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567434:2567434 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567434:2567434 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567434:2567434 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2567440:2567440 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567440:2567440 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567440:2567440 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567440:2567440 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567440:2567440 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567440:2567440 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2567438:2567438 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567438:2567438 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567438:2567438 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567438:2567438 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567438:2567438 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567438:2567438 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2567439:2567439 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2567439:2567439 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567439:2567439 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567439:2567439 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2567439:2567439 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2567439:2567439 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO ncclCommInitRank comm 0x55ad947245f0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO ncclCommInitRank comm 0x55b854ef7e40 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO ncclCommInitRank comm 0x5641d32bf1e0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO ncclCommInitRank comm 0x556ea4233140 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO ncclCommInitRank comm 0x5571e3125c50 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO ncclCommInitRank comm 0x55a081addb20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO ncclCommInitRank comm 0x55708a5457f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO ncclCommInitRank comm 0x561b1aa10a20 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xd03b4426064226f4 - Init START
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO comm 0x55ad947245f0 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO comm 0x561b1aa10a20 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO comm 0x55708a5457f0 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO comm 0x55a081addb20 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO comm 0x55b854ef7e40 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO comm 0x5571e3125c50 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO comm 0x5641d32bf1e0 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO comm 0x556ea4233140 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567435:2569107 [2] NCCL INFO ncclCommInitRank comm 0x561b1aa10a20 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567436:2569104 [3] NCCL INFO ncclCommInitRank comm 0x5641d32bf1e0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567438:2569110 [5] NCCL INFO ncclCommInitRank comm 0x55708a5457f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567440:2569109 [7] NCCL INFO ncclCommInitRank comm 0x556ea4233140 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567439:2569111 [6] NCCL INFO ncclCommInitRank comm 0x55ad947245f0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567433:2569085 [0] NCCL INFO ncclCommInitRank comm 0x5571e3125c50 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567437:2569106 [4] NCCL INFO ncclCommInitRank comm 0x55a081addb20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xd03b4426064226f4 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2567434:2569108 [1] NCCL INFO ncclCommInitRank comm 0x55b854ef7e40 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xd03b4426064226f4 - Init COMPLETE
+[2025-10-16 17:29:58,547] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 17:30:00,319] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_towerLoading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=5.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=5.000000
+Pre-training init connector._connector.0.scores: Mean=5.000005
+Pre-training init connector._connector.2.scores: Mean=4.999970
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 17:30:18,498 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 17:30:18,504 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO ncclCommInitRank comm 0x7ff14806b800 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO ncclCommInitRank comm 0x7f467c06b630 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO ncclCommInitRank comm 0x7f266006af80 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO ncclCommInitRank comm 0x7f944006b770 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO ncclCommInitRank comm 0x7f39a406b450 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO ncclCommInitRank comm 0x7f890c06b6a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO ncclCommInitRank comm 0x7fa0f8069590 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO ncclCommInitRank comm 0x7f7aa806ab60 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x7eb50242cc691459 - Init START
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO comm 0x7f944006b770 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO comm 0x7f39a406b450 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO comm 0x7f467c06b630 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO comm 0x7ff14806b800 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO comm 0x7f7aa806ab60 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO comm 0x7f266006af80 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO comm 0x7f890c06b6a0 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO comm 0x7fa0f8069590 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2567436:2574140 [3] NCCL INFO ncclCommInitRank comm 0x7fa0f8069590 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567440:2574137 [7] NCCL INFO ncclCommInitRank comm 0x7f266006af80 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567437:2574136 [4] NCCL INFO ncclCommInitRank comm 0x7f890c06b6a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567438:2574139 [5] NCCL INFO ncclCommInitRank comm 0x7f39a406b450 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567439:2574138 [6] NCCL INFO ncclCommInitRank comm 0x7f944006b770 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567433:2574134 [0] NCCL INFO ncclCommInitRank comm 0x7f467c06b630 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567434:2574135 [1] NCCL INFO ncclCommInitRank comm 0x7ff14806b800 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x7eb50242cc691459 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2567435:2574141 [2] NCCL INFO ncclCommInitRank comm 0x7f7aa806ab60 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x7eb50242cc691459 - Init COMPLETE
+
  0%|          | 1/520 [00:14<2:02:08, 14.12s/it]
                                                 
{'loss': 2.4889, 'grad_norm': 0.034117914840541295, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:14<2:02:08, 14.12s/it]
  0%|          | 2/520 [00:17<1:08:51,  7.98s/it]
                                                 
{'loss': 2.3915, 'grad_norm': 0.03318996905783698, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:17<1:08:51,  7.98s/it]
  1%|          | 3/520 [00:21<51:35,  5.99s/it]  
                                               
{'loss': 2.5844, 'grad_norm': 0.03728094767269909, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:21<51:35,  5.99s/it]
  1%|          | 4/520 [00:25<43:31,  5.06s/it]
                                               
{'loss': 1.8698, 'grad_norm': 0.013857067078865946, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:25<43:31,  5.06s/it]
  1%|          | 5/520 [00:28<39:14,  4.57s/it]
                                               
{'loss': 1.8275, 'grad_norm': 0.009118504609227371, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:28<39:14,  4.57s/it]
  1%|          | 6/520 [00:32<36:29,  4.26s/it]
                                               
{'loss': 1.634, 'grad_norm': 0.0086278424138866, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:32<36:29,  4.26s/it]
  1%|▏         | 7/520 [00:36<34:39,  4.05s/it]
                                               
{'loss': 1.6167, 'grad_norm': 0.010997697384319372, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:36<34:39,  4.05s/it]
  2%|▏         | 8/520 [00:40<35:11,  4.12s/it]
                                               
{'loss': 1.5809, 'grad_norm': 0.006832587963471146, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:40<35:11,  4.12s/it]
  2%|▏         | 9/520 [00:44<35:09,  4.13s/it]
                                               
{'loss': 1.612, 'grad_norm': 0.003983003738002454, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:44<35:09,  4.13s/it]
  2%|▏         | 10/520 [00:48<33:49,  3.98s/it]
                                                
{'loss': 1.4433, 'grad_norm': 0.00406189141483461, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:48<33:49,  3.98s/it]
  2%|▏         | 11/520 [00:51<33:13,  3.92s/it]
                                                
{'loss': 1.5115, 'grad_norm': 0.004722202872027895, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [00:51<33:13,  3.92s/it]
  2%|▏         | 12/520 [00:55<32:28,  3.84s/it]
                                                
{'loss': 1.4431, 'grad_norm': 0.003584805611352278, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [00:55<32:28,  3.84s/it][2025-10-16 17:31:22,606] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [00:59<33:36,  3.98s/it]
                                                
{'loss': 1.4522, 'grad_norm': 0.0028384332928513825, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [00:59<33:36,  3.98s/it]
  3%|▎         | 14/520 [01:03<32:38,  3.87s/it]
                                                
{'loss': 1.4985, 'grad_norm': 0.003418030096093348, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:03<32:38,  3.87s/it]
  3%|▎         | 15/520 [01:07<31:58,  3.80s/it]
                                                
{'loss': 1.4842, 'grad_norm': 0.0026287801609011013, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:07<31:58,  3.80s/it]
  3%|▎         | 16/520 [01:10<31:31,  3.75s/it]
                                                
{'loss': 1.4319, 'grad_norm': 0.0024033319917720034, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:10<31:31,  3.75s/it]
  3%|▎         | 17/520 [01:14<31:07,  3.71s/it]
                                                
{'loss': 1.535, 'grad_norm': 0.002912203994977872, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:14<31:07,  3.71s/it]
  3%|▎         | 18/520 [01:17<30:50,  3.69s/it]
                                                
{'loss': 1.378, 'grad_norm': 0.0024639521139079193, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:17<30:50,  3.69s/it]
  4%|▎         | 19/520 [01:21<30:35,  3.66s/it]
                                                
{'loss': 1.433, 'grad_norm': 0.002194461473190259, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:21<30:35,  3.66s/it]
  4%|▍         | 20/520 [01:25<30:32,  3.66s/it]
                                                
{'loss': 1.3553, 'grad_norm': 0.0022160502136900073, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:25<30:32,  3.66s/it]
  4%|▍         | 21/520 [01:28<30:29,  3.67s/it]
                                                
{'loss': 1.4376, 'grad_norm': 0.0026465352779182017, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:28<30:29,  3.67s/it]
  4%|▍         | 22/520 [01:32<30:18,  3.65s/it]
                                                
{'loss': 1.5153, 'grad_norm': 0.0020487541138578974, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:32<30:18,  3.65s/it]
  4%|▍         | 23/520 [01:36<30:12,  3.65s/it]
                                                
{'loss': 1.4508, 'grad_norm': 0.0021818731751151866, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:36<30:12,  3.65s/it]
  5%|▍         | 24/520 [01:39<30:06,  3.64s/it]
                                                
{'loss': 1.4019, 'grad_norm': 0.001999347891523532, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [01:39<30:06,  3.64s/it]
  5%|▍         | 25/520 [01:43<30:04,  3.65s/it]
                                                
{'loss': 1.4467, 'grad_norm': 0.0020565473299810557, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [01:43<30:04,  3.65s/it]
  5%|▌         | 26/520 [01:47<30:16,  3.68s/it]
                                                
{'loss': 1.4295, 'grad_norm': 0.0020348956138845466, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [01:47<30:16,  3.68s/it]
  5%|▌         | 27/520 [01:51<30:32,  3.72s/it]
                                                
{'loss': 1.3351, 'grad_norm': 0.001955606026989035, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [01:51<30:32,  3.72s/it]
  5%|▌         | 28/520 [01:54<30:43,  3.75s/it]
                                                
{'loss': 1.3463, 'grad_norm': 0.0018581111708713097, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [01:54<30:43,  3.75s/it]
  6%|▌         | 29/520 [01:58<30:56,  3.78s/it]
                                                
{'loss': 1.373, 'grad_norm': 0.002086643079088805, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [01:58<30:56,  3.78s/it]
  6%|▌         | 30/520 [02:02<30:59,  3.80s/it]
                                                
{'loss': 1.4794, 'grad_norm': 0.0018899226864143193, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:02<30:59,  3.80s/it]
  6%|▌         | 31/520 [02:06<31:00,  3.81s/it]
                                                
{'loss': 1.3472, 'grad_norm': 0.0017225845431423391, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:06<31:00,  3.81s/it]
  6%|▌         | 32/520 [02:10<31:00,  3.81s/it]
                                                
{'loss': 1.3517, 'grad_norm': 0.0021492071983031363, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:10<31:00,  3.81s/it]
  6%|▋         | 33/520 [02:14<31:11,  3.84s/it]
                                                
{'loss': 1.3538, 'grad_norm': 0.0019246253565235335, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:14<31:11,  3.84s/it]
  7%|▋         | 34/520 [02:17<31:09,  3.85s/it]
                                                
{'loss': 1.3374, 'grad_norm': 0.001949978452899857, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:17<31:09,  3.85s/it]
  7%|▋         | 35/520 [02:21<31:07,  3.85s/it]
                                                
{'loss': 1.3494, 'grad_norm': 0.002066838729104068, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:21<31:07,  3.85s/it]
  7%|▋         | 36/520 [02:25<31:07,  3.86s/it]
                                                
{'loss': 1.4548, 'grad_norm': 0.0021242033285857266, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:25<31:07,  3.86s/it]
  7%|▋         | 37/520 [02:29<31:04,  3.86s/it]
                                                
{'loss': 1.4435, 'grad_norm': 0.0017534028739266858, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:29<31:04,  3.86s/it]
  7%|▋         | 38/520 [02:33<31:00,  3.86s/it]
                                                
{'loss': 1.5208, 'grad_norm': 0.0017982609824231698, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:33<31:00,  3.86s/it]
  8%|▊         | 39/520 [02:37<30:53,  3.85s/it]
                                                
{'loss': 1.3716, 'grad_norm': 0.002238708804917386, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [02:37<30:53,  3.85s/it]
  8%|▊         | 40/520 [02:41<30:48,  3.85s/it]
                                                
{'loss': 1.4, 'grad_norm': 0.0019407175826250923, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [02:41<30:48,  3.85s/it]
  8%|▊         | 41/520 [02:44<30:46,  3.86s/it]
                                                
{'loss': 1.383, 'grad_norm': 0.0018830669824330668, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [02:44<30:46,  3.86s/it]
  8%|▊         | 42/520 [02:48<30:45,  3.86s/it]
                                                
{'loss': 1.3976, 'grad_norm': 0.0024517112477869672, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [02:48<30:45,  3.86s/it]
  8%|▊         | 43/520 [02:52<30:42,  3.86s/it]
                                                
{'loss': 1.3435, 'grad_norm': 0.0015866183902868583, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [02:52<30:42,  3.86s/it]
  8%|▊         | 44/520 [02:56<30:41,  3.87s/it]
                                                
{'loss': 1.4437, 'grad_norm': 0.0017439119713607409, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [02:56<30:41,  3.87s/it]
  9%|▊         | 45/520 [03:00<30:39,  3.87s/it]
                                                
{'loss': 1.4012, 'grad_norm': 0.0018385687144403982, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [03:00<30:39,  3.87s/it]
  9%|▉         | 46/520 [03:04<30:32,  3.87s/it]
                                                
{'loss': 1.5005, 'grad_norm': 0.0017232614710622652, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:04<30:32,  3.87s/it]
  9%|▉         | 47/520 [03:08<30:30,  3.87s/it]
                                                
{'loss': 1.3841, 'grad_norm': 0.0018147485061474702, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:08<30:30,  3.87s/it]
  9%|▉         | 48/520 [03:12<30:21,  3.86s/it]
                                                
{'loss': 1.3674, 'grad_norm': 0.002171904577385132, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:12<30:21,  3.86s/it]
  9%|▉         | 49/520 [03:15<30:27,  3.88s/it]
                                                
{'loss': 1.4092, 'grad_norm': 0.0017195853329619421, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:15<30:27,  3.88s/it]
 10%|▉         | 50/520 [03:19<30:21,  3.88s/it]
                                                
{'loss': 1.3955, 'grad_norm': 0.0016725090036306734, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:19<30:21,  3.88s/it]
 10%|▉         | 51/520 [03:23<30:08,  3.86s/it]
                                                
{'loss': 1.3279, 'grad_norm': 0.001933724645510598, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:23<30:08,  3.86s/it]
 10%|█         | 52/520 [03:27<30:04,  3.86s/it]
                                                
{'loss': 1.4538, 'grad_norm': 0.0020504107112645463, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:27<30:04,  3.86s/it]
 10%|█         | 53/520 [03:31<30:01,  3.86s/it]
                                                
{'loss': 1.4429, 'grad_norm': 0.0018619541391614095, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:31<30:01,  3.86s/it]
 10%|█         | 54/520 [03:35<29:57,  3.86s/it]
                                                
{'loss': 1.3558, 'grad_norm': 0.0018784761274908163, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:35<29:57,  3.86s/it]
 11%|█         | 55/520 [03:39<29:53,  3.86s/it]
                                                
{'loss': 1.3286, 'grad_norm': 0.002118995508938791, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [03:39<29:53,  3.86s/it]
 11%|█         | 56/520 [03:42<29:48,  3.85s/it]
                                                
{'loss': 1.4586, 'grad_norm': 0.0018453254565428806, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [03:42<29:48,  3.85s/it]
 11%|█         | 57/520 [03:46<29:46,  3.86s/it]
                                                
{'loss': 1.3292, 'grad_norm': 0.002439879990634269, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [03:46<29:46,  3.86s/it]
 11%|█         | 58/520 [03:50<29:43,  3.86s/it]
                                                
{'loss': 1.4842, 'grad_norm': 0.0018926411876376299, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [03:50<29:43,  3.86s/it]
 11%|█▏        | 59/520 [03:54<29:42,  3.87s/it]
                                                
{'loss': 1.3012, 'grad_norm': 0.0016354612549416153, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [03:54<29:42,  3.87s/it]
 12%|█▏        | 60/520 [03:58<29:35,  3.86s/it]
                                                
{'loss': 1.3911, 'grad_norm': 0.0018506754628049567, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [03:58<29:35,  3.86s/it]
 12%|█▏        | 61/520 [04:02<29:29,  3.85s/it]
                                                
{'loss': 1.407, 'grad_norm': 0.004108017381885451, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [04:02<29:29,  3.85s/it]
 12%|█▏        | 62/520 [04:06<29:31,  3.87s/it]
                                                
{'loss': 1.3713, 'grad_norm': 0.0018776436012686594, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [04:06<29:31,  3.87s/it]
 12%|█▏        | 63/520 [04:09<29:12,  3.83s/it]
                                                
{'loss': 1.3614, 'grad_norm': 0.0016747086439085096, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:09<29:12,  3.83s/it]
 12%|█▏        | 64/520 [04:13<29:10,  3.84s/it]
                                                
{'loss': 1.3895, 'grad_norm': 0.001962083494836413, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:13<29:10,  3.84s/it]
 12%|█▎        | 65/520 [04:17<28:43,  3.79s/it]
                                                
{'loss': 1.3921, 'grad_norm': 0.0022307217597973706, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:17<28:43,  3.79s/it]
 13%|█▎        | 66/520 [04:21<28:21,  3.75s/it]
                                                
{'loss': 1.3455, 'grad_norm': 0.0014797934550546568, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:21<28:21,  3.75s/it]
 13%|█▎        | 67/520 [04:24<28:12,  3.74s/it]
                                                
{'loss': 1.2513, 'grad_norm': 0.0017039299089650697, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:24<28:12,  3.74s/it]
 13%|█▎        | 68/520 [04:28<28:00,  3.72s/it]
                                                
{'loss': 1.3215, 'grad_norm': 0.0020708257292662985, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:28<28:00,  3.72s/it]
 13%|█▎        | 69/520 [04:32<27:45,  3.69s/it]
                                                
{'loss': 1.2939, 'grad_norm': 0.0016605165693155482, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:32<27:45,  3.69s/it]
 13%|█▎        | 70/520 [04:35<27:39,  3.69s/it]
                                                
{'loss': 1.3306, 'grad_norm': 0.0017074848827810468, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [04:35<27:39,  3.69s/it]
 14%|█▎        | 71/520 [04:39<27:29,  3.67s/it]
                                                
{'loss': 1.266, 'grad_norm': 0.0016487733182715093, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [04:39<27:29,  3.67s/it]
 14%|█▍        | 72/520 [04:43<27:23,  3.67s/it]
                                                
{'loss': 1.4212, 'grad_norm': 0.0019378878859265618, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [04:43<27:23,  3.67s/it]
 14%|█▍        | 73/520 [04:46<27:17,  3.66s/it]
                                                
{'loss': 1.2464, 'grad_norm': 0.0015785050960742952, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [04:46<27:17,  3.66s/it]
 14%|█▍        | 74/520 [04:50<27:14,  3.67s/it]
                                                
{'loss': 1.3597, 'grad_norm': 0.0019978061274433883, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [04:50<27:14,  3.67s/it]
 14%|█▍        | 75/520 [04:53<27:05,  3.65s/it]
                                                
{'loss': 1.2638, 'grad_norm': 0.0016139674933715463, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [04:53<27:05,  3.65s/it]
 15%|█▍        | 76/520 [04:57<27:00,  3.65s/it]
                                                
{'loss': 1.5018, 'grad_norm': 0.004879833734050724, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [04:57<27:00,  3.65s/it]
 15%|█▍        | 77/520 [05:01<26:54,  3.65s/it]
                                                
{'loss': 1.1871, 'grad_norm': 0.0018032162599355163, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [05:01<26:54,  3.65s/it]
 15%|█▌        | 78/520 [05:05<27:09,  3.69s/it]
                                                
{'loss': 1.3078, 'grad_norm': 0.0017499524490831898, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:05<27:09,  3.69s/it]
 15%|█▌        | 79/520 [05:08<27:00,  3.68s/it]
                                                
{'loss': 1.3016, 'grad_norm': 0.001680729850351058, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:08<27:00,  3.68s/it]
 15%|█▌        | 80/520 [05:12<26:58,  3.68s/it]
                                                
{'loss': 1.5335, 'grad_norm': 0.009150655407931453, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:12<26:58,  3.68s/it]
 16%|█▌        | 81/520 [05:16<26:50,  3.67s/it]
                                                
{'loss': 1.4414, 'grad_norm': 0.0022376796441349735, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:16<26:50,  3.67s/it]
 16%|█▌        | 82/520 [05:19<26:47,  3.67s/it]
                                                
{'loss': 1.3663, 'grad_norm': 0.0017510302217860485, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:19<26:47,  3.67s/it]
 16%|█▌        | 83/520 [05:23<26:43,  3.67s/it]
                                                
{'loss': 1.3926, 'grad_norm': 0.0018020435762380667, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:23<26:43,  3.67s/it]
 16%|█▌        | 84/520 [05:27<26:38,  3.67s/it]
                                                
{'loss': 1.4026, 'grad_norm': 0.0017427551748271494, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:27<26:38,  3.67s/it]
 16%|█▋        | 85/520 [05:30<26:33,  3.66s/it]
                                                
{'loss': 1.4336, 'grad_norm': 0.0018313197205301445, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:30<26:33,  3.66s/it]
 17%|█▋        | 86/520 [05:34<26:30,  3.66s/it]
                                                
{'loss': 1.4365, 'grad_norm': 0.0018862888728798852, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:34<26:30,  3.66s/it]
 17%|█▋        | 87/520 [05:38<26:28,  3.67s/it]
                                                
{'loss': 1.5142, 'grad_norm': 0.0031048866996752976, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [05:38<26:28,  3.67s/it]
 17%|█▋        | 88/520 [05:41<26:52,  3.73s/it]
                                                
{'loss': 1.4925, 'grad_norm': 0.004199088152774969, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [05:41<26:52,  3.73s/it]
 17%|█▋        | 89/520 [05:45<26:46,  3.73s/it]
                                                
{'loss': 1.3931, 'grad_norm': 0.0018160627231183198, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [05:45<26:46,  3.73s/it]
 17%|█▋        | 90/520 [05:49<26:34,  3.71s/it]
                                                
{'loss': 1.3259, 'grad_norm': 0.0016711320601694906, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [05:49<26:34,  3.71s/it]
 18%|█▊        | 91/520 [05:52<26:28,  3.70s/it]
                                                
{'loss': 1.3911, 'grad_norm': 0.0015638840076470889, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [05:52<26:28,  3.70s/it]
 18%|█▊        | 92/520 [05:56<26:20,  3.69s/it]
                                                
{'loss': 1.3267, 'grad_norm': 0.0016851725861916629, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [05:56<26:20,  3.69s/it]
 18%|█▊        | 93/520 [06:00<26:16,  3.69s/it]
                                                
{'loss': 1.3553, 'grad_norm': 0.001973123380508318, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [06:00<26:16,  3.69s/it]
 18%|█▊        | 94/520 [06:04<26:14,  3.70s/it]
                                                
{'loss': 1.434, 'grad_norm': 0.001647791476333874, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [06:04<26:14,  3.70s/it]
 18%|█▊        | 95/520 [06:07<26:08,  3.69s/it]
                                                
{'loss': 1.3228, 'grad_norm': 0.0018659136691115423, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:07<26:08,  3.69s/it]
 18%|█▊        | 96/520 [06:11<26:23,  3.74s/it]
                                                
{'loss': 1.3416, 'grad_norm': 0.001481546990933611, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:11<26:23,  3.74s/it]
 19%|█▊        | 97/520 [06:15<26:41,  3.79s/it]
                                                
{'loss': 1.3074, 'grad_norm': 0.002101673967185568, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:15<26:41,  3.79s/it]
 19%|█▉        | 98/520 [06:19<26:43,  3.80s/it]
                                                
{'loss': 1.3096, 'grad_norm': 0.0014088319165165258, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:19<26:43,  3.80s/it]
 19%|█▉        | 99/520 [06:23<26:45,  3.81s/it]
                                                
{'loss': 1.3187, 'grad_norm': 0.0016672280184794634, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:23<26:45,  3.81s/it]
 19%|█▉        | 100/520 [06:27<26:49,  3.83s/it]
                                                 
{'loss': 1.4003, 'grad_norm': 0.0021753409103958984, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:27<26:49,  3.83s/it]
 19%|█▉        | 101/520 [06:30<26:48,  3.84s/it]
                                                 
{'loss': 1.3234, 'grad_norm': 0.0016693826695038387, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:30<26:48,  3.84s/it]
 20%|█▉        | 102/520 [06:34<26:50,  3.85s/it]
                                                 
{'loss': 1.3296, 'grad_norm': 0.0018221859370047431, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [06:34<26:50,  3.85s/it]
 20%|█▉        | 103/520 [06:38<26:45,  3.85s/it]
                                                 
{'loss': 1.2595, 'grad_norm': 0.0015006869407040953, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [06:38<26:45,  3.85s/it]
 20%|██        | 104/520 [06:42<26:42,  3.85s/it]
                                                 
{'loss': 1.332, 'grad_norm': 0.0015715504385197194, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [06:42<26:42,  3.85s/it]
 20%|██        | 105/520 [06:46<26:27,  3.82s/it]
                                                 
{'loss': 1.3178, 'grad_norm': 0.0014543441931720374, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [06:46<26:27,  3.82s/it]
 20%|██        | 106/520 [06:49<26:03,  3.78s/it]
                                                 
{'loss': 1.3741, 'grad_norm': 0.0015961325836305873, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [06:49<26:03,  3.78s/it]
 21%|██        | 107/520 [06:53<25:47,  3.75s/it]
                                                 
{'loss': 1.3616, 'grad_norm': 0.001842290985544134, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [06:53<25:47,  3.75s/it]
 21%|██        | 108/520 [06:57<25:36,  3.73s/it]
                                                 
{'loss': 1.278, 'grad_norm': 0.0015821759463509242, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [06:57<25:36,  3.73s/it]
 21%|██        | 109/520 [07:00<25:25,  3.71s/it]
                                                 
{'loss': 1.3278, 'grad_norm': 0.0014786191391221867, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [07:00<25:25,  3.71s/it]
 21%|██        | 110/520 [07:04<25:14,  3.69s/it]
                                                 
{'loss': 1.4609, 'grad_norm': 0.0017764492839841375, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [07:04<25:14,  3.69s/it]
 21%|██▏       | 111/520 [07:08<25:05,  3.68s/it]
                                                 
{'loss': 1.4638, 'grad_norm': 0.0016690328713515757, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:08<25:05,  3.68s/it]
 22%|██▏       | 112/520 [07:11<25:07,  3.69s/it]
                                                 
{'loss': 1.3421, 'grad_norm': 0.0014820997963842918, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:11<25:07,  3.69s/it]
 22%|██▏       | 113/520 [07:15<25:01,  3.69s/it]
                                                 
{'loss': 1.2329, 'grad_norm': 0.0015928867267388, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:15<25:01,  3.69s/it]
 22%|██▏       | 114/520 [07:19<24:53,  3.68s/it]
                                                 
{'loss': 1.3277, 'grad_norm': 0.0015398781772253919, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:19<24:53,  3.68s/it]
 22%|██▏       | 115/520 [07:22<24:51,  3.68s/it]
                                                 
{'loss': 1.4282, 'grad_norm': 0.0014916440976100477, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:22<24:51,  3.68s/it]
 22%|██▏       | 116/520 [07:26<24:47,  3.68s/it]
                                                 
{'loss': 1.4476, 'grad_norm': 0.0015346764761530072, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:26<24:47,  3.68s/it]
 22%|██▎       | 117/520 [07:30<24:41,  3.68s/it]
                                                 
{'loss': 1.4155, 'grad_norm': 0.0015812416344319128, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:30<24:41,  3.68s/it]
 23%|██▎       | 118/520 [07:34<24:39,  3.68s/it]
                                                 
{'loss': 1.3138, 'grad_norm': 0.0014379395326357893, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [07:34<24:39,  3.68s/it]
 23%|██▎       | 119/520 [07:37<24:33,  3.67s/it]
                                                 
{'loss': 1.2733, 'grad_norm': 0.0015629824339689399, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [07:37<24:33,  3.67s/it]
 23%|██▎       | 120/520 [07:41<24:29,  3.67s/it]
                                                 
{'loss': 1.285, 'grad_norm': 0.0018614599204379385, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [07:41<24:29,  3.67s/it]
 23%|██▎       | 121/520 [07:44<24:22,  3.67s/it]
                                                 
{'loss': 1.3386, 'grad_norm': 0.0015565357818869785, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [07:44<24:22,  3.67s/it]
 23%|██▎       | 122/520 [07:48<24:21,  3.67s/it]
                                                 
{'loss': 1.2428, 'grad_norm': 0.0013921402070649533, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [07:48<24:21,  3.67s/it]
 24%|██▎       | 123/520 [07:52<24:19,  3.68s/it]
                                                 
{'loss': 1.4055, 'grad_norm': 0.0020174241473528377, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [07:52<24:19,  3.68s/it]
 24%|██▍       | 124/520 [07:56<24:19,  3.69s/it]
                                                 
{'loss': 1.3081, 'grad_norm': 0.0015481067348131316, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [07:56<24:19,  3.69s/it]
 24%|██▍       | 125/520 [07:59<24:15,  3.69s/it]
                                                 
{'loss': 1.2982, 'grad_norm': 0.0015152299987903216, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [07:59<24:15,  3.69s/it]
 24%|██▍       | 126/520 [08:04<25:32,  3.89s/it]
                                                 
{'loss': 1.3493, 'grad_norm': 0.0016322724531626148, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [08:04<25:32,  3.89s/it]
 24%|██▍       | 127/520 [08:07<25:08,  3.84s/it]
                                                 
{'loss': 1.2888, 'grad_norm': 0.0017488720225891446, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:07<25:08,  3.84s/it]
 25%|██▍       | 128/520 [08:11<24:45,  3.79s/it]
                                                 
{'loss': 1.3384, 'grad_norm': 0.00154492545298961, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:11<24:45,  3.79s/it]
 25%|██▍       | 129/520 [08:15<24:33,  3.77s/it]
                                                 
{'loss': 1.2768, 'grad_norm': 0.0013565094602205723, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:15<24:33,  3.77s/it]
 25%|██▌       | 130/520 [08:18<24:17,  3.74s/it]
                                                 
{'loss': 1.3184, 'grad_norm': 0.0012744115868850206, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:18<24:17,  3.74s/it]
 25%|██▌       | 131/520 [08:22<24:17,  3.75s/it]
                                                 
{'loss': 1.2912, 'grad_norm': 0.001431056340398805, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:22<24:17,  3.75s/it]
 25%|██▌       | 132/520 [08:26<24:04,  3.72s/it]
                                                 
{'loss': 1.3665, 'grad_norm': 0.0015334825891211979, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:26<24:04,  3.72s/it]
 26%|██▌       | 133/520 [08:30<23:56,  3.71s/it]
                                                 
{'loss': 1.2784, 'grad_norm': 0.001562679032858556, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:30<23:56,  3.71s/it]
 26%|██▌       | 134/520 [08:33<23:48,  3.70s/it]
                                                 
{'loss': 1.3608, 'grad_norm': 0.0015285223187060056, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [08:33<23:48,  3.70s/it]
 26%|██▌       | 135/520 [08:37<23:39,  3.69s/it]
                                                 
{'loss': 1.4195, 'grad_norm': 0.001474695778519471, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [08:37<23:39,  3.69s/it]
 26%|██▌       | 136/520 [08:41<23:32,  3.68s/it]
                                                 
{'loss': 1.348, 'grad_norm': 0.0014461461779558924, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [08:41<23:32,  3.68s/it]
 26%|██▋       | 137/520 [08:44<23:27,  3.68s/it]
                                                 
{'loss': 1.2685, 'grad_norm': 0.0016301532867702326, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [08:44<23:27,  3.68s/it]
 27%|██▋       | 138/520 [08:48<23:21,  3.67s/it]
                                                 
{'loss': 1.2809, 'grad_norm': 0.0014591740450084268, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [08:48<23:21,  3.67s/it]
 27%|██▋       | 139/520 [08:52<23:17,  3.67s/it]
                                                 
{'loss': 1.1997, 'grad_norm': 0.0014724745537565457, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [08:52<23:17,  3.67s/it]
 27%|██▋       | 140/520 [08:55<23:14,  3.67s/it]
                                                 
{'loss': 1.3301, 'grad_norm': 0.0013289500090518582, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [08:55<23:14,  3.67s/it]
 27%|██▋       | 141/520 [08:59<23:08,  3.66s/it]
                                                 
{'loss': 1.3904, 'grad_norm': 0.0014699664429603206, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [08:59<23:08,  3.66s/it]
 27%|██▋       | 142/520 [09:03<23:06,  3.67s/it]
                                                 
{'loss': 1.3649, 'grad_norm': 0.001534572529290386, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [09:03<23:06,  3.67s/it]
 28%|██▊       | 143/520 [09:06<23:04,  3.67s/it]
                                                 
{'loss': 1.3053, 'grad_norm': 0.001530198491269999, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:06<23:04,  3.67s/it]
 28%|██▊       | 144/520 [09:10<23:03,  3.68s/it]
                                                 
{'loss': 1.2686, 'grad_norm': 0.0015499239895923383, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:10<23:03,  3.68s/it]
 28%|██▊       | 145/520 [09:14<23:06,  3.70s/it]
                                                 
{'loss': 1.2043, 'grad_norm': 0.0012891626905055266, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:14<23:06,  3.70s/it]
 28%|██▊       | 146/520 [09:17<23:18,  3.74s/it]
                                                 
{'loss': 1.4071, 'grad_norm': 0.0014811000706424237, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:17<23:18,  3.74s/it]
 28%|██▊       | 147/520 [09:21<23:27,  3.77s/it]
                                                 
{'loss': 1.2501, 'grad_norm': 0.0014134715310112702, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:21<23:27,  3.77s/it]
 28%|██▊       | 148/520 [09:25<23:30,  3.79s/it]
                                                 
{'loss': 1.2827, 'grad_norm': 0.001417253030626718, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:25<23:30,  3.79s/it]
 29%|██▊       | 149/520 [09:29<23:34,  3.81s/it]
                                                 
{'loss': 1.2224, 'grad_norm': 0.0013710326328173971, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:29<23:34,  3.81s/it]
 29%|██▉       | 150/520 [09:33<23:33,  3.82s/it]
                                                 
{'loss': 1.4603, 'grad_norm': 0.0015341923121528743, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [09:33<23:33,  3.82s/it]
 29%|██▉       | 151/520 [09:37<23:21,  3.80s/it]
                                                 
{'loss': 1.2729, 'grad_norm': 0.0014275580467147172, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [09:37<23:21,  3.80s/it]
 29%|██▉       | 152/520 [09:40<23:04,  3.76s/it]
                                                 
{'loss': 1.2419, 'grad_norm': 0.0014808764185753043, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [09:40<23:04,  3.76s/it]
 29%|██▉       | 153/520 [09:44<22:52,  3.74s/it]
                                                 
{'loss': 1.2757, 'grad_norm': 0.0014251739857261418, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [09:44<22:52,  3.74s/it]
 30%|██▉       | 154/520 [09:48<22:43,  3.72s/it]
                                                 
{'loss': 1.3576, 'grad_norm': 0.001367128439978376, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [09:48<22:43,  3.72s/it]
 30%|██▉       | 155/520 [09:51<22:32,  3.71s/it]
                                                 
{'loss': 1.2715, 'grad_norm': 0.001445756452454189, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [09:51<22:32,  3.71s/it]
 30%|███       | 156/520 [09:55<22:22,  3.69s/it]
                                                 
{'loss': 1.3009, 'grad_norm': 0.0015554362289822111, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [09:55<22:22,  3.69s/it]
 30%|███       | 157/520 [09:59<22:14,  3.68s/it]
                                                 
{'loss': 1.4291, 'grad_norm': 0.0016192024529584595, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [09:59<22:14,  3.68s/it]
 30%|███       | 158/520 [10:02<22:06,  3.66s/it]
                                                 
{'loss': 1.2804, 'grad_norm': 0.0017279350566854535, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [10:02<22:06,  3.66s/it]
 31%|███       | 159/520 [10:06<22:02,  3.66s/it]
                                                 
{'loss': 1.318, 'grad_norm': 0.0014518198290868099, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [10:06<22:02,  3.66s/it]
 31%|███       | 160/520 [10:10<21:55,  3.66s/it]
                                                 
{'loss': 1.3306, 'grad_norm': 0.0014778592361748015, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:10<21:55,  3.66s/it]
 31%|███       | 161/520 [10:13<21:54,  3.66s/it]
                                                 
{'loss': 1.3121, 'grad_norm': 0.0014084344435855544, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:13<21:54,  3.66s/it]
 31%|███       | 162/520 [10:17<21:51,  3.66s/it]
                                                 
{'loss': 1.3389, 'grad_norm': 0.0015028148060898241, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:17<21:51,  3.66s/it]
 31%|███▏      | 163/520 [10:21<21:49,  3.67s/it]
                                                 
{'loss': 1.2011, 'grad_norm': 0.0017911256003576162, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:21<21:49,  3.67s/it]
 32%|███▏      | 164/520 [10:24<21:44,  3.66s/it]
                                                 
{'loss': 1.1695, 'grad_norm': 0.001491082898938922, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:24<21:44,  3.66s/it]
 32%|███▏      | 165/520 [10:28<21:40,  3.66s/it]
                                                 
{'loss': 1.3073, 'grad_norm': 0.001343437164227205, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:28<21:40,  3.66s/it]
 32%|███▏      | 166/520 [10:32<21:40,  3.67s/it]
                                                 
{'loss': 1.2895, 'grad_norm': 0.0016717329124416, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:32<21:40,  3.67s/it]
 32%|███▏      | 167/520 [10:35<21:35,  3.67s/it]
                                                 
{'loss': 1.2839, 'grad_norm': 0.0014752757496956806, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [10:35<21:35,  3.67s/it]
 32%|███▏      | 168/520 [10:39<21:29,  3.66s/it]
                                                 
{'loss': 1.2201, 'grad_norm': 0.001401035586423785, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [10:39<21:29,  3.66s/it]
 32%|███▎      | 169/520 [10:43<21:26,  3.67s/it]
                                                 
{'loss': 1.3033, 'grad_norm': 0.001427395883871406, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [10:43<21:26,  3.67s/it]
 33%|███▎      | 170/520 [10:46<21:24,  3.67s/it]
                                                 
{'loss': 1.2764, 'grad_norm': 0.0013270271522370645, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [10:46<21:24,  3.67s/it]
 33%|███▎      | 171/520 [10:50<21:23,  3.68s/it]
                                                 
{'loss': 1.2319, 'grad_norm': 0.0015526336599104202, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [10:50<21:23,  3.68s/it]
 33%|███▎      | 172/520 [10:54<21:15,  3.67s/it]
                                                 
{'loss': 1.3025, 'grad_norm': 0.0013510816510301513, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [10:54<21:15,  3.67s/it]
 33%|███▎      | 173/520 [10:57<21:14,  3.67s/it]
                                                 
{'loss': 1.249, 'grad_norm': 0.0013951338357375543, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [10:57<21:14,  3.67s/it]
 33%|███▎      | 174/520 [11:01<21:14,  3.68s/it]
                                                 
{'loss': 1.3008, 'grad_norm': 0.0014393736530503491, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [11:01<21:14,  3.68s/it]
 34%|███▎      | 175/520 [11:05<21:11,  3.68s/it]
                                                 
{'loss': 1.2077, 'grad_norm': 0.0013381362054015404, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [11:05<21:11,  3.68s/it]
 34%|███▍      | 176/520 [11:08<21:05,  3.68s/it]
                                                 
{'loss': 1.3539, 'grad_norm': 0.0014373349009664056, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:08<21:05,  3.68s/it]
 34%|███▍      | 177/520 [11:12<21:02,  3.68s/it]
                                                 
{'loss': 1.2195, 'grad_norm': 0.0014744350678994919, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:12<21:02,  3.68s/it]
 34%|███▍      | 178/520 [11:16<20:59,  3.68s/it]
                                                 
{'loss': 1.2764, 'grad_norm': 0.001496790069424375, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:16<20:59,  3.68s/it]
 34%|███▍      | 179/520 [11:19<20:54,  3.68s/it]
                                                 
{'loss': 1.3603, 'grad_norm': 0.0013221242639962034, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:19<20:54,  3.68s/it]
 35%|███▍      | 180/520 [11:23<21:11,  3.74s/it]
                                                 
{'loss': 1.2823, 'grad_norm': 0.0015561647731438023, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:23<21:11,  3.74s/it]
 35%|███▍      | 181/520 [11:27<21:22,  3.78s/it]
                                                 
{'loss': 1.2534, 'grad_norm': 0.0012890459220726846, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:27<21:22,  3.78s/it]
 35%|███▌      | 182/520 [11:31<21:31,  3.82s/it]
                                                 
{'loss': 1.2642, 'grad_norm': 0.001443782986377428, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:31<21:31,  3.82s/it]
 35%|███▌      | 183/520 [11:35<21:35,  3.85s/it]
                                                 
{'loss': 1.2926, 'grad_norm': 0.0014033433590010472, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [11:35<21:35,  3.85s/it]
 35%|███▌      | 184/520 [11:39<21:38,  3.87s/it]
                                                 
{'loss': 1.2195, 'grad_norm': 0.0015152723003280064, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [11:39<21:38,  3.87s/it]
 36%|███▌      | 185/520 [11:43<21:36,  3.87s/it]
                                                 
{'loss': 1.3705, 'grad_norm': 0.0014213518338356244, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [11:43<21:36,  3.87s/it]
 36%|███▌      | 186/520 [11:47<21:31,  3.87s/it]
                                                 
{'loss': 1.2418, 'grad_norm': 0.0014845896171257996, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [11:47<21:31,  3.87s/it]
 36%|███▌      | 187/520 [11:50<21:25,  3.86s/it]
                                                 
{'loss': 1.2424, 'grad_norm': 0.0016971828428856773, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [11:50<21:25,  3.86s/it]
 36%|███▌      | 188/520 [11:54<21:22,  3.86s/it]
                                                 
{'loss': 1.3254, 'grad_norm': 0.0014581709031031796, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [11:54<21:22,  3.86s/it]
 36%|███▋      | 189/520 [11:58<21:09,  3.83s/it]
                                                 
{'loss': 1.3356, 'grad_norm': 0.0012670584509213234, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [11:58<21:09,  3.83s/it]
 37%|███▋      | 190/520 [12:02<20:52,  3.80s/it]
                                                 
{'loss': 1.2606, 'grad_norm': 0.0015281070863610174, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [12:02<20:52,  3.80s/it]
 37%|███▋      | 191/520 [12:06<20:41,  3.77s/it]
                                                 
{'loss': 1.2045, 'grad_norm': 0.001276767205313577, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [12:06<20:41,  3.77s/it]
 37%|███▋      | 192/520 [12:09<20:30,  3.75s/it]
                                                 
{'loss': 1.2875, 'grad_norm': 0.001375236418108708, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:09<20:30,  3.75s/it]
 37%|███▋      | 193/520 [12:13<20:21,  3.73s/it]
                                                 
{'loss': 1.2757, 'grad_norm': 0.001686829841804752, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:13<20:21,  3.73s/it]
 37%|███▋      | 194/520 [12:17<20:14,  3.73s/it]
                                                 
{'loss': 1.1669, 'grad_norm': 0.0014685445675146457, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:17<20:14,  3.73s/it]
 38%|███▊      | 195/520 [12:20<20:06,  3.71s/it]
                                                 
{'loss': 1.3123, 'grad_norm': 0.0014416120302318109, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:20<20:06,  3.71s/it]
 38%|███▊      | 196/520 [12:24<20:02,  3.71s/it]
                                                 
{'loss': 1.2885, 'grad_norm': 0.0016258070618044688, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:24<20:02,  3.71s/it]
 38%|███▊      | 197/520 [12:28<20:11,  3.75s/it]
                                                 
{'loss': 1.242, 'grad_norm': 0.0013846110812946104, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:28<20:11,  3.75s/it]
 38%|███▊      | 198/520 [12:32<20:18,  3.78s/it]
                                                 
{'loss': 1.3069, 'grad_norm': 0.001442789198750971, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:32<20:18,  3.78s/it]
 38%|███▊      | 199/520 [12:36<20:26,  3.82s/it]
                                                 
{'loss': 1.2267, 'grad_norm': 0.0014409626626523274, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [12:36<20:26,  3.82s/it]
 38%|███▊      | 200/520 [12:39<20:24,  3.83s/it]
                                                 
{'loss': 1.2243, 'grad_norm': 0.0014242625785130757, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [12:39<20:24,  3.83s/it]
 39%|███▊      | 201/520 [12:43<20:25,  3.84s/it]
                                                 
{'loss': 1.2411, 'grad_norm': 0.001198844003983005, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [12:43<20:25,  3.84s/it]
 39%|███▉      | 202/520 [12:47<20:20,  3.84s/it]
                                                 
{'loss': 1.2231, 'grad_norm': 0.0013846863223938174, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [12:47<20:20,  3.84s/it]
 39%|███▉      | 203/520 [12:51<20:16,  3.84s/it]
                                                 
{'loss': 1.2802, 'grad_norm': 0.0014514421019395346, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [12:51<20:16,  3.84s/it]
 39%|███▉      | 204/520 [12:55<20:15,  3.85s/it]
                                                 
{'loss': 1.2923, 'grad_norm': 0.001452635093066012, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [12:55<20:15,  3.85s/it]
 39%|███▉      | 205/520 [12:59<20:11,  3.85s/it]
                                                 
{'loss': 1.2547, 'grad_norm': 0.001330767317235516, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [12:59<20:11,  3.85s/it]
 40%|███▉      | 206/520 [13:02<19:59,  3.82s/it]
                                                 
{'loss': 1.3293, 'grad_norm': 0.0014003297426125065, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [13:02<19:59,  3.82s/it]
 40%|███▉      | 207/520 [13:06<19:42,  3.78s/it]
                                                 
{'loss': 1.2278, 'grad_norm': 0.00125087726408685, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [13:06<19:42,  3.78s/it]
 40%|████      | 208/520 [13:10<19:29,  3.75s/it]
                                                 
{'loss': 1.3161, 'grad_norm': 0.0015555740654420057, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [13:10<19:29,  3.75s/it]
 40%|████      | 209/520 [13:14<19:20,  3.73s/it]
                                                 
{'loss': 1.232, 'grad_norm': 0.0013298309159200114, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:14<19:20,  3.73s/it]
 40%|████      | 210/520 [13:17<19:10,  3.71s/it]
                                                 
{'loss': 1.3068, 'grad_norm': 0.0014628368916660465, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:17<19:10,  3.71s/it]
 41%|████      | 211/520 [13:21<19:03,  3.70s/it]
                                                 
{'loss': 1.3075, 'grad_norm': 0.0013258938562634155, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:21<19:03,  3.70s/it]
 41%|████      | 212/520 [13:25<18:56,  3.69s/it]
                                                 
{'loss': 1.2919, 'grad_norm': 0.001327346682523084, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:25<18:56,  3.69s/it]
 41%|████      | 213/520 [13:28<18:52,  3.69s/it]
                                                 
{'loss': 1.2511, 'grad_norm': 0.0016155164768023736, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:28<18:52,  3.69s/it]
 41%|████      | 214/520 [13:32<18:51,  3.70s/it]
                                                 
{'loss': 1.2361, 'grad_norm': 0.0014165997929560911, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:32<18:51,  3.70s/it]
 41%|████▏     | 215/520 [13:36<19:02,  3.75s/it]
                                                 
{'loss': 1.1733, 'grad_norm': 0.0013138061188935088, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [13:36<19:02,  3.75s/it]
 42%|████▏     | 216/520 [13:40<19:12,  3.79s/it]
                                                 
{'loss': 1.1589, 'grad_norm': 0.0013756206556769957, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [13:40<19:12,  3.79s/it]
 42%|████▏     | 217/520 [13:44<19:18,  3.82s/it]
                                                 
{'loss': 1.2821, 'grad_norm': 0.0014179835262041018, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [13:44<19:18,  3.82s/it]
 42%|████▏     | 218/520 [13:47<19:17,  3.83s/it]
                                                 
{'loss': 1.2746, 'grad_norm': 0.0014647274548435934, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [13:47<19:17,  3.83s/it]
 42%|████▏     | 219/520 [13:51<19:13,  3.83s/it]
                                                 
{'loss': 1.2688, 'grad_norm': 0.0012621284516320583, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [13:51<19:13,  3.83s/it]
 42%|████▏     | 220/520 [13:55<19:10,  3.83s/it]
                                                 
{'loss': 1.2265, 'grad_norm': 0.0013562217591934096, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [13:55<19:10,  3.83s/it]
 42%|████▎     | 221/520 [13:59<19:08,  3.84s/it]
                                                 
{'loss': 1.2762, 'grad_norm': 0.0013789612329226334, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [13:59<19:08,  3.84s/it]
 43%|████▎     | 222/520 [14:03<19:07,  3.85s/it]
                                                 
{'loss': 1.2018, 'grad_norm': 0.0013455077207207764, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [14:03<19:07,  3.85s/it]
 43%|████▎     | 223/520 [14:07<19:05,  3.86s/it]
                                                 
{'loss': 1.202, 'grad_norm': 0.0013258959567089064, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [14:07<19:05,  3.86s/it]
 43%|████▎     | 224/520 [14:11<19:02,  3.86s/it]
                                                 
{'loss': 1.3126, 'grad_norm': 0.0013476717788569656, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [14:11<19:02,  3.86s/it]
 43%|████▎     | 225/520 [14:14<18:58,  3.86s/it]
                                                 
{'loss': 1.2093, 'grad_norm': 0.0013525623571829868, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:14<18:58,  3.86s/it]
 43%|████▎     | 226/520 [14:18<18:46,  3.83s/it]
                                                 
{'loss': 1.3114, 'grad_norm': 0.0013259553446124982, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:18<18:46,  3.83s/it]
 44%|████▎     | 227/520 [14:22<18:28,  3.78s/it]
                                                 
{'loss': 1.2963, 'grad_norm': 0.001283158609393399, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:22<18:28,  3.78s/it]
 44%|████▍     | 228/520 [14:26<18:16,  3.75s/it]
                                                 
{'loss': 1.3487, 'grad_norm': 0.0013935053962703496, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:26<18:16,  3.75s/it]
 44%|████▍     | 229/520 [14:29<18:05,  3.73s/it]
                                                 
{'loss': 1.2681, 'grad_norm': 0.0012742877437406396, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:29<18:05,  3.73s/it]
 44%|████▍     | 230/520 [14:33<18:10,  3.76s/it]
                                                 
{'loss': 1.1569, 'grad_norm': 0.0012558990001342047, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:33<18:10,  3.76s/it]
 44%|████▍     | 231/520 [14:37<18:13,  3.78s/it]
                                                 
{'loss': 1.217, 'grad_norm': 0.001251961078427533, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:37<18:13,  3.78s/it]
 45%|████▍     | 232/520 [14:41<18:12,  3.79s/it]
                                                 
{'loss': 1.3747, 'grad_norm': 0.0014206110854976344, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [14:41<18:12,  3.79s/it]
 45%|████▍     | 233/520 [14:45<18:11,  3.80s/it]
                                                 
{'loss': 1.2595, 'grad_norm': 0.0014857421214563252, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [14:45<18:11,  3.80s/it]
 45%|████▌     | 234/520 [14:48<18:09,  3.81s/it]
                                                 
{'loss': 1.1716, 'grad_norm': 0.0014516193774041267, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [14:48<18:09,  3.81s/it]
 45%|████▌     | 235/520 [14:52<18:10,  3.83s/it]
                                                 
{'loss': 1.2274, 'grad_norm': 0.0014413190168509917, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [14:52<18:10,  3.83s/it]
 45%|████▌     | 236/520 [14:56<18:06,  3.83s/it]
                                                 
{'loss': 1.308, 'grad_norm': 0.0012933458240087128, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [14:56<18:06,  3.83s/it]
 46%|████▌     | 237/520 [15:00<18:03,  3.83s/it]
                                                 
{'loss': 1.3033, 'grad_norm': 0.001333166940411186, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [15:00<18:03,  3.83s/it]
 46%|████▌     | 238/520 [15:04<17:59,  3.83s/it]
                                                 
{'loss': 1.233, 'grad_norm': 0.001400070491188469, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [15:04<17:59,  3.83s/it]
 46%|████▌     | 239/520 [15:08<17:56,  3.83s/it]
                                                 
{'loss': 1.3166, 'grad_norm': 0.0013907332364822983, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [15:08<17:56,  3.83s/it]
 46%|████▌     | 240/520 [15:11<17:51,  3.83s/it]
                                                 
{'loss': 1.1139, 'grad_norm': 0.0013116505742006633, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:11<17:51,  3.83s/it]
 46%|████▋     | 241/520 [15:15<17:47,  3.83s/it]
                                                 
{'loss': 1.2033, 'grad_norm': 0.0013165700622889969, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:15<17:47,  3.83s/it]
 47%|████▋     | 242/520 [15:19<17:43,  3.83s/it]
                                                 
{'loss': 1.2166, 'grad_norm': 0.0012537494541068198, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:19<17:43,  3.83s/it]
 47%|████▋     | 243/520 [15:23<17:44,  3.84s/it]
                                                 
{'loss': 1.2099, 'grad_norm': 0.0013067940006480441, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:23<17:44,  3.84s/it]
 47%|████▋     | 244/520 [15:27<17:39,  3.84s/it]
                                                 
{'loss': 1.3249, 'grad_norm': 0.0014282945845008458, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:27<17:39,  3.84s/it]
 47%|████▋     | 245/520 [15:31<17:34,  3.84s/it]
                                                 
{'loss': 1.193, 'grad_norm': 0.0013995211042324713, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:31<17:34,  3.84s/it]
 47%|████▋     | 246/520 [15:34<17:28,  3.83s/it]
                                                 
{'loss': 1.3471, 'grad_norm': 0.0013661438599332173, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:34<17:28,  3.83s/it]
 48%|████▊     | 247/520 [15:38<17:23,  3.82s/it]
                                                 
{'loss': 1.3672, 'grad_norm': 0.0013668099148076796, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:38<17:23,  3.82s/it]
 48%|████▊     | 248/520 [15:42<17:18,  3.82s/it]
                                                 
{'loss': 1.1972, 'grad_norm': 0.0013360541046733474, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [15:42<17:18,  3.82s/it]
 48%|████▊     | 249/520 [15:46<17:13,  3.81s/it]
                                                 
{'loss': 1.289, 'grad_norm': 0.0013137029415866523, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [15:46<17:13,  3.81s/it]
 48%|████▊     | 250/520 [15:50<17:08,  3.81s/it]
                                                 
{'loss': 1.2255, 'grad_norm': 0.0014084314088494308, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [15:50<17:08,  3.81s/it]
 48%|████▊     | 251/520 [15:53<17:03,  3.81s/it]
                                                 
{'loss': 1.2894, 'grad_norm': 0.0012722255063093069, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [15:53<17:03,  3.81s/it]
 48%|████▊     | 252/520 [15:57<16:59,  3.80s/it]
                                                 
{'loss': 1.2501, 'grad_norm': 0.0012965926950505356, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [15:57<16:59,  3.80s/it]
 49%|████▊     | 253/520 [16:01<16:56,  3.81s/it]
                                                 
{'loss': 1.2837, 'grad_norm': 0.0015496296134128666, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [16:01<16:56,  3.81s/it]
 49%|████▉     | 254/520 [16:05<16:52,  3.81s/it]
                                                 
{'loss': 1.2206, 'grad_norm': 0.0012416766750985922, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [16:05<16:52,  3.81s/it]
 49%|████▉     | 255/520 [16:09<16:48,  3.80s/it]
                                                 
{'loss': 1.2257, 'grad_norm': 0.001490525012600932, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [16:09<16:48,  3.80s/it]
 49%|████▉     | 256/520 [16:12<16:43,  3.80s/it]
                                                 
{'loss': 1.2767, 'grad_norm': 0.0014222544617591814, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:12<16:43,  3.80s/it]
 49%|████▉     | 257/520 [16:16<16:40,  3.80s/it]
                                                 
{'loss': 1.2574, 'grad_norm': 0.0013850108869630552, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:16<16:40,  3.80s/it]
 50%|████▉     | 258/520 [16:20<16:36,  3.80s/it]
                                                 
{'loss': 1.2666, 'grad_norm': 0.001212424407118745, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:20<16:36,  3.80s/it]
 50%|████▉     | 259/520 [16:24<16:31,  3.80s/it]
                                                 
{'loss': 1.3476, 'grad_norm': 0.0016108191425413404, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:24<16:31,  3.80s/it]
 50%|█████     | 260/520 [16:28<16:27,  3.80s/it]
                                                 
{'loss': 1.2974, 'grad_norm': 0.0012721063936162924, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:28<16:27,  3.80s/it]
 50%|█████     | 261/520 [16:31<16:21,  3.79s/it]
                                                 
{'loss': 1.2431, 'grad_norm': 0.001388853222934649, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:31<16:21,  3.79s/it]
 50%|█████     | 262/520 [16:35<16:22,  3.81s/it]
                                                 
{'loss': 1.1907, 'grad_norm': 0.0013186809290719086, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:35<16:22,  3.81s/it]
 51%|█████     | 263/520 [16:39<16:17,  3.80s/it]
                                                 
{'loss': 1.2585, 'grad_norm': 0.0013244499017562206, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:39<16:17,  3.80s/it]
 51%|█████     | 264/520 [16:43<16:16,  3.82s/it]
                                                 
{'loss': 1.2927, 'grad_norm': 0.0013829216599853796, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [16:43<16:16,  3.82s/it]
 51%|█████     | 265/520 [16:47<16:10,  3.81s/it]
                                                 
{'loss': 1.2055, 'grad_norm': 0.0014671197026816353, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [16:47<16:10,  3.81s/it]
 51%|█████     | 266/520 [16:50<16:06,  3.81s/it]
                                                 
{'loss': 1.0759, 'grad_norm': 0.0012275006642452763, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [16:50<16:06,  3.81s/it]
 51%|█████▏    | 267/520 [16:54<16:04,  3.81s/it]
                                                 
{'loss': 1.1983, 'grad_norm': 0.0012969728896405264, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [16:54<16:04,  3.81s/it]
 52%|█████▏    | 268/520 [16:58<15:58,  3.80s/it]
                                                 
{'loss': 1.3765, 'grad_norm': 0.0018794885392100754, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [16:58<15:58,  3.80s/it]
 52%|█████▏    | 269/520 [17:02<15:53,  3.80s/it]
                                                 
{'loss': 1.3087, 'grad_norm': 0.0014125049422015635, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [17:02<15:53,  3.80s/it]
 52%|█████▏    | 270/520 [17:06<15:49,  3.80s/it]
                                                 
{'loss': 1.1845, 'grad_norm': 0.0012799795232976698, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [17:06<15:49,  3.80s/it]
 52%|█████▏    | 271/520 [17:09<15:46,  3.80s/it]
                                                 
{'loss': 1.2914, 'grad_norm': 0.001358486588856644, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [17:09<15:46,  3.80s/it]
 52%|█████▏    | 272/520 [17:13<15:43,  3.81s/it]
                                                 
{'loss': 1.2035, 'grad_norm': 0.001356501831282949, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:13<15:43,  3.81s/it]
 52%|█████▎    | 273/520 [17:17<15:38,  3.80s/it]
                                                 
{'loss': 1.3253, 'grad_norm': 0.0013401413535074644, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:17<15:38,  3.80s/it]
 53%|█████▎    | 274/520 [17:21<15:34,  3.80s/it]
                                                 
{'loss': 1.2681, 'grad_norm': 0.0014971170689644059, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:21<15:34,  3.80s/it]
 53%|█████▎    | 275/520 [17:25<15:32,  3.80s/it]
                                                 
{'loss': 1.2123, 'grad_norm': 0.0014704144100498755, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:25<15:32,  3.80s/it]
 53%|█████▎    | 276/520 [17:28<15:29,  3.81s/it]
                                                 
{'loss': 1.279, 'grad_norm': 0.0015527215706458658, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:28<15:29,  3.81s/it]
 53%|█████▎    | 277/520 [17:32<15:30,  3.83s/it]
                                                 
{'loss': 1.3188, 'grad_norm': 0.001253784286627094, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:32<15:30,  3.83s/it]
 53%|█████▎    | 278/520 [17:36<15:28,  3.84s/it]
                                                 
{'loss': 1.1599, 'grad_norm': 0.0012395555822064745, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:36<15:28,  3.84s/it]
 54%|█████▎    | 279/520 [17:40<15:22,  3.83s/it]
                                                 
{'loss': 1.2105, 'grad_norm': 0.001487953874980471, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:40<15:22,  3.83s/it]
 54%|█████▍    | 280/520 [17:44<15:17,  3.82s/it]
                                                 
{'loss': 1.209, 'grad_norm': 0.0015415607747661351, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [17:44<15:17,  3.82s/it]
 54%|█████▍    | 281/520 [17:48<15:12,  3.82s/it]
                                                 
{'loss': 1.3137, 'grad_norm': 0.001471888599793642, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [17:48<15:12,  3.82s/it]
 54%|█████▍    | 282/520 [17:51<15:06,  3.81s/it]
                                                 
{'loss': 1.1741, 'grad_norm': 0.0012707389516322391, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [17:51<15:06,  3.81s/it]
 54%|█████▍    | 283/520 [17:55<15:02,  3.81s/it]
                                                 
{'loss': 1.3265, 'grad_norm': 0.0015001530883195445, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [17:55<15:02,  3.81s/it]
 55%|█████▍    | 284/520 [17:59<14:56,  3.80s/it]
                                                 
{'loss': 1.199, 'grad_norm': 0.0013590690767206082, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [17:59<14:56,  3.80s/it]
 55%|█████▍    | 285/520 [18:03<14:52,  3.80s/it]
                                                 
{'loss': 1.2002, 'grad_norm': 0.0013513014272254781, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [18:03<14:52,  3.80s/it]
 55%|█████▌    | 286/520 [18:07<14:51,  3.81s/it]
                                                 
{'loss': 1.0799, 'grad_norm': 0.001545894125218819, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [18:07<14:51,  3.81s/it]
 55%|█████▌    | 287/520 [18:10<14:49,  3.82s/it]
                                                 
{'loss': 1.3071, 'grad_norm': 0.0013814796310562898, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [18:10<14:49,  3.82s/it]
 55%|█████▌    | 288/520 [18:14<14:45,  3.82s/it]
                                                 
{'loss': 1.3427, 'grad_norm': 0.0013709563289400044, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:14<14:45,  3.82s/it]
 56%|█████▌    | 289/520 [18:18<14:41,  3.82s/it]
                                                 
{'loss': 1.2127, 'grad_norm': 0.0012775540860405468, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:18<14:41,  3.82s/it]
 56%|█████▌    | 290/520 [18:22<14:38,  3.82s/it]
                                                 
{'loss': 1.1348, 'grad_norm': 0.0012599662387145023, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:22<14:38,  3.82s/it]
 56%|█████▌    | 291/520 [18:26<14:33,  3.81s/it]
                                                 
{'loss': 1.1862, 'grad_norm': 0.0013583713449205038, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:26<14:33,  3.81s/it]
 56%|█████▌    | 292/520 [18:30<14:30,  3.82s/it]
                                                 
{'loss': 1.2458, 'grad_norm': 0.0013500658829469613, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:30<14:30,  3.82s/it]
 56%|█████▋    | 293/520 [18:33<14:30,  3.84s/it]
                                                 
{'loss': 1.1864, 'grad_norm': 0.0014479824782472424, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:33<14:30,  3.84s/it]
 57%|█████▋    | 294/520 [18:37<14:26,  3.83s/it]
                                                 
{'loss': 1.2078, 'grad_norm': 0.001446540730929726, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:37<14:26,  3.83s/it]
 57%|█████▋    | 295/520 [18:41<14:21,  3.83s/it]
                                                 
{'loss': 1.2454, 'grad_norm': 0.0013177117565162462, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:41<14:21,  3.83s/it]
 57%|█████▋    | 296/520 [18:45<14:16,  3.83s/it]
                                                 
{'loss': 1.1552, 'grad_norm': 0.0013867552310700842, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [18:45<14:16,  3.83s/it]
 57%|█████▋    | 297/520 [18:49<14:14,  3.83s/it]
                                                 
{'loss': 1.2849, 'grad_norm': 0.001454519964676377, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [18:49<14:14,  3.83s/it]
 57%|█████▋    | 298/520 [18:53<14:13,  3.84s/it]
                                                 
{'loss': 1.2511, 'grad_norm': 0.0012796452089739617, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [18:53<14:13,  3.84s/it]
 57%|█████▊    | 299/520 [18:56<14:09,  3.84s/it]
                                                 
{'loss': 1.2738, 'grad_norm': 0.0012436254073286228, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [18:56<14:09,  3.84s/it]
 58%|█████▊    | 300/520 [19:00<14:04,  3.84s/it]
                                                 
{'loss': 1.3006, 'grad_norm': 0.0013165447378796942, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [19:00<14:04,  3.84s/it]
 58%|█████▊    | 301/520 [19:04<14:01,  3.84s/it]
                                                 
{'loss': 1.2792, 'grad_norm': 0.0013326472851437933, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [19:04<14:01,  3.84s/it]
 58%|█████▊    | 302/520 [19:08<13:56,  3.84s/it]
                                                 
{'loss': 1.2897, 'grad_norm': 0.001333906336377743, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [19:08<13:56,  3.84s/it]
 58%|█████▊    | 303/520 [19:12<13:52,  3.83s/it]
                                                 
{'loss': 1.2081, 'grad_norm': 0.001513927007899538, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [19:12<13:52,  3.83s/it]
 58%|█████▊    | 304/520 [19:16<13:49,  3.84s/it]
                                                 
{'loss': 1.1915, 'grad_norm': 0.0014412411471342913, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:16<13:49,  3.84s/it]
 59%|█████▊    | 305/520 [19:20<13:46,  3.85s/it]
                                                 
{'loss': 1.3078, 'grad_norm': 0.001497982344263626, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:20<13:46,  3.85s/it]
 59%|█████▉    | 306/520 [19:23<13:42,  3.84s/it]
                                                 
{'loss': 1.257, 'grad_norm': 0.0013885831923587666, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:23<13:42,  3.84s/it]
 59%|█████▉    | 307/520 [19:27<13:56,  3.93s/it]
                                                 
{'loss': 1.1908, 'grad_norm': 0.0012558819130713938, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:27<13:56,  3.93s/it]
 59%|█████▉    | 308/520 [19:31<13:47,  3.90s/it]
                                                 
{'loss': 1.3083, 'grad_norm': 0.001366342708807733, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:31<13:47,  3.90s/it]
 59%|█████▉    | 309/520 [19:35<13:40,  3.89s/it]
                                                 
{'loss': 1.1967, 'grad_norm': 0.0012350419637831135, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:35<13:40,  3.89s/it]
 60%|█████▉    | 310/520 [19:39<13:33,  3.87s/it]
                                                 
{'loss': 1.1726, 'grad_norm': 0.001362710378332638, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:39<13:33,  3.87s/it]
 60%|█████▉    | 311/520 [19:43<13:28,  3.87s/it]
                                                 
{'loss': 1.1459, 'grad_norm': 0.0013593253215597088, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [19:43<13:28,  3.87s/it]
 60%|██████    | 312/520 [19:47<13:23,  3.87s/it]
                                                 
{'loss': 1.1372, 'grad_norm': 0.0015158500796348562, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [19:47<13:23,  3.87s/it]
 60%|██████    | 313/520 [19:51<13:18,  3.86s/it]
                                                 
{'loss': 1.1256, 'grad_norm': 0.0012002363236157262, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [19:51<13:18,  3.86s/it]
 60%|██████    | 314/520 [19:55<13:39,  3.98s/it]
                                                 
{'loss': 1.1671, 'grad_norm': 0.0012571458723607378, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [19:55<13:39,  3.98s/it]
 61%|██████    | 315/520 [19:59<13:26,  3.93s/it]
                                                 
{'loss': 1.2385, 'grad_norm': 0.0015748260850795847, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [19:59<13:26,  3.93s/it]
 61%|██████    | 316/520 [20:03<13:45,  4.05s/it]
                                                 
{'loss': 1.1451, 'grad_norm': 0.001584429192106464, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [20:03<13:45,  4.05s/it]
 61%|██████    | 317/520 [20:07<13:29,  3.99s/it]
                                                 
{'loss': 1.1583, 'grad_norm': 0.0012525822729273224, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [20:07<13:29,  3.99s/it]
 61%|██████    | 318/520 [20:11<13:16,  3.94s/it]
                                                 
{'loss': 1.2704, 'grad_norm': 0.001490042220206525, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [20:11<13:16,  3.94s/it]
 61%|██████▏   | 319/520 [20:15<13:32,  4.04s/it]
                                                 
{'loss': 1.1445, 'grad_norm': 0.0012302872164355115, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [20:15<13:32,  4.04s/it]
 62%|██████▏   | 320/520 [20:19<13:17,  3.99s/it]
                                                 
{'loss': 1.0922, 'grad_norm': 0.0013358403278745901, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:19<13:17,  3.99s/it]
 62%|██████▏   | 321/520 [20:23<13:04,  3.94s/it]
                                                 
{'loss': 1.2878, 'grad_norm': 0.0014592745614526528, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:23<13:04,  3.94s/it]
 62%|██████▏   | 322/520 [20:26<12:53,  3.91s/it]
                                                 
{'loss': 1.133, 'grad_norm': 0.0013121784999797503, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:26<12:53,  3.91s/it]
 62%|██████▏   | 323/520 [20:30<12:45,  3.89s/it]
                                                 
{'loss': 1.2085, 'grad_norm': 0.0014208915932582633, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:30<12:45,  3.89s/it]
 62%|██████▏   | 324/520 [20:34<12:40,  3.88s/it]
                                                 
{'loss': 1.2264, 'grad_norm': 0.001335204171533113, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:34<12:40,  3.88s/it]
 62%|██████▎   | 325/520 [20:38<12:34,  3.87s/it]
                                                 
{'loss': 1.23, 'grad_norm': 0.0013918091927954531, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:38<12:34,  3.87s/it]
 63%|██████▎   | 326/520 [20:42<12:32,  3.88s/it]
                                                 
{'loss': 1.2242, 'grad_norm': 0.0013934021490694885, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:42<12:32,  3.88s/it]
 63%|██████▎   | 327/520 [20:46<12:24,  3.86s/it]
                                                 
{'loss': 1.2513, 'grad_norm': 0.0014598382866702194, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [20:46<12:24,  3.86s/it]
 63%|██████▎   | 328/520 [20:50<12:21,  3.86s/it]
                                                 
{'loss': 1.271, 'grad_norm': 0.0013896244197546166, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [20:50<12:21,  3.86s/it]
 63%|██████▎   | 329/520 [20:53<12:15,  3.85s/it]
                                                 
{'loss': 1.1451, 'grad_norm': 0.0011638913686225552, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [20:53<12:15,  3.85s/it]
 63%|██████▎   | 330/520 [20:57<12:08,  3.84s/it]
                                                 
{'loss': 1.2171, 'grad_norm': 0.0012468397780071255, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [20:57<12:08,  3.84s/it]
 64%|██████▎   | 331/520 [21:01<12:03,  3.83s/it]
                                                 
{'loss': 1.1795, 'grad_norm': 0.001282225791493693, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [21:01<12:03,  3.83s/it]
 64%|██████▍   | 332/520 [21:05<11:58,  3.82s/it]
                                                 
{'loss': 1.2777, 'grad_norm': 0.001284038969251486, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [21:05<11:58,  3.82s/it]
 64%|██████▍   | 333/520 [21:09<11:54,  3.82s/it]
                                                 
{'loss': 1.3199, 'grad_norm': 0.001414718743030883, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [21:09<11:54,  3.82s/it]
 64%|██████▍   | 334/520 [21:12<11:50,  3.82s/it]
                                                 
{'loss': 1.2261, 'grad_norm': 0.0014016743341685384, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [21:12<11:50,  3.82s/it]
 64%|██████▍   | 335/520 [21:16<11:48,  3.83s/it]
                                                 
{'loss': 1.2286, 'grad_norm': 0.0012365890796778368, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [21:16<11:48,  3.83s/it]
 65%|██████▍   | 336/520 [21:20<11:38,  3.79s/it]
                                                 
{'loss': 1.1253, 'grad_norm': 0.001506700904016684, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:20<11:38,  3.79s/it]
 65%|██████▍   | 337/520 [21:24<11:29,  3.77s/it]
                                                 
{'loss': 1.1078, 'grad_norm': 0.001276924121050403, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:24<11:29,  3.77s/it]
 65%|██████▌   | 338/520 [21:27<11:20,  3.74s/it]
                                                 
{'loss': 1.2248, 'grad_norm': 0.0012986254576040728, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:27<11:20,  3.74s/it]
 65%|██████▌   | 339/520 [21:31<11:13,  3.72s/it]
                                                 
{'loss': 1.1756, 'grad_norm': 0.0013065965063321875, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:31<11:13,  3.72s/it]
 65%|██████▌   | 340/520 [21:35<11:10,  3.73s/it]
                                                 
{'loss': 1.1672, 'grad_norm': 0.0013147780620713335, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:35<11:10,  3.73s/it]
 66%|██████▌   | 341/520 [21:39<11:04,  3.71s/it]
                                                 
{'loss': 1.1868, 'grad_norm': 0.0014128890544196058, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:39<11:04,  3.71s/it]
 66%|██████▌   | 342/520 [21:42<11:02,  3.72s/it]
                                                 
{'loss': 1.2447, 'grad_norm': 0.0015370170181413264, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:42<11:02,  3.72s/it]
 66%|██████▌   | 343/520 [21:46<11:01,  3.74s/it]
                                                 
{'loss': 1.1982, 'grad_norm': 0.001260411619210508, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [21:46<11:01,  3.74s/it]
 66%|██████▌   | 344/520 [21:50<11:03,  3.77s/it]
                                                 
{'loss': 1.1407, 'grad_norm': 0.001276562114324582, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [21:50<11:03,  3.77s/it]
 66%|██████▋   | 345/520 [21:54<11:00,  3.77s/it]
                                                 
{'loss': 1.2544, 'grad_norm': 0.0014338525482333047, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [21:54<11:00,  3.77s/it]
 67%|██████▋   | 346/520 [21:57<10:50,  3.74s/it]
                                                 
{'loss': 1.2094, 'grad_norm': 0.0012311964951036561, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [21:57<10:50,  3.74s/it]
 67%|██████▋   | 347/520 [22:01<10:41,  3.71s/it]
                                                 
{'loss': 1.1585, 'grad_norm': 0.0012315748170868444, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [22:01<10:41,  3.71s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [22:05<10:35,  3.69s/it]
                                                 
{'loss': 1.1118, 'grad_norm': 0.001496988882110305, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [22:05<10:35,  3.69s/it]
 67%|██████▋   | 349/520 [22:08<10:29,  3.68s/it]
                                                 
{'loss': 1.1545, 'grad_norm': 0.0013321100308003385, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [22:08<10:29,  3.68s/it]
 67%|██████▋   | 350/520 [22:12<10:24,  3.68s/it]
                                                 
{'loss': 1.1987, 'grad_norm': 0.0014429136190854822, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [22:12<10:24,  3.68s/it]
 68%|██████▊   | 351/520 [22:16<10:21,  3.68s/it]
                                                 
{'loss': 1.1079, 'grad_norm': 0.0012920691244930724, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [22:16<10:21,  3.68s/it]
 68%|██████▊   | 352/520 [22:19<10:18,  3.68s/it]
                                                 
{'loss': 1.2311, 'grad_norm': 0.0012985380597232647, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [22:19<10:18,  3.68s/it]
 68%|██████▊   | 353/520 [22:23<10:17,  3.69s/it]
                                                 
{'loss': 1.1626, 'grad_norm': 0.0011146709359956006, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:23<10:17,  3.69s/it]
 68%|██████▊   | 354/520 [22:27<10:11,  3.68s/it]
                                                 
{'loss': 1.2855, 'grad_norm': 0.0012215570355233276, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:27<10:11,  3.68s/it]
 68%|██████▊   | 355/520 [22:30<10:05,  3.67s/it]
                                                 
{'loss': 1.1694, 'grad_norm': 0.001301692436378191, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:30<10:05,  3.67s/it]
 68%|██████▊   | 356/520 [22:34<10:11,  3.73s/it]
                                                 
{'loss': 1.1698, 'grad_norm': 0.001342741732585784, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:34<10:11,  3.73s/it]
 69%|██████▊   | 357/520 [22:38<10:15,  3.78s/it]
                                                 
{'loss': 1.2002, 'grad_norm': 0.0012699476467568332, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:38<10:15,  3.78s/it]
 69%|██████▉   | 358/520 [22:42<10:18,  3.82s/it]
                                                 
{'loss': 1.1317, 'grad_norm': 0.0012715559733760889, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:42<10:18,  3.82s/it]
 69%|██████▉   | 359/520 [22:46<10:07,  3.77s/it]
                                                 
{'loss': 1.2155, 'grad_norm': 0.0013340351539265554, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [22:46<10:07,  3.77s/it]
 69%|██████▉   | 360/520 [22:49<09:58,  3.74s/it]
                                                 
{'loss': 1.2237, 'grad_norm': 0.0013417338749314484, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [22:49<09:58,  3.74s/it]
 69%|██████▉   | 361/520 [22:53<09:53,  3.73s/it]
                                                 
{'loss': 1.2335, 'grad_norm': 0.0012103880645066413, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [22:53<09:53,  3.73s/it]
 70%|██████▉   | 362/520 [22:57<09:44,  3.70s/it]
                                                 
{'loss': 1.1857, 'grad_norm': 0.0015515288277296148, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [22:57<09:44,  3.70s/it]
 70%|██████▉   | 363/520 [23:00<09:39,  3.69s/it]
                                                 
{'loss': 1.2065, 'grad_norm': 0.001301259815835577, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [23:00<09:39,  3.69s/it]
 70%|███████   | 364/520 [23:04<09:37,  3.70s/it]
                                                 
{'loss': 1.2542, 'grad_norm': 0.00125795986917179, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [23:04<09:37,  3.70s/it]
 70%|███████   | 365/520 [23:08<09:30,  3.68s/it]
                                                 
{'loss': 1.2682, 'grad_norm': 0.0014242976279781596, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [23:08<09:30,  3.68s/it]
 70%|███████   | 366/520 [23:11<09:26,  3.68s/it]
                                                 
{'loss': 1.2232, 'grad_norm': 0.0013247648348886658, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [23:11<09:26,  3.68s/it]
 71%|███████   | 367/520 [23:15<09:24,  3.69s/it]
                                                 
{'loss': 1.2222, 'grad_norm': 0.0013059355341866199, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [23:15<09:24,  3.69s/it]
 71%|███████   | 368/520 [23:19<09:20,  3.69s/it]
                                                 
{'loss': 1.0775, 'grad_norm': 0.0013778116179042706, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [23:19<09:20,  3.69s/it]
 71%|███████   | 369/520 [23:22<09:15,  3.68s/it]
                                                 
{'loss': 1.2084, 'grad_norm': 0.0011627616937706765, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:22<09:15,  3.68s/it]
 71%|███████   | 370/520 [23:26<09:11,  3.68s/it]
                                                 
{'loss': 1.1356, 'grad_norm': 0.0012116995535653378, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:26<09:11,  3.68s/it]
 71%|███████▏  | 371/520 [23:30<09:06,  3.67s/it]
                                                 
{'loss': 1.1344, 'grad_norm': 0.0013464187217560944, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:30<09:06,  3.67s/it]
 72%|███████▏  | 372/520 [23:33<09:03,  3.67s/it]
                                                 
{'loss': 1.2862, 'grad_norm': 0.0011838860934081112, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:33<09:03,  3.67s/it]
 72%|███████▏  | 373/520 [23:37<09:00,  3.67s/it]
                                                 
{'loss': 1.1731, 'grad_norm': 0.001344499386851558, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:37<09:00,  3.67s/it]
 72%|███████▏  | 374/520 [23:41<08:56,  3.68s/it]
                                                 
{'loss': 1.2187, 'grad_norm': 0.001291136781446381, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:41<08:56,  3.68s/it]
 72%|███████▏  | 375/520 [23:44<08:53,  3.68s/it]
                                                 
{'loss': 1.1393, 'grad_norm': 0.0013252103563978223, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [23:44<08:53,  3.68s/it]
 72%|███████▏  | 376/520 [23:48<08:51,  3.69s/it]
                                                 
{'loss': 1.2556, 'grad_norm': 0.001296907290187708, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [23:48<08:51,  3.69s/it]
 72%|███████▎  | 377/520 [23:52<08:46,  3.68s/it]
                                                 
{'loss': 1.1842, 'grad_norm': 0.0013234620410272529, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [23:52<08:46,  3.68s/it]
 73%|███████▎  | 378/520 [23:56<08:42,  3.68s/it]
                                                 
{'loss': 1.2446, 'grad_norm': 0.001289751626076243, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [23:56<08:42,  3.68s/it]
 73%|███████▎  | 379/520 [23:59<08:39,  3.68s/it]
                                                 
{'loss': 1.2224, 'grad_norm': 0.0012367067523967874, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [23:59<08:39,  3.68s/it]
 73%|███████▎  | 380/520 [24:03<08:34,  3.68s/it]
                                                 
{'loss': 1.2681, 'grad_norm': 0.0013315924691227167, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [24:03<08:34,  3.68s/it]
 73%|███████▎  | 381/520 [24:07<08:31,  3.68s/it]
                                                 
{'loss': 1.2241, 'grad_norm': 0.001274546209818559, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [24:07<08:31,  3.68s/it]
 73%|███████▎  | 382/520 [24:10<08:28,  3.68s/it]
                                                 
{'loss': 1.2201, 'grad_norm': 0.0012684265019709301, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [24:10<08:28,  3.68s/it]
 74%|███████▎  | 383/520 [24:14<08:23,  3.68s/it]
                                                 
{'loss': 1.059, 'grad_norm': 0.001399157750695273, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [24:14<08:23,  3.68s/it]
 74%|███████▍  | 384/520 [24:18<08:18,  3.67s/it]
                                                 
{'loss': 1.2753, 'grad_norm': 0.0011799524769092482, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [24:18<08:18,  3.67s/it]
 74%|███████▍  | 385/520 [24:21<08:13,  3.65s/it]
                                                 
{'loss': 1.1994, 'grad_norm': 0.0012248490001901984, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:21<08:13,  3.65s/it]
 74%|███████▍  | 386/520 [24:25<08:10,  3.66s/it]
                                                 
{'loss': 1.1514, 'grad_norm': 0.0011285913824787054, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:25<08:10,  3.66s/it]
 74%|███████▍  | 387/520 [24:29<08:07,  3.66s/it]
                                                 
{'loss': 1.2831, 'grad_norm': 0.0012441563006599962, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:29<08:07,  3.66s/it]
 75%|███████▍  | 388/520 [24:32<08:03,  3.66s/it]
                                                 
{'loss': 1.1027, 'grad_norm': 0.001213860812641313, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:32<08:03,  3.66s/it]
 75%|███████▍  | 389/520 [24:36<07:59,  3.66s/it]
                                                 
{'loss': 1.1556, 'grad_norm': 0.0015495538052014948, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}
+
 75%|███████▍  | 389/520 [24:36<07:59,  3.66s/it]
 75%|███████▌  | 390/520 [24:40<07:57,  3.68s/it]
                                                 
{'loss': 1.2216, 'grad_norm': 0.0012882243286992793, 'learning_rate': 0.031074166151605298, 'epoch': 0.75}
+
 75%|███████▌  | 390/520 [24:40<07:57,  3.68s/it]
 75%|███████▌  | 391/520 [24:43<07:55,  3.69s/it]
                                                 
{'loss': 1.2952, 'grad_norm': 0.0013652915810865261, 'learning_rate': 0.030623894397145836, 'epoch': 0.75}
+
 75%|███████▌  | 391/520 [24:43<07:55,  3.69s/it]
 75%|███████▌  | 392/520 [24:47<07:51,  3.69s/it]
                                                 
{'loss': 1.1104, 'grad_norm': 0.0012794685739945398, 'learning_rate': 0.03017631819139273, 'epoch': 0.75}
+
 75%|███████▌  | 392/520 [24:47<07:51,  3.69s/it]
 76%|███████▌  | 393/520 [24:51<07:47,  3.68s/it]
                                                 
{'loss': 1.1191, 'grad_norm': 0.001123559623475409, 'learning_rate': 0.029731454924533086, 'epoch': 0.76}
+
 76%|███████▌  | 393/520 [24:51<07:47,  3.68s/it]
 76%|███████▌  | 394/520 [24:54<07:43,  3.68s/it]
                                                 
{'loss': 1.1722, 'grad_norm': 0.0014466760000275266, 'learning_rate': 0.029289321881345254, 'epoch': 0.76}
+
 76%|███████▌  | 394/520 [24:54<07:43,  3.68s/it]
 76%|███████▌  | 395/520 [24:58<07:39,  3.67s/it]
                                                 
{'loss': 1.1366, 'grad_norm': 0.0013803711673005667, 'learning_rate': 0.028849936240527008, 'epoch': 0.76}
+
 76%|███████▌  | 395/520 [24:58<07:39,  3.67s/it]
 76%|███████▌  | 396/520 [25:02<07:35,  3.67s/it]
                                                 
{'loss': 1.2185, 'grad_norm': 0.0013802406720506983, 'learning_rate': 0.028413315074028157, 'epoch': 0.76}
+
 76%|███████▌  | 396/520 [25:02<07:35,  3.67s/it]
 76%|███████▋  | 397/520 [25:05<07:33,  3.69s/it]
                                                 
{'loss': 1.199, 'grad_norm': 0.0012301428108791527, 'learning_rate': 0.027979475346387363, 'epoch': 0.76}
+
 76%|███████▋  | 397/520 [25:05<07:33,  3.69s/it]
 77%|███████▋  | 398/520 [25:09<07:30,  3.69s/it]
                                                 
{'loss': 1.1976, 'grad_norm': 0.001362583984132521, 'learning_rate': 0.027548433914072735, 'epoch': 0.77}
+
 77%|███████▋  | 398/520 [25:09<07:30,  3.69s/it]
 77%|███████▋  | 399/520 [25:13<07:27,  3.70s/it]
                                                 
{'loss': 1.1658, 'grad_norm': 0.0012651420867766765, 'learning_rate': 0.027120207524827168, 'epoch': 0.77}
+
 77%|███████▋  | 399/520 [25:13<07:27,  3.70s/it]
 77%|███████▋  | 400/520 [25:17<07:25,  3.71s/it]
                                                 
{'loss': 1.1961, 'grad_norm': 0.0011876347880315071, 'learning_rate': 0.02669481281701739, 'epoch': 0.77}
+
 77%|███████▋  | 400/520 [25:17<07:25,  3.71s/it]
 77%|███████▋  | 401/520 [25:20<07:20,  3.70s/it]
                                                 
{'loss': 1.0361, 'grad_norm': 0.0014145300943584922, 'learning_rate': 0.026272266318987603, 'epoch': 0.77}
+
 77%|███████▋  | 401/520 [25:20<07:20,  3.70s/it]
 77%|███████▋  | 402/520 [25:24<07:17,  3.70s/it]
                                                 
{'loss': 1.156, 'grad_norm': 0.0013319696665051457, 'learning_rate': 0.02585258444841733, 'epoch': 0.77}
+
 77%|███████▋  | 402/520 [25:24<07:17,  3.70s/it]
 78%|███████▊  | 403/520 [25:28<07:14,  3.71s/it]
                                                 
{'loss': 1.1833, 'grad_norm': 0.0014350731842111778, 'learning_rate': 0.025435783511683442, 'epoch': 0.78}
+
 78%|███████▊  | 403/520 [25:28<07:14,  3.71s/it]
 78%|███████▊  | 404/520 [25:31<07:09,  3.71s/it]
                                                 
{'loss': 1.089, 'grad_norm': 0.0015460077773308406, 'learning_rate': 0.02502187970322657, 'epoch': 0.78}
+
 78%|███████▊  | 404/520 [25:31<07:09,  3.71s/it]
 78%|███████▊  | 405/520 [25:35<07:06,  3.71s/it]
                                                 
{'loss': 1.1728, 'grad_norm': 0.0013308508387264676, 'learning_rate': 0.02461088910492202, 'epoch': 0.78}
+
 78%|███████▊  | 405/520 [25:35<07:06,  3.71s/it]
 78%|███████▊  | 406/520 [25:39<07:02,  3.71s/it]
                                                 
{'loss': 1.0934, 'grad_norm': 0.001540996308576223, 'learning_rate': 0.02420282768545469, 'epoch': 0.78}
+
 78%|███████▊  | 406/520 [25:39<07:02,  3.71s/it]
 78%|███████▊  | 407/520 [25:42<07:00,  3.72s/it]
                                                 
{'loss': 1.2667, 'grad_norm': 0.0013380837788429707, 'learning_rate': 0.02379771129969892, 'epoch': 0.78}
+
 78%|███████▊  | 407/520 [25:43<07:00,  3.72s/it]
 78%|███████▊  | 408/520 [25:46<06:56,  3.72s/it]
                                                 
{'loss': 1.1691, 'grad_norm': 0.001393614302105086, 'learning_rate': 0.023395555688102213, 'epoch': 0.78}
+
 78%|███████▊  | 408/520 [25:46<06:56,  3.72s/it]
 79%|███████▊  | 409/520 [25:50<06:51,  3.71s/it]
                                                 
{'loss': 1.2886, 'grad_norm': 0.0014089680682511628, 'learning_rate': 0.02299637647607372, 'epoch': 0.79}
+
 79%|███████▊  | 409/520 [25:50<06:51,  3.71s/it]
 79%|███████▉  | 410/520 [25:54<06:46,  3.70s/it]
                                                 
{'loss': 1.0215, 'grad_norm': 0.001280735054159039, 'learning_rate': 0.022600189173377264, 'epoch': 0.79}
+
 79%|███████▉  | 410/520 [25:54<06:46,  3.70s/it]
 79%|███████▉  | 411/520 [25:57<06:41,  3.68s/it]
                                                 
{'loss': 1.2665, 'grad_norm': 0.0014729329851244758, 'learning_rate': 0.022207009173528525, 'epoch': 0.79}
+
 79%|███████▉  | 411/520 [25:57<06:41,  3.68s/it]
 79%|███████▉  | 412/520 [26:01<06:37,  3.68s/it]
                                                 
{'loss': 1.1792, 'grad_norm': 0.00130283717699585, 'learning_rate': 0.02181685175319702, 'epoch': 0.79}
+
 79%|███████▉  | 412/520 [26:01<06:37,  3.68s/it]
 79%|███████▉  | 413/520 [26:05<06:35,  3.70s/it]
                                                 
{'loss': 1.1861, 'grad_norm': 0.0011910187881029122, 'learning_rate': 0.021429732071612653, 'epoch': 0.79}
+
 79%|███████▉  | 413/520 [26:05<06:35,  3.70s/it]
 80%|███████▉  | 414/520 [26:08<06:31,  3.69s/it]
                                                 
{'loss': 0.9969, 'grad_norm': 0.0010834270677189293, 'learning_rate': 0.02104566516997647, 'epoch': 0.8}
+
 80%|███████▉  | 414/520 [26:08<06:31,  3.69s/it]
 80%|███████▉  | 415/520 [26:12<06:28,  3.70s/it]
                                                 
{'loss': 1.1532, 'grad_norm': 0.0012175346673235095, 'learning_rate': 0.020664665970876496, 'epoch': 0.8}
+
 80%|███████▉  | 415/520 [26:12<06:28,  3.70s/it]
 80%|████████  | 416/520 [26:16<06:24,  3.70s/it]
                                                 
{'loss': 1.0729, 'grad_norm': 0.0013670489640025642, 'learning_rate': 0.020286749277707784, 'epoch': 0.8}
+
 80%|████████  | 416/520 [26:16<06:24,  3.70s/it]
 80%|████████  | 417/520 [26:19<06:20,  3.69s/it]
                                                 
{'loss': 1.2342, 'grad_norm': 0.0013685204092066365, 'learning_rate': 0.019911929774097215, 'epoch': 0.8}
+
 80%|████████  | 417/520 [26:19<06:20,  3.69s/it]
 80%|████████  | 418/520 [26:23<06:17,  3.70s/it]
                                                 
{'loss': 1.2197, 'grad_norm': 0.001272719414895202, 'learning_rate': 0.019540222023333165, 'epoch': 0.8}
+
 80%|████████  | 418/520 [26:23<06:17,  3.70s/it]
 81%|████████  | 419/520 [26:27<06:14,  3.71s/it]
                                                 
{'loss': 1.2119, 'grad_norm': 0.0014111846812866337, 'learning_rate': 0.01917164046779948, 'epoch': 0.81}
+
 81%|████████  | 419/520 [26:27<06:14,  3.71s/it]
 81%|████████  | 420/520 [26:31<06:10,  3.70s/it]
                                                 
{'loss': 1.1041, 'grad_norm': 0.001399748654040505, 'learning_rate': 0.018806199428414352, 'epoch': 0.81}
+
 81%|████████  | 420/520 [26:31<06:10,  3.70s/it]
 81%|████████  | 421/520 [26:34<06:05,  3.69s/it]
                                                 
{'loss': 1.0368, 'grad_norm': 0.0015145541716717734, 'learning_rate': 0.018443913104073985, 'epoch': 0.81}
+
 81%|████████  | 421/520 [26:34<06:05,  3.69s/it]
 81%|████████  | 422/520 [26:38<06:00,  3.68s/it]
                                                 
{'loss': 1.1622, 'grad_norm': 0.0013332058743592502, 'learning_rate': 0.01808479557110081, 'epoch': 0.81}
+
 81%|████████  | 422/520 [26:38<06:00,  3.68s/it]
 81%|████████▏ | 423/520 [26:42<05:56,  3.67s/it]
                                                 
{'loss': 1.1397, 'grad_norm': 0.0014439305848749488, 'learning_rate': 0.017728860782696667, 'epoch': 0.81}
+
 81%|████████▏ | 423/520 [26:42<05:56,  3.67s/it]
 82%|████████▏ | 424/520 [26:45<05:53,  3.68s/it]
                                                 
{'loss': 1.2765, 'grad_norm': 0.0013263615712303853, 'learning_rate': 0.017376122568400532, 'epoch': 0.82}
+
 82%|████████▏ | 424/520 [26:45<05:53,  3.68s/it]
 82%|████████▏ | 425/520 [26:49<05:49,  3.68s/it]
                                                 
{'loss': 1.154, 'grad_norm': 0.0012762560309160138, 'learning_rate': 0.017026594633551252, 'epoch': 0.82}
+
 82%|████████▏ | 425/520 [26:49<05:49,  3.68s/it]
 82%|████████▏ | 426/520 [26:53<05:45,  3.67s/it]
                                                 
{'loss': 1.1742, 'grad_norm': 0.001625188278294579, 'learning_rate': 0.01668029055875512, 'epoch': 0.82}
+
 82%|████████▏ | 426/520 [26:53<05:45,  3.67s/it]
 82%|████████▏ | 427/520 [26:56<05:41,  3.67s/it]
                                                 
{'loss': 1.0843, 'grad_norm': 0.0012435597203359478, 'learning_rate': 0.016337223799358026, 'epoch': 0.82}
+
 82%|████████▏ | 427/520 [26:56<05:41,  3.67s/it]
 82%|████████▏ | 428/520 [27:00<05:37,  3.67s/it]
                                                 
{'loss': 1.0678, 'grad_norm': 0.001345614114169129, 'learning_rate': 0.01599740768492286, 'epoch': 0.82}
+
 82%|████████▏ | 428/520 [27:00<05:37,  3.67s/it]
 82%|████████▎ | 429/520 [27:04<05:33,  3.67s/it]
                                                 
{'loss': 1.1609, 'grad_norm': 0.0012963863207711018, 'learning_rate': 0.015660855418711452, 'epoch': 0.82}
+
 82%|████████▎ | 429/520 [27:04<05:33,  3.67s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2076 > 2048). Running this sequence through the model will result in indexing errors
+
 83%|████████▎ | 430/520 [27:07<05:30,  3.67s/it]
                                                 
{'loss': 1.1663, 'grad_norm': 0.001210379349762174, 'learning_rate': 0.015327580077171589, 'epoch': 0.83}
+
 83%|████████▎ | 430/520 [27:07<05:30,  3.67s/it]
 83%|████████▎ | 431/520 [27:11<05:27,  3.68s/it]
                                                 
{'loss': 1.1583, 'grad_norm': 0.0013499311077755528, 'learning_rate': 0.014997594609429088, 'epoch': 0.83}
+
 83%|████████▎ | 431/520 [27:11<05:27,  3.68s/it]
 83%|████████▎ | 432/520 [27:15<05:23,  3.68s/it]
                                                 
{'loss': 1.0737, 'grad_norm': 0.001344825090422715, 'learning_rate': 0.01467091183678444, 'epoch': 0.83}
+
 83%|████████▎ | 432/520 [27:15<05:23,  3.68s/it]
 83%|████████▎ | 433/520 [27:18<05:20,  3.69s/it]
                                                 
{'loss': 1.2019, 'grad_norm': 0.001255078167465542, 'learning_rate': 0.014347544452214867, 'epoch': 0.83}
+
 83%|████████▎ | 433/520 [27:18<05:20,  3.69s/it]
 83%|████████▎ | 434/520 [27:22<05:19,  3.71s/it]
                                                 
{'loss': 0.9536, 'grad_norm': 0.0013257211332379723, 'learning_rate': 0.014027505019880971, 'epoch': 0.83}
+
 83%|████████▎ | 434/520 [27:22<05:19,  3.71s/it]
 84%|████████▎ | 435/520 [27:26<05:18,  3.74s/it]
                                                 
{'loss': 1.2431, 'grad_norm': 0.0014506019635190622, 'learning_rate': 0.013710805974638696, 'epoch': 0.84}
+
 84%|████████▎ | 435/520 [27:26<05:18,  3.74s/it]
 84%|████████▍ | 436/520 [27:30<05:14,  3.74s/it]
                                                 
{'loss': 1.0416, 'grad_norm': 0.0013821849589997276, 'learning_rate': 0.01339745962155613, 'epoch': 0.84}
+
 84%|████████▍ | 436/520 [27:30<05:14,  3.74s/it]
 84%|████████▍ | 437/520 [27:33<05:08,  3.72s/it]
                                                 
{'loss': 1.2644, 'grad_norm': 0.001318848146842099, 'learning_rate': 0.01308747813543536, 'epoch': 0.84}
+
 84%|████████▍ | 437/520 [27:33<05:08,  3.72s/it]
 84%|████████▍ | 438/520 [27:37<05:03,  3.70s/it]
                                                 
{'loss': 1.0822, 'grad_norm': 0.0012513975177630573, 'learning_rate': 0.012780873560339467, 'epoch': 0.84}
+
 84%|████████▍ | 438/520 [27:37<05:03,  3.70s/it]
 84%|████████▍ | 439/520 [27:41<04:59,  3.70s/it]
                                                 
{'loss': 1.1389, 'grad_norm': 0.0010805295590861803, 'learning_rate': 0.012477657809124632, 'epoch': 0.84}
+
 84%|████████▍ | 439/520 [27:41<04:59,  3.70s/it]
 85%|████████▍ | 440/520 [27:44<04:56,  3.70s/it]
                                                 
{'loss': 1.1147, 'grad_norm': 0.0012843896990488526, 'learning_rate': 0.012177842662977134, 'epoch': 0.85}
+
 85%|████████▍ | 440/520 [27:44<04:56,  3.70s/it]
 85%|████████▍ | 441/520 [27:48<04:51,  3.69s/it]
                                                 
{'loss': 1.1649, 'grad_norm': 0.0012880965580576097, 'learning_rate': 0.01188143977095576, 'epoch': 0.85}
+
 85%|████████▍ | 441/520 [27:48<04:51,  3.69s/it]
 85%|████████▌ | 442/520 [27:52<04:47,  3.69s/it]
                                                 
{'loss': 1.183, 'grad_norm': 0.0014301015054219975, 'learning_rate': 0.011588460649539035, 'epoch': 0.85}
+
 85%|████████▌ | 442/520 [27:52<04:47,  3.69s/it]
 85%|████████▌ | 443/520 [27:55<04:43,  3.69s/it]
                                                 
{'loss': 1.1958, 'grad_norm': 0.0012876968177064029, 'learning_rate': 0.011298916682177829, 'epoch': 0.85}
+
 85%|████████▌ | 443/520 [27:55<04:43,  3.69s/it]
 85%|████████▌ | 444/520 [27:59<04:39,  3.68s/it]
                                                 
{'loss': 1.1593, 'grad_norm': 0.0012113305300910139, 'learning_rate': 0.011012819118853146, 'epoch': 0.85}
+
 85%|████████▌ | 444/520 [27:59<04:39,  3.68s/it]
 86%|████████▌ | 445/520 [28:03<04:35,  3.67s/it]
                                                 
{'loss': 1.0862, 'grad_norm': 0.001276674279417714, 'learning_rate': 0.01073017907563887, 'epoch': 0.86}
+
 86%|████████▌ | 445/520 [28:03<04:35,  3.67s/it]
 86%|████████▌ | 446/520 [28:06<04:31,  3.67s/it]
                                                 
{'loss': 1.234, 'grad_norm': 0.0012423106045434812, 'learning_rate': 0.010451007534269908, 'epoch': 0.86}
+
 86%|████████▌ | 446/520 [28:06<04:31,  3.67s/it]
 86%|████████▌ | 447/520 [28:10<04:28,  3.68s/it]
                                                 
{'loss': 1.1662, 'grad_norm': 0.0013119501574560572, 'learning_rate': 0.010175315341715598, 'epoch': 0.86}
+
 86%|████████▌ | 447/520 [28:10<04:28,  3.68s/it]
 86%|████████▌ | 448/520 [28:14<04:25,  3.69s/it]
                                                 
{'loss': 1.1554, 'grad_norm': 0.0013685066310742033, 'learning_rate': 0.009903113209758098, 'epoch': 0.86}
+
 86%|████████▌ | 448/520 [28:14<04:25,  3.69s/it]
 86%|████████▋ | 449/520 [28:17<04:22,  3.69s/it]
                                                 
{'loss': 1.1887, 'grad_norm': 0.0013003844791035354, 'learning_rate': 0.009634411714576352, 'epoch': 0.86}
+
 86%|████████▋ | 449/520 [28:18<04:22,  3.69s/it]
 87%|████████▋ | 450/520 [28:21<04:18,  3.69s/it]
                                                 
{'loss': 1.1873, 'grad_norm': 0.001324704371513428, 'learning_rate': 0.009369221296335007, 'epoch': 0.87}
+
 87%|████████▋ | 450/520 [28:21<04:18,  3.69s/it]
 87%|████████▋ | 451/520 [28:25<04:14,  3.69s/it]
                                                 
{'loss': 1.1832, 'grad_norm': 0.0013156664333136016, 'learning_rate': 0.009107552258778906, 'epoch': 0.87}
+
 87%|████████▋ | 451/520 [28:25<04:14,  3.69s/it]
 87%|████████▋ | 452/520 [28:29<04:10,  3.68s/it]
                                                 
{'loss': 1.2267, 'grad_norm': 0.001210483160214824, 'learning_rate': 0.008849414768832687, 'epoch': 0.87}
+
 87%|████████▋ | 452/520 [28:29<04:10,  3.68s/it]
 87%|████████▋ | 453/520 [28:32<04:06,  3.67s/it]
                                                 
{'loss': 1.2071, 'grad_norm': 0.0012947655433409293, 'learning_rate': 0.008594818856205699, 'epoch': 0.87}
+
 87%|████████▋ | 453/520 [28:32<04:06,  3.67s/it]
 87%|████████▋ | 454/520 [28:36<04:02,  3.67s/it]
                                                 
{'loss': 1.0947, 'grad_norm': 0.0013331537711642604, 'learning_rate': 0.00834377441300238, 'epoch': 0.87}
+
 87%|████████▋ | 454/520 [28:36<04:02,  3.67s/it]
 88%|████████▊ | 455/520 [28:40<03:58,  3.67s/it]
                                                 
{'loss': 1.2314, 'grad_norm': 0.0012994154852157024, 'learning_rate': 0.008096291193337934, 'epoch': 0.88}
+
 88%|████████▊ | 455/520 [28:40<03:58,  3.67s/it]
 88%|████████▊ | 456/520 [28:43<03:55,  3.68s/it]
                                                 
{'loss': 1.1556, 'grad_norm': 0.001330782275002396, 'learning_rate': 0.007852378812959226, 'epoch': 0.88}
+
 88%|████████▊ | 456/520 [28:43<03:55,  3.68s/it]
 88%|████████▊ | 457/520 [28:47<03:51,  3.68s/it]
                                                 
{'loss': 1.1284, 'grad_norm': 0.0011707125255720815, 'learning_rate': 0.007612046748871327, 'epoch': 0.88}
+
 88%|████████▊ | 457/520 [28:47<03:51,  3.68s/it]
 88%|████████▊ | 458/520 [28:51<03:48,  3.69s/it]
                                                 
{'loss': 1.2889, 'grad_norm': 0.0014261874483088202, 'learning_rate': 0.007375304338969136, 'epoch': 0.88}
+
 88%|████████▊ | 458/520 [28:51<03:48,  3.69s/it]
 88%|████████▊ | 459/520 [28:54<03:45,  3.69s/it]
                                                 
{'loss': 1.2208, 'grad_norm': 0.001360417902500949, 'learning_rate': 0.007142160781674645, 'epoch': 0.88}
+
 88%|████████▊ | 459/520 [28:54<03:45,  3.69s/it]
 88%|████████▊ | 460/520 [28:58<03:41,  3.69s/it]
                                                 
{'loss': 1.1089, 'grad_norm': 0.0013199670410274616, 'learning_rate': 0.006912625135579587, 'epoch': 0.88}
+
 88%|████████▊ | 460/520 [28:58<03:41,  3.69s/it]
 89%|████████▊ | 461/520 [29:02<03:38,  3.70s/it]
                                                 
{'loss': 1.2038, 'grad_norm': 0.0010146409777413668, 'learning_rate': 0.0066867063190933496, 'epoch': 0.89}
+
 89%|████████▊ | 461/520 [29:02<03:38,  3.70s/it]
 89%|████████▉ | 462/520 [29:05<03:34,  3.69s/it]
                                                 
{'loss': 1.2817, 'grad_norm': 0.001249725145730227, 'learning_rate': 0.006464413110096601, 'epoch': 0.89}
+
 89%|████████▉ | 462/520 [29:05<03:34,  3.69s/it]
 89%|████████▉ | 463/520 [29:09<03:30,  3.69s/it]
                                                 
{'loss': 1.0669, 'grad_norm': 0.001351941298238371, 'learning_rate': 0.006245754145600091, 'epoch': 0.89}
+
 89%|████████▉ | 463/520 [29:09<03:30,  3.69s/it]
 89%|████████▉ | 464/520 [29:13<03:27,  3.70s/it]
                                                 
{'loss': 1.2048, 'grad_norm': 0.0014001046044584723, 'learning_rate': 0.006030737921409169, 'epoch': 0.89}
+
 89%|████████▉ | 464/520 [29:13<03:27,  3.70s/it]
 89%|████████▉ | 465/520 [29:17<03:23,  3.71s/it]
                                                 
{'loss': 1.31, 'grad_norm': 0.0014130976924363004, 'learning_rate': 0.005819372791793654, 'epoch': 0.89}
+
 89%|████████▉ | 465/520 [29:17<03:23,  3.71s/it]
 90%|████████▉ | 466/520 [29:20<03:20,  3.71s/it]
                                                 
{'loss': 1.1909, 'grad_norm': 0.0012000185923608162, 'learning_rate': 0.005611666969163243, 'epoch': 0.9}
+
 90%|████████▉ | 466/520 [29:20<03:20,  3.71s/it]
 90%|████████▉ | 467/520 [29:24<03:16,  3.70s/it]
                                                 
{'loss': 1.1643, 'grad_norm': 0.0011740337034183508, 'learning_rate': 0.005407628523748398, 'epoch': 0.9}
+
 90%|████████▉ | 467/520 [29:24<03:16,  3.70s/it]
 90%|█████████ | 468/520 [29:28<03:12,  3.70s/it]
                                                 
{'loss': 1.1681, 'grad_norm': 0.0014528498631169117, 'learning_rate': 0.00520726538328683, 'epoch': 0.9}
+
 90%|█████████ | 468/520 [29:28<03:12,  3.70s/it]
 90%|█████████ | 469/520 [29:31<03:08,  3.70s/it]
                                                 
{'loss': 1.2259, 'grad_norm': 0.0014344523614635326, 'learning_rate': 0.005010585332715401, 'epoch': 0.9}
+
 90%|█████████ | 469/520 [29:31<03:08,  3.70s/it]
 90%|█████████ | 470/520 [29:35<03:04,  3.69s/it]
                                                 
{'loss': 1.1067, 'grad_norm': 0.0011915035851414995, 'learning_rate': 0.004817596013867765, 'epoch': 0.9}
+
 90%|█████████ | 470/520 [29:35<03:04,  3.69s/it]
 91%|█████████ | 471/520 [29:39<03:00,  3.69s/it]
                                                 
{'loss': 1.1299, 'grad_norm': 0.001363362484525896, 'learning_rate': 0.004628304925177318, 'epoch': 0.91}
+
 91%|█████████ | 471/520 [29:39<03:00,  3.69s/it]
 91%|█████████ | 472/520 [29:42<02:58,  3.71s/it]
                                                 
{'loss': 1.0952, 'grad_norm': 0.001432619489988451, 'learning_rate': 0.004442719421385921, 'epoch': 0.91}
+
 91%|█████████ | 472/520 [29:42<02:58,  3.71s/it]
 91%|█████████ | 473/520 [29:46<02:54,  3.71s/it]
                                                 
{'loss': 1.1634, 'grad_norm': 0.0013510043026271827, 'learning_rate': 0.004260846713258193, 'epoch': 0.91}
+
 91%|█████████ | 473/520 [29:46<02:54,  3.71s/it]
 91%|█████████ | 474/520 [29:50<02:50,  3.71s/it]
                                                 
{'loss': 1.2023, 'grad_norm': 0.0012177871015096652, 'learning_rate': 0.004082693867301224, 'epoch': 0.91}
+
 91%|█████████ | 474/520 [29:50<02:50,  3.71s/it]
 91%|█████████▏| 475/520 [29:54<02:46,  3.70s/it]
                                                 
{'loss': 1.1185, 'grad_norm': 0.0012024370134693543, 'learning_rate': 0.003908267805490051, 'epoch': 0.91}
+
 91%|█████████▏| 475/520 [29:54<02:46,  3.70s/it]
 92%|█████████▏| 476/520 [29:57<02:42,  3.70s/it]
                                                 
{'loss': 1.1488, 'grad_norm': 0.0013192069909029082, 'learning_rate': 0.003737575304998797, 'epoch': 0.92}
+
 92%|█████████▏| 476/520 [29:57<02:42,  3.70s/it]
 92%|█████████▏| 477/520 [30:01<02:38,  3.69s/it]
                                                 
{'loss': 1.139, 'grad_norm': 0.0014496833993358426, 'learning_rate': 0.003570622997937234, 'epoch': 0.92}
+
 92%|█████████▏| 477/520 [30:01<02:38,  3.69s/it]
 92%|█████████▏| 478/520 [30:05<02:34,  3.69s/it]
                                                 
{'loss': 1.0938, 'grad_norm': 0.0013083212531977953, 'learning_rate': 0.00340741737109318, 'epoch': 0.92}
+
 92%|█████████▏| 478/520 [30:05<02:34,  3.69s/it]
 92%|█████████▏| 479/520 [30:08<02:30,  3.68s/it]
                                                 
{'loss': 1.1673, 'grad_norm': 0.0013656821348002291, 'learning_rate': 0.003247964765680389, 'epoch': 0.92}
+
 92%|█████████▏| 479/520 [30:08<02:30,  3.68s/it]
 92%|█████████▏| 480/520 [30:12<02:27,  3.68s/it]
                                                 
{'loss': 1.1887, 'grad_norm': 0.0011884932924553571, 'learning_rate': 0.0030922713770922153, 'epoch': 0.92}
+
 92%|█████████▏| 480/520 [30:12<02:27,  3.68s/it]
 92%|█████████▎| 481/520 [30:16<02:23,  3.69s/it]
                                                 
{'loss': 1.1814, 'grad_norm': 0.001198765885503492, 'learning_rate': 0.0029403432546609046, 'epoch': 0.93}
+
 92%|█████████▎| 481/520 [30:16<02:23,  3.69s/it]
 93%|█████████▎| 482/520 [30:19<02:20,  3.69s/it]
                                                 
{'loss': 1.1961, 'grad_norm': 0.0012482689871213986, 'learning_rate': 0.0027921863014225504, 'epoch': 0.93}
+
 93%|█████████▎| 482/520 [30:19<02:20,  3.69s/it]
 93%|█████████▎| 483/520 [30:23<02:16,  3.69s/it]
                                                 
{'loss': 1.16, 'grad_norm': 0.0013644143919224905, 'learning_rate': 0.002647806273887665, 'epoch': 0.93}
+
 93%|█████████▎| 483/520 [30:23<02:16,  3.69s/it]
 93%|█████████▎| 484/520 [30:27<02:13,  3.70s/it]
                                                 
{'loss': 1.1697, 'grad_norm': 0.00132840293953154, 'learning_rate': 0.0025072087818176383, 'epoch': 0.93}
+
 93%|█████████▎| 484/520 [30:27<02:13,  3.70s/it]
 93%|█████████▎| 485/520 [30:30<02:09,  3.70s/it]
                                                 
{'loss': 1.1207, 'grad_norm': 0.0012634402229455743, 'learning_rate': 0.002370399288006664, 'epoch': 0.93}
+
 93%|█████████▎| 485/520 [30:30<02:09,  3.70s/it]
 93%|█████████▎| 486/520 [30:34<02:05,  3.69s/it]
                                                 
{'loss': 1.2433, 'grad_norm': 0.0013651504138910242, 'learning_rate': 0.0022373831080695463, 'epoch': 0.93}
+
 93%|█████████▎| 486/520 [30:34<02:05,  3.69s/it]
 94%|█████████▎| 487/520 [30:38<02:01,  3.70s/it]
                                                 
{'loss': 1.0946, 'grad_norm': 0.001255030765285037, 'learning_rate': 0.0021081654102351635, 'epoch': 0.94}
+
 94%|█████████▎| 487/520 [30:38<02:01,  3.70s/it]
 94%|█████████▍| 488/520 [30:42<01:58,  3.69s/it]
                                                 
{'loss': 1.0405, 'grad_norm': 0.001400422217504602, 'learning_rate': 0.0019827512151456175, 'epoch': 0.94}
+
 94%|█████████▍| 488/520 [30:42<01:58,  3.69s/it]
 94%|█████████▍| 489/520 [30:45<01:54,  3.69s/it]
                                                 
{'loss': 1.1959, 'grad_norm': 0.0010981772335067884, 'learning_rate': 0.0018611453956612345, 'epoch': 0.94}
+
 94%|█████████▍| 489/520 [30:45<01:54,  3.69s/it]
 94%|█████████▍| 490/520 [30:49<01:50,  3.70s/it]
                                                 
{'loss': 1.1601, 'grad_norm': 0.0013191415726667617, 'learning_rate': 0.0017433526766711727, 'epoch': 0.94}
+
 94%|█████████▍| 490/520 [30:49<01:50,  3.70s/it]
 94%|█████████▍| 491/520 [30:53<01:47,  3.70s/it]
                                                 
{'loss': 1.1237, 'grad_norm': 0.0013500432983040412, 'learning_rate': 0.0016293776349098677, 'epoch': 0.94}
+
 94%|█████████▍| 491/520 [30:53<01:47,  3.70s/it]
 95%|█████████▍| 492/520 [30:56<01:43,  3.69s/it]
                                                 
{'loss': 1.2382, 'grad_norm': 0.0013741286862158676, 'learning_rate': 0.0015192246987791981, 'epoch': 0.95}
+
 95%|█████████▍| 492/520 [30:56<01:43,  3.69s/it]
 95%|█████████▍| 493/520 [31:00<01:39,  3.68s/it]
                                                 
{'loss': 1.2097, 'grad_norm': 0.001316658482458519, 'learning_rate': 0.0014128981481764114, 'epoch': 0.95}
+
 95%|█████████▍| 493/520 [31:00<01:39,  3.68s/it]
 95%|█████████▌| 494/520 [31:04<01:35,  3.68s/it]
                                                 
{'loss': 1.1767, 'grad_norm': 0.0011993131121554607, 'learning_rate': 0.0013104021143278911, 'epoch': 0.95}
+
 95%|█████████▌| 494/520 [31:04<01:35,  3.68s/it]
 95%|█████████▌| 495/520 [31:07<01:31,  3.67s/it]
                                                 
{'loss': 1.1448, 'grad_norm': 0.0013085833548957294, 'learning_rate': 0.0012117405796285285, 'epoch': 0.95}
+
 95%|█████████▌| 495/520 [31:07<01:31,  3.67s/it]
 95%|█████████▌| 496/520 [31:11<01:28,  3.68s/it]
                                                 
{'loss': 1.0602, 'grad_norm': 0.0013743333495448967, 'learning_rate': 0.0011169173774871477, 'epoch': 0.95}
+
 95%|█████████▌| 496/520 [31:11<01:28,  3.68s/it]
 96%|█████████▌| 497/520 [31:15<01:24,  3.68s/it]
                                                 
{'loss': 1.1307, 'grad_norm': 0.0011552448108171453, 'learning_rate': 0.0010259361921774012, 'epoch': 0.96}
+
 96%|█████████▌| 497/520 [31:15<01:24,  3.68s/it]
 96%|█████████▌| 498/520 [31:18<01:20,  3.67s/it]
                                                 
{'loss': 1.14, 'grad_norm': 0.0012982546643861466, 'learning_rate': 0.000938800558694719, 'epoch': 0.96}
+
 96%|█████████▌| 498/520 [31:18<01:20,  3.67s/it]
 96%|█████████▌| 499/520 [31:22<01:17,  3.68s/it]
                                                 
{'loss': 1.2675, 'grad_norm': 0.0013297679499253232, 'learning_rate': 0.0008555138626189618, 'epoch': 0.96}
+
 96%|█████████▌| 499/520 [31:22<01:17,  3.68s/it]
 96%|█████████▌| 500/520 [31:26<01:13,  3.68s/it]
                                                 
{'loss': 1.2606, 'grad_norm': 0.0015064680625154598, 'learning_rate': 0.0007760793399827937, 'epoch': 0.96}
+
 96%|█████████▌| 500/520 [31:26<01:13,  3.68s/it]
 96%|█████████▋| 501/520 [31:29<01:09,  3.67s/it]
                                                 
{'loss': 1.1759, 'grad_norm': 0.0014143606077368096, 'learning_rate': 0.000700500077146038, 'epoch': 0.96}
+
 96%|█████████▋| 501/520 [31:29<01:09,  3.67s/it]
 97%|█████████▋| 502/520 [31:33<01:06,  3.69s/it]
                                                 
{'loss': 1.1778, 'grad_norm': 0.0012308232037555275, 'learning_rate': 0.0006287790106757397, 'epoch': 0.97}
+
 97%|█████████▋| 502/520 [31:33<01:06,  3.69s/it]
 97%|█████████▋| 503/520 [31:37<01:02,  3.68s/it]
                                                 
{'loss': 1.1635, 'grad_norm': 0.0013317942395385445, 'learning_rate': 0.0005609189272320237, 'epoch': 0.97}
+
 97%|█████████▋| 503/520 [31:37<01:02,  3.68s/it]
 97%|█████████▋| 504/520 [31:40<00:58,  3.68s/it]
                                                 
{'loss': 1.1708, 'grad_norm': 0.0015642985730192494, 'learning_rate': 0.000496922463459859, 'epoch': 0.97}
+
 97%|█████████▋| 504/520 [31:40<00:58,  3.68s/it]
 97%|█████████▋| 505/520 [31:44<00:55,  3.68s/it]
                                                 
{'loss': 1.2052, 'grad_norm': 0.001305342640292229, 'learning_rate': 0.0004367921058866187, 'epoch': 0.97}
+
 97%|█████████▋| 505/520 [31:44<00:55,  3.68s/it]
 97%|█████████▋| 506/520 [31:48<00:51,  3.68s/it]
                                                 
{'loss': 1.13, 'grad_norm': 0.0013300623339534909, 'learning_rate': 0.0003805301908254455, 'epoch': 0.97}
+
 97%|█████████▋| 506/520 [31:48<00:51,  3.68s/it]
 98%|█████████▊| 507/520 [31:51<00:47,  3.68s/it]
                                                 
{'loss': 1.3063, 'grad_norm': 0.0011875170597056472, 'learning_rate': 0.0003281389042844918, 'epoch': 0.97}
+
 98%|█████████▊| 507/520 [31:51<00:47,  3.68s/it]
 98%|█████████▊| 508/520 [31:55<00:44,  3.67s/it]
                                                 
{'loss': 1.2513, 'grad_norm': 0.001338940480414633, 'learning_rate': 0.00027962028188198705, 'epoch': 0.98}
+
 98%|█████████▊| 508/520 [31:55<00:44,  3.67s/it]
 98%|█████████▊| 509/520 [31:59<00:40,  3.67s/it]
                                                 
{'loss': 1.219, 'grad_norm': 0.0013013801242119244, 'learning_rate': 0.00023497620876711256, 'epoch': 0.98}
+
 98%|█████████▊| 509/520 [31:59<00:40,  3.67s/it]
 98%|█████████▊| 510/520 [32:02<00:36,  3.68s/it]
                                                 
{'loss': 1.166, 'grad_norm': 0.0012804271260643764, 'learning_rate': 0.00019420841954681523, 'epoch': 0.98}
+
 98%|█████████▊| 510/520 [32:02<00:36,  3.68s/it]
 98%|█████████▊| 511/520 [32:06<00:33,  3.68s/it]
                                                 
{'loss': 1.1423, 'grad_norm': 0.0012334220649314988, 'learning_rate': 0.00015731849821833956, 'epoch': 0.98}
+
 98%|█████████▊| 511/520 [32:06<00:33,  3.68s/it]
 98%|█████████▊| 512/520 [32:10<00:29,  3.67s/it]
                                                 
{'loss': 1.0259, 'grad_norm': 0.0013298017227305154, 'learning_rate': 0.00012430787810776555, 'epoch': 0.98}
+
 98%|█████████▊| 512/520 [32:10<00:29,  3.67s/it]
 99%|█████████▊| 513/520 [32:13<00:25,  3.68s/it]
                                                 
{'loss': 1.2252, 'grad_norm': 0.0015063139877772545, 'learning_rate': 9.517784181422018e-05, 'epoch': 0.99}
+
 99%|█████████▊| 513/520 [32:13<00:25,  3.68s/it]
 99%|█████████▉| 514/520 [32:17<00:22,  3.68s/it]
                                                 
{'loss': 1.1953, 'grad_norm': 0.0012097495618544382, 'learning_rate': 6.992952116013917e-05, 'epoch': 0.99}
+
 99%|█████████▉| 514/520 [32:17<00:22,  3.68s/it]
 99%|█████████▉| 515/520 [32:21<00:18,  3.68s/it]
                                                 
{'loss': 1.2427, 'grad_norm': 0.001491031107596644, 'learning_rate': 4.856389714723575e-05, 'epoch': 0.99}
+
 99%|█████████▉| 515/520 [32:21<00:18,  3.68s/it]
 99%|█████████▉| 516/520 [32:25<00:14,  3.68s/it]
                                                 
{'loss': 1.1467, 'grad_norm': 0.0012671762005214164, 'learning_rate': 3.108179991837545e-05, 'epoch': 0.99}
+
 99%|█████████▉| 516/520 [32:25<00:14,  3.68s/it]
 99%|█████████▉| 517/520 [32:28<00:10,  3.66s/it]
                                                 
{'loss': 1.2008, 'grad_norm': 0.00124990227849876, 'learning_rate': 1.7483908725357544e-05, 'epoch': 0.99}
+
 99%|█████████▉| 517/520 [32:28<00:10,  3.66s/it]
100%|█████████▉| 518/520 [32:32<00:07,  3.64s/it]
                                                 
{'loss': 1.1607, 'grad_norm': 0.001354160069870508, 'learning_rate': 7.770751902513862e-06, 'epoch': 1.0}
+
100%|█████████▉| 518/520 [32:32<00:07,  3.64s/it]
100%|█████████▉| 519/520 [32:35<00:03,  3.64s/it]
                                                 
{'loss': 1.1698, 'grad_norm': 0.001257731166174193, 'learning_rate': 1.9427068461808086e-06, 'epoch': 1.0}
+
100%|█████████▉| 519/520 [32:35<00:03,  3.64s/it]
100%|██████████| 520/520 [32:40<00:00,  3.90s/it]
                                                 
{'loss': 1.1832, 'grad_norm': 0.0012429836458360264, 'learning_rate': 0.0, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:40<00:00,  3.90s/it]
                                                 
{'train_runtime': 1960.3849, 'train_samples_per_second': 33.937, 'train_steps_per_second': 0.265, 'train_loss': 1.267326688881104, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:40<00:00,  3.90s/it]
100%|██████████| 520/520 [32:40<00:00,  3.77s/it]
+[2025-10-16 18:03:09,060] [INFO] [launch.py:348:main] Process 2567440 exits successfully.
+[2025-10-16 18:03:10,062] [INFO] [launch.py:348:main] Process 2567439 exits successfully.
+[2025-10-16 18:03:10,063] [INFO] [launch.py:348:main] Process 2567438 exits successfully.
+[2025-10-16 18:03:10,063] [INFO] [launch.py:348:main] Process 2567436 exits successfully.
+[2025-10-16 18:03:10,063] [INFO] [launch.py:348:main] Process 2567437 exits successfully.
+[2025-10-16 18:03:10,064] [INFO] [launch.py:348:main] Process 2567435 exits successfully.
+[2025-10-16 18:03:10,064] [INFO] [launch.py:348:main] Process 2567434 exits successfully.
+[2025-10-16 18:03:14,069] [INFO] [launch.py:348:main] Process 2567433 exits successfully.
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-5.0_2.9_2e-1_connector-5.0_2.9_2e-1_ablation_20251016_172855.log
+Timestamp: 2025-10-16 18:03:16
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation_20251016_180316.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation_20251016_180316.log
new file mode 100644
index 0000000000000000000000000000000000000000..870efab86653b8c262c0701040fd07be97c8560a
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation_20251016_180316.log
@@ -0,0 +1,2312 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation_20251016_180316.log
+Timestamp: 2025-10-16 18:03:16
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 18:03:19,399] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:22,117] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 18:03:22,119] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 7.0 --temperature_attn_text 0.9 --temperature_mlp_text 0.9 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 7.0 --temperature_attn_vision 0.9 --temperature_mlp_vision 0.9 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 7.0 --temperature_connector 0.9 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 18:03:24,706] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:25,754] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 18:03:25,754] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 18:03:25,754] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 18:03:25,754] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 18:03:25,754] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 18:03:25,754] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 18:03:25,754] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 18:03:25,757] [INFO] [launch.py:253:main] process 2589801 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,758] [INFO] [launch.py:253:main] process 2589802 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,760] [INFO] [launch.py:253:main] process 2589803 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,762] [INFO] [launch.py:253:main] process 2589804 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,764] [INFO] [launch.py:253:main] process 2589805 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,766] [INFO] [launch.py:253:main] process 2589806 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,768] [INFO] [launch.py:253:main] process 2589807 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 18:03:25,770] [INFO] [launch.py:253:main] process 2589808 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '0.9', '--temperature_mlp_text', '0.9', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '0.9', '--temperature_mlp_vision', '0.9', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '0.9', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 18:03:32,323] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,618] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,630] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,714] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,722] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:32,729] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,729] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,735] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:32,749] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 18:03:33,023] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:33,033] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:33,121] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:33,131] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:33,132] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:33,132] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-10-16 18:03:33,140] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 18:03:33,152] [INFO] [comm.py:637:init_distributed] cdb=None
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 0.9, 'temperature_mlp': 0.9, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 0.9, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 0.9,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 0.9,
+    "temperature_mlp": 0.9,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2589801:2589801 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589801:2589801 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589801:2589801 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589801:2589801 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589801:2589801 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2589801:2589801 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2589808:2589808 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589808:2589808 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589808:2589808 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589808:2589808 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589808:2589808 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589808:2589808 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2589802:2589802 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589802:2589802 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589802:2589802 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589807:2589807 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589807:2589807 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589807:2589807 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589802:2589802 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589802:2589802 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589802:2589802 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2589807:2589807 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589807:2589807 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589807:2589807 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2589804:2589804 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589804:2589804 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589804:2589804 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589804:2589804 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589804:2589804 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589804:2589804 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589806:2589806 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589806:2589806 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589806:2589806 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589806:2589806 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589806:2589806 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589806:2589806 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2589803:2589803 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589803:2589803 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589803:2589803 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589803:2589803 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589803:2589803 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589803:2589803 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2589805:2589805 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2589805:2589805 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589805:2589805 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589805:2589805 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2589805:2589805 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2589805:2589805 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO ncclCommInitRank comm 0x55f84335ac20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO ncclCommInitRank comm 0x55a59954db10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO ncclCommInitRank comm 0x557d04754e80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO ncclCommInitRank comm 0x55d7be71f4e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO ncclCommInitRank comm 0x55cd8e62a1c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO ncclCommInitRank comm 0x555e1a53d7e0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO ncclCommInitRank comm 0x564ee48d6d80 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO ncclCommInitRank comm 0x562880a5a2c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xa2a2f61a36fb31c3 - Init START
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO comm 0x564ee48d6d80 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO comm 0x55f84335ac20 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO comm 0x55cd8e62a1c0 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO comm 0x557d04754e80 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO comm 0x55a59954db10 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO comm 0x55d7be71f4e0 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO comm 0x562880a5a2c0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO comm 0x555e1a53d7e0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589803:2591408 [2] NCCL INFO ncclCommInitRank comm 0x562880a5a2c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589804:2591406 [3] NCCL INFO ncclCommInitRank comm 0x557d04754e80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589802:2591404 [1] NCCL INFO ncclCommInitRank comm 0x55d7be71f4e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2589801:2591402 [0] NCCL INFO ncclCommInitRank comm 0x555e1a53d7e0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2589805:2591409 [4] NCCL INFO ncclCommInitRank comm 0x55f84335ac20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589806:2591407 [5] NCCL INFO ncclCommInitRank comm 0x55cd8e62a1c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589807:2591405 [6] NCCL INFO ncclCommInitRank comm 0x55a59954db10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589808:2591403 [7] NCCL INFO ncclCommInitRank comm 0x564ee48d6d80 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xa2a2f61a36fb31c3 - Init COMPLETE
+[2025-10-16 18:04:15,730] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 18:04:17,505] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=7.000000
+Pre-training init connector._connector.0.scores: Mean=7.000005
+Pre-training init connector._connector.2.scores: Mean=6.999971
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 18:28:05,645 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 18:28:05,651 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO ncclCommInitRank comm 0x7efa5406a910 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO ncclCommInitRank comm 0x7fdb34069f90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO ncclCommInitRank comm 0x7f7f0406b320 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO ncclCommInitRank comm 0x7faac406b170 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO ncclCommInitRank comm 0x7fca4006ac50 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO ncclCommInitRank comm 0x7ef76806b040 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO ncclCommInitRank comm 0x7fc6e806b340 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO ncclCommInitRank comm 0x7f784c06aeb0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x24f0794c1f15e32c - Init START
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO comm 0x7faac406b170 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO comm 0x7fdb34069f90 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO comm 0x7fc6e806b340 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO comm 0x7ef76806b040 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO comm 0x7efa5406a910 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO comm 0x7f7f0406b320 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO comm 0x7f784c06aeb0 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO comm 0x7fca4006ac50 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2589804:2596777 [3] NCCL INFO ncclCommInitRank comm 0x7f784c06aeb0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589808:2596774 [7] NCCL INFO ncclCommInitRank comm 0x7faac406b170 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589805:2596778 [4] NCCL INFO ncclCommInitRank comm 0x7fca4006ac50 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589802:2596775 [1] NCCL INFO ncclCommInitRank comm 0x7efa5406a910 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589801:2596771 [0] NCCL INFO ncclCommInitRank comm 0x7f7f0406b320 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589806:2596776 [5] NCCL INFO ncclCommInitRank comm 0x7ef76806b040 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589807:2596773 [6] NCCL INFO ncclCommInitRank comm 0x7fc6e806b340 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x24f0794c1f15e32c - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2589803:2596772 [2] NCCL INFO ncclCommInitRank comm 0x7fdb34069f90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x24f0794c1f15e32c - Init COMPLETE
+
  0%|          | 1/520 [00:14<2:04:53, 14.44s/it]
                                                 
{'loss': 2.0497, 'grad_norm': 0.0, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:14<2:04:53, 14.44s/it]
  0%|          | 2/520 [00:18<1:10:54,  8.21s/it]
                                                 
{'loss': 2.06, 'grad_norm': 0.0, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:18<1:10:54,  8.21s/it]
  1%|          | 3/520 [00:22<53:32,  6.21s/it]  
                                               
{'loss': 2.1958, 'grad_norm': 0.0, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:22<53:32,  6.21s/it]
  1%|          | 4/520 [00:25<44:56,  5.23s/it]
                                               
{'loss': 2.0688, 'grad_norm': 0.0, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:25<44:56,  5.23s/it]
  1%|          | 5/520 [00:29<39:58,  4.66s/it]
                                               
{'loss': 2.2403, 'grad_norm': 0.0, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:29<39:58,  4.66s/it]
  1%|          | 6/520 [00:33<36:56,  4.31s/it]
                                               
{'loss': 1.6782, 'grad_norm': 0.0, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:33<36:56,  4.31s/it]
  1%|▏         | 7/520 [00:36<34:56,  4.09s/it]
                                               
{'loss': 2.0829, 'grad_norm': 0.0, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:36<34:56,  4.09s/it]
  2%|▏         | 8/520 [00:40<35:15,  4.13s/it]
                                               
{'loss': 2.0585, 'grad_norm': 0.0, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:40<35:15,  4.13s/it]
  2%|▏         | 9/520 [00:45<35:06,  4.12s/it]
                                               
{'loss': 2.1936, 'grad_norm': 0.0, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:45<35:06,  4.12s/it]
  2%|▏         | 10/520 [00:48<33:40,  3.96s/it]
                                                
{'loss': 2.0887, 'grad_norm': 0.0, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:48<33:40,  3.96s/it]
  2%|▏         | 11/520 [00:52<33:24,  3.94s/it]
                                                
{'loss': 2.0637, 'grad_norm': 0.0, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [00:52<33:24,  3.94s/it]
  2%|▏         | 12/520 [00:56<33:08,  3.91s/it]
                                                
{'loss': 1.8848, 'grad_norm': 0.0, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [00:56<33:08,  3.91s/it][2025-10-16 18:29:10,844] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:00<34:24,  4.07s/it]
                                                
{'loss': 2.0728, 'grad_norm': 0.0, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:00<34:24,  4.07s/it]
  3%|▎         | 14/520 [01:04<33:17,  3.95s/it]
                                                
{'loss': 2.1118, 'grad_norm': 0.0, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:04<33:17,  3.95s/it]
  3%|▎         | 15/520 [01:08<32:26,  3.85s/it]
                                                
{'loss': 1.7478, 'grad_norm': 0.0, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:08<32:26,  3.85s/it]
  3%|▎         | 16/520 [01:11<31:48,  3.79s/it]
                                                
{'loss': 1.8954, 'grad_norm': 0.0, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:11<31:48,  3.79s/it]
  3%|▎         | 17/520 [01:15<31:15,  3.73s/it]
                                                
{'loss': 2.1158, 'grad_norm': 0.0, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:15<31:15,  3.73s/it]
  3%|▎         | 18/520 [01:18<30:52,  3.69s/it]
                                                
{'loss': 2.1718, 'grad_norm': 0.0, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:18<30:52,  3.69s/it]
  4%|▎         | 19/520 [01:22<30:36,  3.67s/it]
                                                
{'loss': 1.8467, 'grad_norm': 0.0, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:22<30:36,  3.67s/it]
  4%|▍         | 20/520 [01:26<30:31,  3.66s/it]
                                                
{'loss': 2.2091, 'grad_norm': 0.0, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:26<30:31,  3.66s/it]
  4%|▍         | 21/520 [01:29<30:25,  3.66s/it]
                                                
{'loss': 2.0718, 'grad_norm': 0.0, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:29<30:25,  3.66s/it]
  4%|▍         | 22/520 [01:33<30:14,  3.64s/it]
                                                
{'loss': 2.0488, 'grad_norm': 0.0, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:33<30:14,  3.64s/it]
  4%|▍         | 23/520 [01:37<30:14,  3.65s/it]
                                                
{'loss': 2.0811, 'grad_norm': 0.0, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:37<30:14,  3.65s/it]
  5%|▍         | 24/520 [01:40<30:07,  3.64s/it]
                                                
{'loss': 1.8639, 'grad_norm': 0.0, 'learning_rate': 0.19987569212189224, 'epoch': 0.05}
+
  5%|▍         | 24/520 [01:40<30:07,  3.64s/it]
  5%|▍         | 25/520 [01:44<30:01,  3.64s/it]
                                                
{'loss': 2.2763, 'grad_norm': 0.0, 'learning_rate': 0.19984268150178167, 'epoch': 0.05}
+
  5%|▍         | 25/520 [01:44<30:01,  3.64s/it]
  5%|▌         | 26/520 [01:48<30:02,  3.65s/it]
                                                
{'loss': 1.9761, 'grad_norm': 0.0, 'learning_rate': 0.1998057915804532, 'epoch': 0.05}
+
  5%|▌         | 26/520 [01:48<30:02,  3.65s/it]
  5%|▌         | 27/520 [01:51<29:54,  3.64s/it]
                                                
{'loss': 2.055, 'grad_norm': 0.0, 'learning_rate': 0.1997650237912329, 'epoch': 0.05}
+
  5%|▌         | 27/520 [01:51<29:54,  3.64s/it]
  5%|▌         | 28/520 [01:55<29:54,  3.65s/it]
                                                
{'loss': 2.2197, 'grad_norm': 0.0, 'learning_rate': 0.199720379718118, 'epoch': 0.05}
+
  5%|▌         | 28/520 [01:55<29:54,  3.65s/it]
  6%|▌         | 29/520 [01:59<29:50,  3.65s/it]
                                                
{'loss': 2.145, 'grad_norm': 0.0, 'learning_rate': 0.19967186109571552, 'epoch': 0.06}
+
  6%|▌         | 29/520 [01:59<29:50,  3.65s/it]
  6%|▌         | 30/520 [02:02<30:13,  3.70s/it]
                                                
{'loss': 1.9357, 'grad_norm': 0.0, 'learning_rate': 0.19961946980917455, 'epoch': 0.06}
+
  6%|▌         | 30/520 [02:02<30:13,  3.70s/it]
  6%|▌         | 31/520 [02:06<30:35,  3.75s/it]
                                                
{'loss': 1.9528, 'grad_norm': 0.0, 'learning_rate': 0.1995632078941134, 'epoch': 0.06}
+
  6%|▌         | 31/520 [02:06<30:35,  3.75s/it]
  6%|▌         | 32/520 [02:10<30:36,  3.76s/it]
                                                
{'loss': 1.6773, 'grad_norm': 0.0, 'learning_rate': 0.19950307753654017, 'epoch': 0.06}
+
  6%|▌         | 32/520 [02:10<30:36,  3.76s/it]
  6%|▋         | 33/520 [02:14<30:35,  3.77s/it]
                                                
{'loss': 2.148, 'grad_norm': 0.0, 'learning_rate': 0.19943908107276798, 'epoch': 0.06}
+
  6%|▋         | 33/520 [02:14<30:35,  3.77s/it]
  7%|▋         | 34/520 [02:17<30:10,  3.73s/it]
                                                
{'loss': 2.2337, 'grad_norm': 0.0, 'learning_rate': 0.19937122098932428, 'epoch': 0.07}
+
  7%|▋         | 34/520 [02:17<30:10,  3.73s/it]
  7%|▋         | 35/520 [02:21<29:47,  3.69s/it]
                                                
{'loss': 2.1437, 'grad_norm': 0.0, 'learning_rate': 0.19929949992285395, 'epoch': 0.07}
+
  7%|▋         | 35/520 [02:21<29:47,  3.69s/it]
  7%|▋         | 36/520 [02:25<29:30,  3.66s/it]
                                                
{'loss': 2.0502, 'grad_norm': 0.0, 'learning_rate': 0.19922392066001723, 'epoch': 0.07}
+
  7%|▋         | 36/520 [02:25<29:30,  3.66s/it]
  7%|▋         | 37/520 [02:28<29:17,  3.64s/it]
                                                
{'loss': 1.911, 'grad_norm': 0.0, 'learning_rate': 0.19914448613738106, 'epoch': 0.07}
+
  7%|▋         | 37/520 [02:28<29:17,  3.64s/it]
  7%|▋         | 38/520 [02:32<29:11,  3.63s/it]
                                                
{'loss': 2.0439, 'grad_norm': 0.0, 'learning_rate': 0.1990611994413053, 'epoch': 0.07}
+
  7%|▋         | 38/520 [02:32<29:11,  3.63s/it]
  8%|▊         | 39/520 [02:35<29:03,  3.62s/it]
                                                
{'loss': 2.2716, 'grad_norm': 0.0, 'learning_rate': 0.19897406380782262, 'epoch': 0.07}
+
  8%|▊         | 39/520 [02:35<29:03,  3.62s/it]
  8%|▊         | 40/520 [02:39<28:57,  3.62s/it]
                                                
{'loss': 2.0359, 'grad_norm': 0.0, 'learning_rate': 0.19888308262251286, 'epoch': 0.08}
+
  8%|▊         | 40/520 [02:39<28:57,  3.62s/it]
  8%|▊         | 41/520 [02:43<28:52,  3.62s/it]
                                                
{'loss': 2.1421, 'grad_norm': 0.0, 'learning_rate': 0.19878825942037148, 'epoch': 0.08}
+
  8%|▊         | 41/520 [02:43<28:52,  3.62s/it]
  8%|▊         | 42/520 [02:46<28:50,  3.62s/it]
                                                
{'loss': 2.3152, 'grad_norm': 0.0, 'learning_rate': 0.19868959788567211, 'epoch': 0.08}
+
  8%|▊         | 42/520 [02:46<28:50,  3.62s/it]
  8%|▊         | 43/520 [02:50<28:48,  3.62s/it]
                                                
{'loss': 1.9193, 'grad_norm': 0.0, 'learning_rate': 0.1985871018518236, 'epoch': 0.08}
+
  8%|▊         | 43/520 [02:50<28:48,  3.62s/it]
  8%|▊         | 44/520 [02:54<28:42,  3.62s/it]
                                                
{'loss': 1.9613, 'grad_norm': 0.0, 'learning_rate': 0.19848077530122082, 'epoch': 0.08}
+
  8%|▊         | 44/520 [02:54<28:42,  3.62s/it]
  9%|▊         | 45/520 [02:57<28:48,  3.64s/it]
                                                
{'loss': 2.2185, 'grad_norm': 0.0, 'learning_rate': 0.19837062236509015, 'epoch': 0.09}
+
  9%|▊         | 45/520 [02:57<28:48,  3.64s/it]
  9%|▉         | 46/520 [03:01<28:44,  3.64s/it]
                                                
{'loss': 1.9333, 'grad_norm': 0.0, 'learning_rate': 0.19825664732332884, 'epoch': 0.09}
+
  9%|▉         | 46/520 [03:01<28:44,  3.64s/it]
  9%|▉         | 47/520 [03:04<28:40,  3.64s/it]
                                                
{'loss': 2.1175, 'grad_norm': 0.0, 'learning_rate': 0.19813885460433878, 'epoch': 0.09}
+
  9%|▉         | 47/520 [03:04<28:40,  3.64s/it]
  9%|▉         | 48/520 [03:08<28:33,  3.63s/it]
                                                
{'loss': 2.2907, 'grad_norm': 0.0, 'learning_rate': 0.19801724878485438, 'epoch': 0.09}
+
  9%|▉         | 48/520 [03:08<28:33,  3.63s/it]
  9%|▉         | 49/520 [03:12<28:32,  3.64s/it]
                                                
{'loss': 2.1249, 'grad_norm': 0.0, 'learning_rate': 0.19789183458976486, 'epoch': 0.09}
+
  9%|▉         | 49/520 [03:12<28:32,  3.64s/it]
 10%|▉         | 50/520 [03:15<28:23,  3.62s/it]
                                                
{'loss': 2.1602, 'grad_norm': 0.0, 'learning_rate': 0.19776261689193048, 'epoch': 0.1}
+
 10%|▉         | 50/520 [03:15<28:23,  3.62s/it]
 10%|▉         | 51/520 [03:19<28:23,  3.63s/it]
                                                
{'loss': 2.2575, 'grad_norm': 0.0, 'learning_rate': 0.19762960071199334, 'epoch': 0.1}
+
 10%|▉         | 51/520 [03:19<28:23,  3.63s/it]
 10%|█         | 52/520 [03:23<28:15,  3.62s/it]
                                                
{'loss': 2.2524, 'grad_norm': 0.0, 'learning_rate': 0.19749279121818236, 'epoch': 0.1}
+
 10%|█         | 52/520 [03:23<28:15,  3.62s/it]
 10%|█         | 53/520 [03:26<28:12,  3.62s/it]
                                                
{'loss': 2.0526, 'grad_norm': 0.0, 'learning_rate': 0.19735219372611235, 'epoch': 0.1}
+
 10%|█         | 53/520 [03:26<28:12,  3.62s/it]
 10%|█         | 54/520 [03:30<28:05,  3.62s/it]
                                                
{'loss': 2.1465, 'grad_norm': 0.0, 'learning_rate': 0.19720781369857746, 'epoch': 0.1}
+
 10%|█         | 54/520 [03:30<28:05,  3.62s/it]
 11%|█         | 55/520 [03:33<28:02,  3.62s/it]
                                                
{'loss': 2.2296, 'grad_norm': 0.0, 'learning_rate': 0.1970596567453391, 'epoch': 0.11}
+
 11%|█         | 55/520 [03:33<28:02,  3.62s/it]
 11%|█         | 56/520 [03:37<28:00,  3.62s/it]
                                                
{'loss': 2.1652, 'grad_norm': 0.0, 'learning_rate': 0.1969077286229078, 'epoch': 0.11}
+
 11%|█         | 56/520 [03:37<28:00,  3.62s/it]
 11%|█         | 57/520 [03:41<27:55,  3.62s/it]
                                                
{'loss': 2.1812, 'grad_norm': 0.0, 'learning_rate': 0.19675203523431964, 'epoch': 0.11}
+
 11%|█         | 57/520 [03:41<27:55,  3.62s/it]
 11%|█         | 58/520 [03:44<27:49,  3.61s/it]
                                                
{'loss': 1.9834, 'grad_norm': 0.0, 'learning_rate': 0.19659258262890683, 'epoch': 0.11}
+
 11%|█         | 58/520 [03:44<27:49,  3.61s/it]
 11%|█▏        | 59/520 [03:48<27:58,  3.64s/it]
                                                
{'loss': 1.8767, 'grad_norm': 0.0, 'learning_rate': 0.19642937700206278, 'epoch': 0.11}
+
 11%|█▏        | 59/520 [03:48<27:58,  3.64s/it]
 12%|█▏        | 60/520 [03:52<28:19,  3.70s/it]
                                                
{'loss': 2.072, 'grad_norm': 0.0, 'learning_rate': 0.19626242469500121, 'epoch': 0.12}
+
 12%|█▏        | 60/520 [03:52<28:19,  3.70s/it]
 12%|█▏        | 61/520 [03:56<28:35,  3.74s/it]
                                                
{'loss': 1.939, 'grad_norm': 0.0, 'learning_rate': 0.19609173219450998, 'epoch': 0.12}
+
 12%|█▏        | 61/520 [03:56<28:35,  3.74s/it]
 12%|█▏        | 62/520 [03:59<28:40,  3.76s/it]
                                                
{'loss': 2.1336, 'grad_norm': 0.0, 'learning_rate': 0.19591730613269878, 'epoch': 0.12}
+
 12%|█▏        | 62/520 [03:59<28:40,  3.76s/it]
 12%|█▏        | 63/520 [04:03<28:36,  3.76s/it]
                                                
{'loss': 2.0545, 'grad_norm': 0.0, 'learning_rate': 0.19573915328674182, 'epoch': 0.12}
+
 12%|█▏        | 63/520 [04:03<28:36,  3.76s/it]
 12%|█▏        | 64/520 [04:07<28:40,  3.77s/it]
                                                
{'loss': 2.096, 'grad_norm': 0.0, 'learning_rate': 0.1955572805786141, 'epoch': 0.12}
+
 12%|█▏        | 64/520 [04:07<28:40,  3.77s/it]
 12%|█▎        | 65/520 [04:11<28:35,  3.77s/it]
                                                
{'loss': 2.2915, 'grad_norm': 0.0, 'learning_rate': 0.1953716950748227, 'epoch': 0.12}
+
 12%|█▎        | 65/520 [04:11<28:35,  3.77s/it]
 13%|█▎        | 66/520 [04:15<28:32,  3.77s/it]
                                                
{'loss': 2.0169, 'grad_norm': 0.0, 'learning_rate': 0.19518240398613226, 'epoch': 0.13}
+
 13%|█▎        | 66/520 [04:15<28:32,  3.77s/it]
 13%|█▎        | 67/520 [04:18<28:26,  3.77s/it]
                                                
{'loss': 2.15, 'grad_norm': 0.0, 'learning_rate': 0.1949894146672846, 'epoch': 0.13}
+
 13%|█▎        | 67/520 [04:18<28:26,  3.77s/it]
 13%|█▎        | 68/520 [04:22<28:23,  3.77s/it]
                                                
{'loss': 2.0439, 'grad_norm': 0.0, 'learning_rate': 0.1947927346167132, 'epoch': 0.13}
+
 13%|█▎        | 68/520 [04:22<28:23,  3.77s/it]
 13%|█▎        | 69/520 [04:26<28:19,  3.77s/it]
                                                
{'loss': 2.2354, 'grad_norm': 0.0, 'learning_rate': 0.1945923714762516, 'epoch': 0.13}
+
 13%|█▎        | 69/520 [04:26<28:19,  3.77s/it]
 13%|█▎        | 70/520 [04:30<28:11,  3.76s/it]
                                                
{'loss': 2.0978, 'grad_norm': 0.0, 'learning_rate': 0.19438833303083677, 'epoch': 0.13}
+
 13%|█▎        | 70/520 [04:30<28:11,  3.76s/it]
 14%|█▎        | 71/520 [04:33<28:06,  3.76s/it]
                                                
{'loss': 2.093, 'grad_norm': 0.0, 'learning_rate': 0.19418062720820636, 'epoch': 0.14}
+
 14%|█▎        | 71/520 [04:33<28:06,  3.76s/it]
 14%|█▍        | 72/520 [04:37<28:07,  3.77s/it]
                                                
{'loss': 2.188, 'grad_norm': 0.0, 'learning_rate': 0.19396926207859086, 'epoch': 0.14}
+
 14%|█▍        | 72/520 [04:37<28:07,  3.77s/it]
 14%|█▍        | 73/520 [04:41<28:01,  3.76s/it]
                                                
{'loss': 2.076, 'grad_norm': 0.0, 'learning_rate': 0.19375424585439993, 'epoch': 0.14}
+
 14%|█▍        | 73/520 [04:41<28:01,  3.76s/it]
 14%|█▍        | 74/520 [04:45<27:56,  3.76s/it]
                                                
{'loss': 2.1694, 'grad_norm': 0.0, 'learning_rate': 0.1935355868899034, 'epoch': 0.14}
+
 14%|█▍        | 74/520 [04:45<27:56,  3.76s/it]
 14%|█▍        | 75/520 [04:48<27:52,  3.76s/it]
                                                
{'loss': 2.015, 'grad_norm': 0.0, 'learning_rate': 0.19331329368090666, 'epoch': 0.14}
+
 14%|█▍        | 75/520 [04:48<27:52,  3.76s/it]
 15%|█▍        | 76/520 [04:52<27:52,  3.77s/it]
                                                
{'loss': 1.7991, 'grad_norm': 0.0, 'learning_rate': 0.19308737486442043, 'epoch': 0.15}
+
 15%|█▍        | 76/520 [04:52<27:52,  3.77s/it]
 15%|█▍        | 77/520 [04:56<27:47,  3.76s/it]
                                                
{'loss': 2.2933, 'grad_norm': 0.0, 'learning_rate': 0.19285783921832536, 'epoch': 0.15}
+
 15%|█▍        | 77/520 [04:56<27:47,  3.76s/it]
 15%|█▌        | 78/520 [05:00<27:41,  3.76s/it]
                                                
{'loss': 2.065, 'grad_norm': 0.0, 'learning_rate': 0.19262469566103088, 'epoch': 0.15}
+
 15%|█▌        | 78/520 [05:00<27:41,  3.76s/it]
 15%|█▌        | 79/520 [05:03<27:36,  3.76s/it]
                                                
{'loss': 1.9893, 'grad_norm': 0.0, 'learning_rate': 0.19238795325112867, 'epoch': 0.15}
+
 15%|█▌        | 79/520 [05:03<27:36,  3.76s/it]
 15%|█▌        | 80/520 [05:07<27:34,  3.76s/it]
                                                
{'loss': 1.9708, 'grad_norm': 0.0, 'learning_rate': 0.19214762118704076, 'epoch': 0.15}
+
 15%|█▌        | 80/520 [05:07<27:34,  3.76s/it]
 16%|█▌        | 81/520 [05:11<27:27,  3.75s/it]
                                                
{'loss': 2.3674, 'grad_norm': 0.0, 'learning_rate': 0.19190370880666208, 'epoch': 0.16}
+
 16%|█▌        | 81/520 [05:11<27:27,  3.75s/it]
 16%|█▌        | 82/520 [05:15<27:22,  3.75s/it]
                                                
{'loss': 2.1439, 'grad_norm': 0.0, 'learning_rate': 0.19165622558699763, 'epoch': 0.16}
+
 16%|█▌        | 82/520 [05:15<27:22,  3.75s/it]
 16%|█▌        | 83/520 [05:18<27:19,  3.75s/it]
                                                
{'loss': 2.1668, 'grad_norm': 0.0, 'learning_rate': 0.19140518114379435, 'epoch': 0.16}
+
 16%|█▌        | 83/520 [05:18<27:19,  3.75s/it]
 16%|█▌        | 84/520 [05:22<27:04,  3.73s/it]
                                                
{'loss': 2.163, 'grad_norm': 0.0, 'learning_rate': 0.19115058523116735, 'epoch': 0.16}
+
 16%|█▌        | 84/520 [05:22<27:04,  3.73s/it]
 16%|█▋        | 85/520 [05:26<26:49,  3.70s/it]
                                                
{'loss': 2.1234, 'grad_norm': 0.0, 'learning_rate': 0.1908924477412211, 'epoch': 0.16}
+
 16%|█▋        | 85/520 [05:26<26:49,  3.70s/it]
 17%|█▋        | 86/520 [05:29<26:31,  3.67s/it]
                                                
{'loss': 2.0339, 'grad_norm': 0.0, 'learning_rate': 0.19063077870366502, 'epoch': 0.17}
+
 17%|█▋        | 86/520 [05:29<26:31,  3.67s/it]
 17%|█▋        | 87/520 [05:33<26:25,  3.66s/it]
                                                
{'loss': 2.0185, 'grad_norm': 0.0, 'learning_rate': 0.1903655882854237, 'epoch': 0.17}
+
 17%|█▋        | 87/520 [05:33<26:25,  3.66s/it]
 17%|█▋        | 88/520 [05:37<26:21,  3.66s/it]
                                                
{'loss': 1.5804, 'grad_norm': 0.0, 'learning_rate': 0.19009688679024192, 'epoch': 0.17}
+
 17%|█▋        | 88/520 [05:37<26:21,  3.66s/it]
 17%|█▋        | 89/520 [05:40<26:12,  3.65s/it]
                                                
{'loss': 2.1898, 'grad_norm': 0.0, 'learning_rate': 0.18982468465828442, 'epoch': 0.17}
+
 17%|█▋        | 89/520 [05:40<26:12,  3.65s/it]
 17%|█▋        | 90/520 [05:44<26:06,  3.64s/it]
                                                
{'loss': 2.1876, 'grad_norm': 0.0, 'learning_rate': 0.1895489924657301, 'epoch': 0.17}
+
 17%|█▋        | 90/520 [05:44<26:06,  3.64s/it]
 18%|█▊        | 91/520 [05:48<26:03,  3.65s/it]
                                                
{'loss': 1.9945, 'grad_norm': 0.0, 'learning_rate': 0.18926982092436118, 'epoch': 0.17}
+
 18%|█▊        | 91/520 [05:48<26:03,  3.65s/it]
 18%|█▊        | 92/520 [05:51<25:57,  3.64s/it]
                                                
{'loss': 2.0336, 'grad_norm': 0.0, 'learning_rate': 0.18898718088114688, 'epoch': 0.18}
+
 18%|█▊        | 92/520 [05:51<25:57,  3.64s/it]
 18%|█▊        | 93/520 [05:55<25:58,  3.65s/it]
                                                
{'loss': 2.2198, 'grad_norm': 0.0, 'learning_rate': 0.18870108331782218, 'epoch': 0.18}
+
 18%|█▊        | 93/520 [05:55<25:58,  3.65s/it]
 18%|█▊        | 94/520 [05:58<25:55,  3.65s/it]
                                                
{'loss': 2.091, 'grad_norm': 0.0, 'learning_rate': 0.18841153935046098, 'epoch': 0.18}
+
 18%|█▊        | 94/520 [05:58<25:55,  3.65s/it]
 18%|█▊        | 95/520 [06:02<25:52,  3.65s/it]
                                                
{'loss': 2.276, 'grad_norm': 0.0, 'learning_rate': 0.18811856022904425, 'epoch': 0.18}
+
 18%|█▊        | 95/520 [06:02<25:52,  3.65s/it]
 18%|█▊        | 96/520 [06:06<25:43,  3.64s/it]
                                                
{'loss': 2.0205, 'grad_norm': 0.0, 'learning_rate': 0.18782215733702287, 'epoch': 0.18}
+
 18%|█▊        | 96/520 [06:06<25:43,  3.64s/it]
 19%|█▊        | 97/520 [06:09<25:37,  3.63s/it]
                                                
{'loss': 2.3655, 'grad_norm': 0.0, 'learning_rate': 0.18752234219087538, 'epoch': 0.19}
+
 19%|█▊        | 97/520 [06:09<25:37,  3.63s/it]
 19%|█▉        | 98/520 [06:13<25:37,  3.64s/it]
                                                
{'loss': 1.9072, 'grad_norm': 0.0, 'learning_rate': 0.18721912643966054, 'epoch': 0.19}
+
 19%|█▉        | 98/520 [06:13<25:37,  3.64s/it]
 19%|█▉        | 99/520 [06:17<25:35,  3.65s/it]
                                                
{'loss': 2.1488, 'grad_norm': 0.0, 'learning_rate': 0.18691252186456464, 'epoch': 0.19}
+
 19%|█▉        | 99/520 [06:17<25:35,  3.65s/it]
 19%|█▉        | 100/520 [06:20<25:31,  3.65s/it]
                                                 
{'loss': 1.9139, 'grad_norm': 0.0, 'learning_rate': 0.1866025403784439, 'epoch': 0.19}
+
 19%|█▉        | 100/520 [06:20<25:31,  3.65s/it]
 19%|█▉        | 101/520 [06:24<25:25,  3.64s/it]
                                                 
{'loss': 2.1148, 'grad_norm': 0.0, 'learning_rate': 0.18628919402536132, 'epoch': 0.19}
+
 19%|█▉        | 101/520 [06:24<25:25,  3.64s/it]
 20%|█▉        | 102/520 [06:28<25:22,  3.64s/it]
                                                 
{'loss': 2.3567, 'grad_norm': 0.0, 'learning_rate': 0.18597249498011903, 'epoch': 0.2}
+
 20%|█▉        | 102/520 [06:28<25:22,  3.64s/it]
 20%|█▉        | 103/520 [06:31<25:14,  3.63s/it]
                                                 
{'loss': 1.9617, 'grad_norm': 0.0, 'learning_rate': 0.18565245554778517, 'epoch': 0.2}
+
 20%|█▉        | 103/520 [06:31<25:14,  3.63s/it]
 20%|██        | 104/520 [06:35<25:15,  3.64s/it]
                                                 
{'loss': 2.2072, 'grad_norm': 0.0, 'learning_rate': 0.18532908816321558, 'epoch': 0.2}
+
 20%|██        | 104/520 [06:35<25:15,  3.64s/it]
 20%|██        | 105/520 [06:38<25:06,  3.63s/it]
                                                 
{'loss': 2.068, 'grad_norm': 0.0, 'learning_rate': 0.18500240539057092, 'epoch': 0.2}
+
 20%|██        | 105/520 [06:38<25:06,  3.63s/it]
 20%|██        | 106/520 [06:42<25:02,  3.63s/it]
                                                 
{'loss': 1.7787, 'grad_norm': 0.0, 'learning_rate': 0.18467241992282843, 'epoch': 0.2}
+
 20%|██        | 106/520 [06:42<25:02,  3.63s/it]
 21%|██        | 107/520 [06:46<24:56,  3.62s/it]
                                                 
{'loss': 1.8298, 'grad_norm': 0.0, 'learning_rate': 0.18433914458128858, 'epoch': 0.21}
+
 21%|██        | 107/520 [06:46<24:56,  3.62s/it]
 21%|██        | 108/520 [06:49<24:57,  3.63s/it]
                                                 
{'loss': 2.1999, 'grad_norm': 0.0, 'learning_rate': 0.18400259231507718, 'epoch': 0.21}
+
 21%|██        | 108/520 [06:49<24:57,  3.63s/it]
 21%|██        | 109/520 [06:53<24:54,  3.64s/it]
                                                 
{'loss': 1.9115, 'grad_norm': 0.0, 'learning_rate': 0.18366277620064198, 'epoch': 0.21}
+
 21%|██        | 109/520 [06:53<24:54,  3.64s/it]
 21%|██        | 110/520 [06:57<24:47,  3.63s/it]
                                                 
{'loss': 2.1085, 'grad_norm': 0.0, 'learning_rate': 0.1833197094412449, 'epoch': 0.21}
+
 21%|██        | 110/520 [06:57<24:47,  3.63s/it]
 21%|██▏       | 111/520 [07:00<24:45,  3.63s/it]
                                                 
{'loss': 2.2082, 'grad_norm': 0.0, 'learning_rate': 0.18297340536644877, 'epoch': 0.21}
+
 21%|██▏       | 111/520 [07:00<24:45,  3.63s/it]
 22%|██▏       | 112/520 [07:04<24:40,  3.63s/it]
                                                 
{'loss': 1.9692, 'grad_norm': 0.0, 'learning_rate': 0.1826238774315995, 'epoch': 0.22}
+
 22%|██▏       | 112/520 [07:04<24:40,  3.63s/it]
 22%|██▏       | 113/520 [07:08<24:41,  3.64s/it]
                                                 
{'loss': 2.0226, 'grad_norm': 0.0, 'learning_rate': 0.18227113921730334, 'epoch': 0.22}
+
 22%|██▏       | 113/520 [07:08<24:41,  3.64s/it]
 22%|██▏       | 114/520 [07:11<24:34,  3.63s/it]
                                                 
{'loss': 2.1343, 'grad_norm': 0.0, 'learning_rate': 0.1819152044288992, 'epoch': 0.22}
+
 22%|██▏       | 114/520 [07:11<24:34,  3.63s/it]
 22%|██▏       | 115/520 [07:15<24:37,  3.65s/it]
                                                 
{'loss': 2.085, 'grad_norm': 0.0, 'learning_rate': 0.18155608689592603, 'epoch': 0.22}
+
 22%|██▏       | 115/520 [07:15<24:37,  3.65s/it]
 22%|██▏       | 116/520 [07:19<24:32,  3.65s/it]
                                                 
{'loss': 2.0739, 'grad_norm': 0.0, 'learning_rate': 0.18119380057158568, 'epoch': 0.22}
+
 22%|██▏       | 116/520 [07:19<24:32,  3.65s/it]
 22%|██▎       | 117/520 [07:22<24:31,  3.65s/it]
                                                 
{'loss': 2.1182, 'grad_norm': 0.0, 'learning_rate': 0.18082835953220056, 'epoch': 0.23}
+
 22%|██▎       | 117/520 [07:22<24:31,  3.65s/it]
 23%|██▎       | 118/520 [07:26<24:53,  3.71s/it]
                                                 
{'loss': 2.0715, 'grad_norm': 0.0, 'learning_rate': 0.18045977797666685, 'epoch': 0.23}
+
 23%|██▎       | 118/520 [07:26<24:53,  3.71s/it]
 23%|██▎       | 119/520 [07:30<25:09,  3.77s/it]
                                                 
{'loss': 2.0071, 'grad_norm': 0.0, 'learning_rate': 0.1800880702259028, 'epoch': 0.23}
+
 23%|██▎       | 119/520 [07:30<25:09,  3.77s/it]
 23%|██▎       | 120/520 [07:34<25:18,  3.80s/it]
                                                 
{'loss': 2.2848, 'grad_norm': 0.0, 'learning_rate': 0.17971325072229227, 'epoch': 0.23}
+
 23%|██▎       | 120/520 [07:34<25:18,  3.80s/it]
 23%|██▎       | 121/520 [07:37<24:55,  3.75s/it]
                                                 
{'loss': 2.0958, 'grad_norm': 0.0, 'learning_rate': 0.17933533402912352, 'epoch': 0.23}
+
 23%|██▎       | 121/520 [07:37<24:55,  3.75s/it]
 23%|██▎       | 122/520 [07:41<25:06,  3.79s/it]
                                                 
{'loss': 2.0953, 'grad_norm': 0.0, 'learning_rate': 0.17895433483002354, 'epoch': 0.23}
+
 23%|██▎       | 122/520 [07:41<25:06,  3.79s/it]
 24%|██▎       | 123/520 [07:45<25:14,  3.82s/it]
                                                 
{'loss': 1.7717, 'grad_norm': 0.0, 'learning_rate': 0.17857026792838737, 'epoch': 0.24}
+
 24%|██▎       | 123/520 [07:45<25:14,  3.82s/it]
 24%|██▍       | 124/520 [07:49<25:15,  3.83s/it]
                                                 
{'loss': 2.236, 'grad_norm': 0.0, 'learning_rate': 0.178183148246803, 'epoch': 0.24}
+
 24%|██▍       | 124/520 [07:49<25:15,  3.83s/it]
 24%|██▍       | 125/520 [07:53<25:15,  3.84s/it]
                                                 
{'loss': 2.1224, 'grad_norm': 0.0, 'learning_rate': 0.1777929908264715, 'epoch': 0.24}
+
 24%|██▍       | 125/520 [07:53<25:15,  3.84s/it]
 24%|██▍       | 126/520 [07:57<26:32,  4.04s/it]
                                                 
{'loss': 1.7811, 'grad_norm': 0.0, 'learning_rate': 0.17739981082662276, 'epoch': 0.24}
+
 24%|██▍       | 126/520 [07:57<26:32,  4.04s/it]
 24%|██▍       | 127/520 [08:01<26:08,  3.99s/it]
                                                 
{'loss': 2.2468, 'grad_norm': 0.0, 'learning_rate': 0.1770036235239263, 'epoch': 0.24}
+
 24%|██▍       | 127/520 [08:01<26:08,  3.99s/it]
 25%|██▍       | 128/520 [08:05<25:49,  3.95s/it]
                                                 
{'loss': 2.138, 'grad_norm': 0.0, 'learning_rate': 0.1766044443118978, 'epoch': 0.25}
+
 25%|██▍       | 128/520 [08:05<25:49,  3.95s/it]
 25%|██▍       | 129/520 [08:09<25:19,  3.89s/it]
                                                 
{'loss': 2.012, 'grad_norm': 0.0, 'learning_rate': 0.17620228870030108, 'epoch': 0.25}
+
 25%|██▍       | 129/520 [08:09<25:19,  3.89s/it]
 25%|██▌       | 130/520 [08:13<24:45,  3.81s/it]
                                                 
{'loss': 2.0714, 'grad_norm': 0.0, 'learning_rate': 0.1757971723145453, 'epoch': 0.25}
+
 25%|██▌       | 130/520 [08:13<24:45,  3.81s/it]
 25%|██▌       | 131/520 [08:16<24:20,  3.75s/it]
                                                 
{'loss': 1.7064, 'grad_norm': 0.0, 'learning_rate': 0.175389110895078, 'epoch': 0.25}
+
 25%|██▌       | 131/520 [08:16<24:20,  3.75s/it]
 25%|██▌       | 132/520 [08:20<24:01,  3.72s/it]
                                                 
{'loss': 2.2907, 'grad_norm': 0.0, 'learning_rate': 0.17497812029677343, 'epoch': 0.25}
+
 25%|██▌       | 132/520 [08:20<24:01,  3.72s/it]
 26%|██▌       | 133/520 [08:23<23:46,  3.69s/it]
                                                 
{'loss': 2.2873, 'grad_norm': 0.0, 'learning_rate': 0.17456421648831655, 'epoch': 0.26}
+
 26%|██▌       | 133/520 [08:23<23:46,  3.69s/it]
 26%|██▌       | 134/520 [08:27<23:37,  3.67s/it]
                                                 
{'loss': 2.1675, 'grad_norm': 0.0, 'learning_rate': 0.17414741555158267, 'epoch': 0.26}
+
 26%|██▌       | 134/520 [08:27<23:37,  3.67s/it]
 26%|██▌       | 135/520 [08:31<23:27,  3.66s/it]
                                                 
{'loss': 2.184, 'grad_norm': 0.0, 'learning_rate': 0.1737277336810124, 'epoch': 0.26}
+
 26%|██▌       | 135/520 [08:31<23:27,  3.66s/it]
 26%|██▌       | 136/520 [08:34<23:16,  3.64s/it]
                                                 
{'loss': 2.109, 'grad_norm': 0.0, 'learning_rate': 0.17330518718298263, 'epoch': 0.26}
+
 26%|██▌       | 136/520 [08:34<23:16,  3.64s/it]
 26%|██▋       | 137/520 [08:38<23:10,  3.63s/it]
                                                 
{'loss': 2.2285, 'grad_norm': 0.0, 'learning_rate': 0.17287979247517285, 'epoch': 0.26}
+
 26%|██▋       | 137/520 [08:38<23:10,  3.63s/it]
 27%|██▋       | 138/520 [08:41<23:05,  3.63s/it]
                                                 
{'loss': 2.0231, 'grad_norm': 0.0, 'learning_rate': 0.17245156608592727, 'epoch': 0.27}
+
 27%|██▋       | 138/520 [08:41<23:05,  3.63s/it]
 27%|██▋       | 139/520 [08:45<23:25,  3.69s/it]
                                                 
{'loss': 1.779, 'grad_norm': 0.0, 'learning_rate': 0.17202052465361267, 'epoch': 0.27}
+
 27%|██▋       | 139/520 [08:45<23:25,  3.69s/it]
 27%|██▋       | 140/520 [08:49<23:25,  3.70s/it]
                                                 
{'loss': 1.8436, 'grad_norm': 0.0, 'learning_rate': 0.17158668492597184, 'epoch': 0.27}
+
 27%|██▋       | 140/520 [08:49<23:25,  3.70s/it]
 27%|██▋       | 141/520 [08:53<23:16,  3.68s/it]
                                                 
{'loss': 1.9943, 'grad_norm': 0.0, 'learning_rate': 0.17115006375947303, 'epoch': 0.27}
+
 27%|██▋       | 141/520 [08:53<23:16,  3.68s/it]
 27%|██▋       | 142/520 [08:56<23:09,  3.68s/it]
                                                 
{'loss': 1.7887, 'grad_norm': 0.0, 'learning_rate': 0.17071067811865476, 'epoch': 0.27}
+
 27%|██▋       | 142/520 [08:56<23:09,  3.68s/it]
 28%|██▊       | 143/520 [09:00<22:57,  3.65s/it]
                                                 
{'loss': 2.1754, 'grad_norm': 0.0, 'learning_rate': 0.17026854507546693, 'epoch': 0.28}
+
 28%|██▊       | 143/520 [09:00<22:57,  3.65s/it]
 28%|██▊       | 144/520 [09:04<22:51,  3.65s/it]
                                                 
{'loss': 2.2613, 'grad_norm': 0.0, 'learning_rate': 0.1698236818086073, 'epoch': 0.28}
+
 28%|██▊       | 144/520 [09:04<22:51,  3.65s/it]
 28%|██▊       | 145/520 [09:07<22:48,  3.65s/it]
                                                 
{'loss': 2.0696, 'grad_norm': 0.0, 'learning_rate': 0.16937610560285418, 'epoch': 0.28}
+
 28%|██▊       | 145/520 [09:07<22:48,  3.65s/it]
 28%|██▊       | 146/520 [09:11<22:43,  3.65s/it]
                                                 
{'loss': 1.935, 'grad_norm': 0.0, 'learning_rate': 0.1689258338483947, 'epoch': 0.28}
+
 28%|██▊       | 146/520 [09:11<22:43,  3.65s/it]
 28%|██▊       | 147/520 [09:14<22:38,  3.64s/it]
                                                 
{'loss': 2.0573, 'grad_norm': 0.0, 'learning_rate': 0.16847288404014937, 'epoch': 0.28}
+
 28%|██▊       | 147/520 [09:14<22:38,  3.64s/it]
 28%|██▊       | 148/520 [09:18<22:32,  3.63s/it]
                                                 
{'loss': 2.0705, 'grad_norm': 0.0, 'learning_rate': 0.16801727377709194, 'epoch': 0.28}
+
 28%|██▊       | 148/520 [09:18<22:32,  3.63s/it]
 29%|██▊       | 149/520 [09:22<22:35,  3.65s/it]
                                                 
{'loss': 2.1033, 'grad_norm': 0.0, 'learning_rate': 0.16755902076156604, 'epoch': 0.29}
+
 29%|██▊       | 149/520 [09:22<22:35,  3.65s/it]
 29%|██▉       | 150/520 [09:25<22:29,  3.65s/it]
                                                 
{'loss': 2.1162, 'grad_norm': 0.0, 'learning_rate': 0.16709814279859703, 'epoch': 0.29}
+
 29%|██▉       | 150/520 [09:25<22:29,  3.65s/it]
 29%|██▉       | 151/520 [09:29<22:23,  3.64s/it]
                                                 
{'loss': 2.1368, 'grad_norm': 0.0, 'learning_rate': 0.1666346577952004, 'epoch': 0.29}
+
 29%|██▉       | 151/520 [09:29<22:23,  3.64s/it]
 29%|██▉       | 152/520 [09:33<22:18,  3.64s/it]
                                                 
{'loss': 2.2102, 'grad_norm': 0.0, 'learning_rate': 0.16616858375968596, 'epoch': 0.29}
+
 29%|██▉       | 152/520 [09:33<22:18,  3.64s/it]
 29%|██▉       | 153/520 [09:36<22:12,  3.63s/it]
                                                 
{'loss': 2.069, 'grad_norm': 0.0, 'learning_rate': 0.16569993880095807, 'epoch': 0.29}
+
 29%|██▉       | 153/520 [09:36<22:12,  3.63s/it]
 30%|██▉       | 154/520 [09:40<22:08,  3.63s/it]
                                                 
{'loss': 2.0422, 'grad_norm': 0.0, 'learning_rate': 0.16522874112781213, 'epoch': 0.3}
+
 30%|██▉       | 154/520 [09:40<22:08,  3.63s/it]
 30%|██▉       | 155/520 [09:44<22:06,  3.63s/it]
                                                 
{'loss': 2.2418, 'grad_norm': 0.0, 'learning_rate': 0.16475500904822704, 'epoch': 0.3}
+
 30%|██▉       | 155/520 [09:44<22:06,  3.63s/it]
 30%|███       | 156/520 [09:47<22:01,  3.63s/it]
                                                 
{'loss': 2.1952, 'grad_norm': 0.0, 'learning_rate': 0.16427876096865393, 'epoch': 0.3}
+
 30%|███       | 156/520 [09:47<22:01,  3.63s/it]
 30%|███       | 157/520 [09:51<22:20,  3.69s/it]
                                                 
{'loss': 1.9614, 'grad_norm': 0.0, 'learning_rate': 0.16380001539330089, 'epoch': 0.3}
+
 30%|███       | 157/520 [09:51<22:20,  3.69s/it]
 30%|███       | 158/520 [09:55<22:13,  3.68s/it]
                                                 
{'loss': 2.0753, 'grad_norm': 0.0, 'learning_rate': 0.163318790923414, 'epoch': 0.3}
+
 30%|███       | 158/520 [09:55<22:13,  3.68s/it]
 31%|███       | 159/520 [09:58<22:01,  3.66s/it]
                                                 
{'loss': 2.1432, 'grad_norm': 0.0, 'learning_rate': 0.16283510625655473, 'epoch': 0.31}
+
 31%|███       | 159/520 [09:58<22:01,  3.66s/it]
 31%|███       | 160/520 [10:02<21:52,  3.64s/it]
                                                 
{'loss': 2.1456, 'grad_norm': 0.0, 'learning_rate': 0.16234898018587338, 'epoch': 0.31}
+
 31%|███       | 160/520 [10:02<21:52,  3.64s/it]
 31%|███       | 161/520 [10:06<21:47,  3.64s/it]
                                                 
{'loss': 2.152, 'grad_norm': 0.0, 'learning_rate': 0.16186043159937882, 'epoch': 0.31}
+
 31%|███       | 161/520 [10:06<21:47,  3.64s/it]
 31%|███       | 162/520 [10:09<21:42,  3.64s/it]
                                                 
{'loss': 1.8926, 'grad_norm': 0.0, 'learning_rate': 0.16136947947920477, 'epoch': 0.31}
+
 31%|███       | 162/520 [10:09<21:42,  3.64s/it]
 31%|███▏      | 163/520 [10:13<21:44,  3.65s/it]
                                                 
{'loss': 2.2871, 'grad_norm': 0.0, 'learning_rate': 0.16087614290087207, 'epoch': 0.31}
+
 31%|███▏      | 163/520 [10:13<21:44,  3.65s/it]
 32%|███▏      | 164/520 [10:17<22:00,  3.71s/it]
                                                 
{'loss': 2.1267, 'grad_norm': 0.0, 'learning_rate': 0.16038044103254775, 'epoch': 0.32}
+
 32%|███▏      | 164/520 [10:17<22:00,  3.71s/it]
 32%|███▏      | 165/520 [10:21<22:08,  3.74s/it]
                                                 
{'loss': 2.1168, 'grad_norm': 0.0, 'learning_rate': 0.15988239313430005, 'epoch': 0.32}
+
 32%|███▏      | 165/520 [10:21<22:08,  3.74s/it]
 32%|███▏      | 166/520 [10:24<22:15,  3.77s/it]
                                                 
{'loss': 2.1565, 'grad_norm': 0.0, 'learning_rate': 0.15938201855735015, 'epoch': 0.32}
+
 32%|███▏      | 166/520 [10:24<22:15,  3.77s/it]
 32%|███▏      | 167/520 [10:28<22:21,  3.80s/it]
                                                 
{'loss': 2.1271, 'grad_norm': 0.0, 'learning_rate': 0.15887933674332047, 'epoch': 0.32}
+
 32%|███▏      | 167/520 [10:28<22:21,  3.80s/it]
 32%|███▏      | 168/520 [10:32<22:20,  3.81s/it]
                                                 
{'loss': 1.9889, 'grad_norm': 0.0, 'learning_rate': 0.158374367223479, 'epoch': 0.32}
+
 32%|███▏      | 168/520 [10:32<22:20,  3.81s/it]
 32%|███▎      | 169/520 [10:36<22:20,  3.82s/it]
                                                 
{'loss': 2.1528, 'grad_norm': 0.0, 'learning_rate': 0.1578671296179806, 'epoch': 0.33}
+
 32%|███▎      | 169/520 [10:36<22:20,  3.82s/it]
 33%|███▎      | 170/520 [10:40<22:19,  3.83s/it]
                                                 
{'loss': 1.809, 'grad_norm': 0.0, 'learning_rate': 0.15735764363510463, 'epoch': 0.33}
+
 33%|███▎      | 170/520 [10:40<22:19,  3.83s/it]
 33%|███▎      | 171/520 [10:44<22:18,  3.83s/it]
                                                 
{'loss': 2.2505, 'grad_norm': 0.0, 'learning_rate': 0.15684592907048925, 'epoch': 0.33}
+
 33%|███▎      | 171/520 [10:44<22:18,  3.83s/it]
 33%|███▎      | 172/520 [10:47<22:14,  3.83s/it]
                                                 
{'loss': 2.0153, 'grad_norm': 0.0, 'learning_rate': 0.1563320058063622, 'epoch': 0.33}
+
 33%|███▎      | 172/520 [10:47<22:14,  3.83s/it]
 33%|███▎      | 173/520 [10:51<22:06,  3.82s/it]
                                                 
{'loss': 2.1214, 'grad_norm': 0.0, 'learning_rate': 0.15581589381076844, 'epoch': 0.33}
+
 33%|███▎      | 173/520 [10:51<22:06,  3.82s/it]
 33%|███▎      | 174/520 [10:55<21:47,  3.78s/it]
                                                 
{'loss': 2.1899, 'grad_norm': 0.0, 'learning_rate': 0.15529761313679394, 'epoch': 0.33}
+
 33%|███▎      | 174/520 [10:55<21:47,  3.78s/it]
 34%|███▎      | 175/520 [10:59<21:46,  3.79s/it]
                                                 
{'loss': 2.0477, 'grad_norm': 0.0, 'learning_rate': 0.15477718392178716, 'epoch': 0.34}
+
 34%|███▎      | 175/520 [10:59<21:46,  3.79s/it]
 34%|███▍      | 176/520 [11:02<21:35,  3.77s/it]
                                                 
{'loss': 1.9286, 'grad_norm': 0.0, 'learning_rate': 0.15425462638657594, 'epoch': 0.34}
+
 34%|███▍      | 176/520 [11:02<21:35,  3.77s/it]
 34%|███▍      | 177/520 [11:06<21:17,  3.72s/it]
                                                 
{'loss': 1.9234, 'grad_norm': 0.0, 'learning_rate': 0.1537299608346824, 'epoch': 0.34}
+
 34%|███▍      | 177/520 [11:06<21:17,  3.72s/it]
 34%|███▍      | 178/520 [11:10<21:00,  3.69s/it]
                                                 
{'loss': 2.1311, 'grad_norm': 0.0, 'learning_rate': 0.15320320765153367, 'epoch': 0.34}
+
 34%|███▍      | 178/520 [11:10<21:00,  3.69s/it]
 34%|███▍      | 179/520 [11:13<20:49,  3.66s/it]
                                                 
{'loss': 2.127, 'grad_norm': 0.0, 'learning_rate': 0.15267438730367008, 'epoch': 0.34}
+
 34%|███▍      | 179/520 [11:13<20:49,  3.66s/it]
 35%|███▍      | 180/520 [11:17<20:40,  3.65s/it]
                                                 
{'loss': 2.0928, 'grad_norm': 0.0, 'learning_rate': 0.1521435203379498, 'epoch': 0.35}
+
 35%|███▍      | 180/520 [11:17<20:40,  3.65s/it]
 35%|███▍      | 181/520 [11:20<20:31,  3.63s/it]
                                                 
{'loss': 1.924, 'grad_norm': 0.0, 'learning_rate': 0.15161062738075068, 'epoch': 0.35}
+
 35%|███▍      | 181/520 [11:20<20:31,  3.63s/it]
 35%|███▌      | 182/520 [11:24<20:27,  3.63s/it]
                                                 
{'loss': 2.173, 'grad_norm': 0.0, 'learning_rate': 0.1510757291371686, 'epoch': 0.35}
+
 35%|███▌      | 182/520 [11:24<20:27,  3.63s/it]
 35%|███▌      | 183/520 [11:28<20:24,  3.63s/it]
                                                 
{'loss': 2.0064, 'grad_norm': 0.0, 'learning_rate': 0.1505388463902131, 'epoch': 0.35}
+
 35%|███▌      | 183/520 [11:28<20:24,  3.63s/it]
 35%|███▌      | 184/520 [11:32<20:39,  3.69s/it]
                                                 
{'loss': 2.2563, 'grad_norm': 0.0, 'learning_rate': 0.15000000000000002, 'epoch': 0.35}
+
 35%|███▌      | 184/520 [11:32<20:39,  3.69s/it]
 36%|███▌      | 185/520 [11:35<20:52,  3.74s/it]
                                                 
{'loss': 2.0567, 'grad_norm': 0.0, 'learning_rate': 0.14945921090294076, 'epoch': 0.36}
+
 36%|███▌      | 185/520 [11:35<20:52,  3.74s/it]
 36%|███▌      | 186/520 [11:39<21:00,  3.77s/it]
                                                 
{'loss': 2.0963, 'grad_norm': 0.0, 'learning_rate': 0.14891650011092894, 'epoch': 0.36}
+
 36%|███▌      | 186/520 [11:39<21:00,  3.77s/it]
 36%|███▌      | 187/520 [11:43<21:05,  3.80s/it]
                                                 
{'loss': 2.2718, 'grad_norm': 0.0, 'learning_rate': 0.14837188871052398, 'epoch': 0.36}
+
 36%|███▌      | 187/520 [11:43<21:05,  3.80s/it]
 36%|███▌      | 188/520 [11:47<20:53,  3.77s/it]
                                                 
{'loss': 2.1637, 'grad_norm': 0.0, 'learning_rate': 0.14782539786213184, 'epoch': 0.36}
+
 36%|███▌      | 188/520 [11:47<20:53,  3.77s/it]
 36%|███▋      | 189/520 [11:50<20:34,  3.73s/it]
                                                 
{'loss': 2.0487, 'grad_norm': 0.0, 'learning_rate': 0.1472770487991827, 'epoch': 0.36}
+
 36%|███▋      | 189/520 [11:50<20:34,  3.73s/it]
 37%|███▋      | 190/520 [11:54<20:17,  3.69s/it]
                                                 
{'loss': 2.2378, 'grad_norm': 0.0, 'learning_rate': 0.1467268628273062, 'epoch': 0.37}
+
 37%|███▋      | 190/520 [11:54<20:17,  3.69s/it]
 37%|███▋      | 191/520 [11:58<20:09,  3.68s/it]
                                                 
{'loss': 2.2167, 'grad_norm': 0.0, 'learning_rate': 0.1461748613235034, 'epoch': 0.37}
+
 37%|███▋      | 191/520 [11:58<20:09,  3.68s/it]
 37%|███▋      | 192/520 [12:01<20:01,  3.66s/it]
                                                 
{'loss': 2.1261, 'grad_norm': 0.0, 'learning_rate': 0.1456210657353163, 'epoch': 0.37}
+
 37%|███▋      | 192/520 [12:01<20:01,  3.66s/it]
 37%|███▋      | 193/520 [12:05<19:51,  3.64s/it]
                                                 
{'loss': 1.9259, 'grad_norm': 0.0, 'learning_rate': 0.14506549757999454, 'epoch': 0.37}
+
 37%|███▋      | 193/520 [12:05<19:51,  3.64s/it]
 37%|███▋      | 194/520 [12:09<19:45,  3.64s/it]
                                                 
{'loss': 1.7848, 'grad_norm': 0.0, 'learning_rate': 0.14450817844365924, 'epoch': 0.37}
+
 37%|███▋      | 194/520 [12:09<19:45,  3.64s/it]
 38%|███▊      | 195/520 [12:12<19:39,  3.63s/it]
                                                 
{'loss': 1.9879, 'grad_norm': 0.0, 'learning_rate': 0.1439491299804645, 'epoch': 0.38}
+
 38%|███▊      | 195/520 [12:12<19:39,  3.63s/it]
 38%|███▊      | 196/520 [12:16<19:31,  3.62s/it]
                                                 
{'loss': 2.2013, 'grad_norm': 0.0, 'learning_rate': 0.14338837391175582, 'epoch': 0.38}
+
 38%|███▊      | 196/520 [12:16<19:31,  3.62s/it]
 38%|███▊      | 197/520 [12:19<19:29,  3.62s/it]
                                                 
{'loss': 2.0407, 'grad_norm': 0.0, 'learning_rate': 0.14282593202522628, 'epoch': 0.38}
+
 38%|███▊      | 197/520 [12:19<19:29,  3.62s/it]
 38%|███▊      | 198/520 [12:23<19:22,  3.61s/it]
                                                 
{'loss': 2.1881, 'grad_norm': 0.0, 'learning_rate': 0.14226182617406996, 'epoch': 0.38}
+
 38%|███▊      | 198/520 [12:23<19:22,  3.61s/it]
 38%|███▊      | 199/520 [12:27<19:22,  3.62s/it]
                                                 
{'loss': 2.1723, 'grad_norm': 0.0, 'learning_rate': 0.14169607827613281, 'epoch': 0.38}
+
 38%|███▊      | 199/520 [12:27<19:22,  3.62s/it]
 38%|███▊      | 200/520 [12:30<19:18,  3.62s/it]
                                                 
{'loss': 1.8883, 'grad_norm': 0.0, 'learning_rate': 0.14112871031306118, 'epoch': 0.38}
+
 38%|███▊      | 200/520 [12:30<19:18,  3.62s/it]
 39%|███▊      | 201/520 [12:34<19:16,  3.63s/it]
                                                 
{'loss': 1.789, 'grad_norm': 0.0, 'learning_rate': 0.1405597443294475, 'epoch': 0.39}
+
 39%|███▊      | 201/520 [12:34<19:16,  3.63s/it]
 39%|███▉      | 202/520 [12:38<19:12,  3.62s/it]
                                                 
{'loss': 2.1073, 'grad_norm': 0.0, 'learning_rate': 0.13998920243197407, 'epoch': 0.39}
+
 39%|███▉      | 202/520 [12:38<19:12,  3.62s/it]
 39%|███▉      | 203/520 [12:41<19:07,  3.62s/it]
                                                 
{'loss': 2.1082, 'grad_norm': 0.0, 'learning_rate': 0.13941710678855396, 'epoch': 0.39}
+
 39%|███▉      | 203/520 [12:41<19:07,  3.62s/it]
 39%|███▉      | 204/520 [12:45<19:03,  3.62s/it]
                                                 
{'loss': 2.124, 'grad_norm': 0.0, 'learning_rate': 0.13884347962746948, 'epoch': 0.39}
+
 39%|███▉      | 204/520 [12:45<19:03,  3.62s/it]
 39%|███▉      | 205/520 [12:48<19:01,  3.62s/it]
                                                 
{'loss': 1.7868, 'grad_norm': 0.0, 'learning_rate': 0.138268343236509, 'epoch': 0.39}
+
 39%|███▉      | 205/520 [12:48<19:01,  3.62s/it]
 40%|███▉      | 206/520 [12:52<18:54,  3.61s/it]
                                                 
{'loss': 2.0823, 'grad_norm': 0.0, 'learning_rate': 0.13769171996210053, 'epoch': 0.4}
+
 40%|███▉      | 206/520 [12:52<18:54,  3.61s/it]
 40%|███▉      | 207/520 [12:56<18:51,  3.61s/it]
                                                 
{'loss': 1.6807, 'grad_norm': 0.0, 'learning_rate': 0.1371136322084438, 'epoch': 0.4}
+
 40%|███▉      | 207/520 [12:56<18:51,  3.61s/it]
 40%|████      | 208/520 [12:59<18:46,  3.61s/it]
                                                 
{'loss': 2.3713, 'grad_norm': 0.0, 'learning_rate': 0.13653410243663952, 'epoch': 0.4}
+
 40%|████      | 208/520 [12:59<18:46,  3.61s/it]
 40%|████      | 209/520 [13:03<18:44,  3.62s/it]
                                                 
{'loss': 2.1462, 'grad_norm': 0.0, 'learning_rate': 0.13595315316381676, 'epoch': 0.4}
+
 40%|████      | 209/520 [13:03<18:44,  3.62s/it]
 40%|████      | 210/520 [13:06<18:42,  3.62s/it]
                                                 
{'loss': 2.0903, 'grad_norm': 0.0, 'learning_rate': 0.13537080696225814, 'epoch': 0.4}
+
 40%|████      | 210/520 [13:06<18:42,  3.62s/it]
 41%|████      | 211/520 [13:10<18:38,  3.62s/it]
                                                 
{'loss': 2.1076, 'grad_norm': 0.0, 'learning_rate': 0.13478708645852272, 'epoch': 0.41}
+
 41%|████      | 211/520 [13:10<18:38,  3.62s/it]
 41%|████      | 212/520 [13:14<18:32,  3.61s/it]
                                                 
{'loss': 1.9716, 'grad_norm': 0.0, 'learning_rate': 0.1342020143325669, 'epoch': 0.41}
+
 41%|████      | 212/520 [13:14<18:32,  3.61s/it]
 41%|████      | 213/520 [13:17<18:31,  3.62s/it]
                                                 
{'loss': 2.3066, 'grad_norm': 0.0, 'learning_rate': 0.13361561331686309, 'epoch': 0.41}
+
 41%|████      | 213/520 [13:17<18:31,  3.62s/it]
 41%|████      | 214/520 [13:21<18:27,  3.62s/it]
                                                 
{'loss': 2.1004, 'grad_norm': 0.0, 'learning_rate': 0.13302790619551672, 'epoch': 0.41}
+
 41%|████      | 214/520 [13:21<18:27,  3.62s/it]
 41%|████▏     | 215/520 [13:25<18:25,  3.63s/it]
                                                 
{'loss': 1.9269, 'grad_norm': 0.0, 'learning_rate': 0.1324389158033807, 'epoch': 0.41}
+
 41%|████▏     | 215/520 [13:25<18:25,  3.63s/it]
 42%|████▏     | 216/520 [13:28<18:25,  3.64s/it]
                                                 
{'loss': 2.091, 'grad_norm': 0.0, 'learning_rate': 0.13184866502516845, 'epoch': 0.42}
+
 42%|████▏     | 216/520 [13:28<18:25,  3.64s/it]
 42%|████▏     | 217/520 [13:32<18:22,  3.64s/it]
                                                 
{'loss': 2.13, 'grad_norm': 0.0, 'learning_rate': 0.13125717679456447, 'epoch': 0.42}
+
 42%|████▏     | 217/520 [13:32<18:22,  3.64s/it]
 42%|████▏     | 218/520 [13:36<18:33,  3.69s/it]
                                                 
{'loss': 2.2033, 'grad_norm': 0.0, 'learning_rate': 0.13066447409333345, 'epoch': 0.42}
+
 42%|████▏     | 218/520 [13:36<18:33,  3.69s/it]
 42%|████▏     | 219/520 [13:40<18:44,  3.74s/it]
                                                 
{'loss': 2.0636, 'grad_norm': 0.0, 'learning_rate': 0.1300705799504273, 'epoch': 0.42}
+
 42%|████▏     | 219/520 [13:40<18:44,  3.74s/it]
 42%|████▏     | 220/520 [13:43<18:49,  3.77s/it]
                                                 
{'loss': 1.8837, 'grad_norm': 0.0, 'learning_rate': 0.12947551744109043, 'epoch': 0.42}
+
 42%|████▏     | 220/520 [13:43<18:49,  3.77s/it]
 42%|████▎     | 221/520 [13:47<18:54,  3.79s/it]
                                                 
{'loss': 2.1677, 'grad_norm': 0.0, 'learning_rate': 0.128879309685963, 'epoch': 0.42}
+
 42%|████▎     | 221/520 [13:47<18:54,  3.79s/it]
 43%|████▎     | 222/520 [13:51<18:55,  3.81s/it]
                                                 
{'loss': 2.1221, 'grad_norm': 0.0, 'learning_rate': 0.12828197985018275, 'epoch': 0.43}
+
 43%|████▎     | 222/520 [13:51<18:55,  3.81s/it]
 43%|████▎     | 223/520 [13:55<18:55,  3.82s/it]
                                                 
{'loss': 2.061, 'grad_norm': 0.0, 'learning_rate': 0.12768355114248495, 'epoch': 0.43}
+
 43%|████▎     | 223/520 [13:55<18:55,  3.82s/it]
 43%|████▎     | 224/520 [13:59<18:56,  3.84s/it]
                                                 
{'loss': 1.6814, 'grad_norm': 0.0, 'learning_rate': 0.12708404681430052, 'epoch': 0.43}
+
 43%|████▎     | 224/520 [13:59<18:56,  3.84s/it]
 43%|████▎     | 225/520 [14:03<18:53,  3.84s/it]
                                                 
{'loss': 2.0363, 'grad_norm': 0.0, 'learning_rate': 0.1264834901588527, 'epoch': 0.43}
+
 43%|████▎     | 225/520 [14:03<18:53,  3.84s/it]
 43%|████▎     | 226/520 [14:06<18:50,  3.85s/it]
                                                 
{'loss': 2.1585, 'grad_norm': 0.0, 'learning_rate': 0.12588190451025208, 'epoch': 0.43}
+
 43%|████▎     | 226/520 [14:06<18:50,  3.85s/it]
 44%|████▎     | 227/520 [14:10<18:48,  3.85s/it]
                                                 
{'loss': 2.081, 'grad_norm': 0.0, 'learning_rate': 0.12527931324258976, 'epoch': 0.44}
+
 44%|████▎     | 227/520 [14:10<18:48,  3.85s/it]
 44%|████▍     | 228/520 [14:14<18:44,  3.85s/it]
                                                 
{'loss': 1.9309, 'grad_norm': 0.0, 'learning_rate': 0.12467573976902935, 'epoch': 0.44}
+
 44%|████▍     | 228/520 [14:14<18:44,  3.85s/it]
 44%|████▍     | 229/520 [14:18<18:39,  3.85s/it]
                                                 
{'loss': 1.963, 'grad_norm': 0.0, 'learning_rate': 0.12407120754089732, 'epoch': 0.44}
+
 44%|████▍     | 229/520 [14:18<18:39,  3.85s/it]
 44%|████▍     | 230/520 [14:22<18:36,  3.85s/it]
                                                 
{'loss': 2.1559, 'grad_norm': 0.0, 'learning_rate': 0.12346574004677154, 'epoch': 0.44}
+
 44%|████▍     | 230/520 [14:22<18:36,  3.85s/it]
 44%|████▍     | 231/520 [14:26<18:32,  3.85s/it]
                                                 
{'loss': 2.0662, 'grad_norm': 0.0, 'learning_rate': 0.12285936081156897, 'epoch': 0.44}
+
 44%|████▍     | 231/520 [14:26<18:32,  3.85s/it]
 45%|████▍     | 232/520 [14:30<18:35,  3.87s/it]
                                                 
{'loss': 1.9045, 'grad_norm': 0.0, 'learning_rate': 0.12225209339563144, 'epoch': 0.45}
+
 45%|████▍     | 232/520 [14:30<18:35,  3.87s/it]
 45%|████▍     | 233/520 [14:34<18:30,  3.87s/it]
                                                 
{'loss': 1.9718, 'grad_norm': 0.0, 'learning_rate': 0.12164396139381028, 'epoch': 0.45}
+
 45%|████▍     | 233/520 [14:34<18:30,  3.87s/it]
 45%|████▌     | 234/520 [14:37<18:26,  3.87s/it]
                                                 
{'loss': 2.1082, 'grad_norm': 0.0, 'learning_rate': 0.12103498843454959, 'epoch': 0.45}
+
 45%|████▌     | 234/520 [14:37<18:26,  3.87s/it]
 45%|████▌     | 235/520 [14:41<18:21,  3.86s/it]
                                                 
{'loss': 2.1515, 'grad_norm': 0.0, 'learning_rate': 0.12042519817896805, 'epoch': 0.45}
+
 45%|████▌     | 235/520 [14:41<18:21,  3.86s/it]
 45%|████▌     | 236/520 [14:45<18:16,  3.86s/it]
                                                 
{'loss': 2.0049, 'grad_norm': 0.0, 'learning_rate': 0.11981461431993977, 'epoch': 0.45}
+
 45%|████▌     | 236/520 [14:45<18:16,  3.86s/it]
 46%|████▌     | 237/520 [14:49<18:14,  3.87s/it]
                                                 
{'loss': 2.0493, 'grad_norm': 0.0, 'learning_rate': 0.11920326058117364, 'epoch': 0.46}
+
 46%|████▌     | 237/520 [14:49<18:14,  3.87s/it]
 46%|████▌     | 238/520 [14:53<18:10,  3.87s/it]
                                                 
{'loss': 2.1513, 'grad_norm': 0.0, 'learning_rate': 0.11859116071629149, 'epoch': 0.46}
+
 46%|████▌     | 238/520 [14:53<18:10,  3.87s/it]
 46%|████▌     | 239/520 [14:57<18:06,  3.87s/it]
                                                 
{'loss': 2.1663, 'grad_norm': 0.0, 'learning_rate': 0.11797833850790528, 'epoch': 0.46}
+
 46%|████▌     | 239/520 [14:57<18:06,  3.87s/it]
 46%|████▌     | 240/520 [15:01<18:26,  3.95s/it]
                                                 
{'loss': 2.0831, 'grad_norm': 0.0, 'learning_rate': 0.11736481776669305, 'epoch': 0.46}
+
 46%|████▌     | 240/520 [15:01<18:26,  3.95s/it]
 46%|████▋     | 241/520 [15:05<18:22,  3.95s/it]
                                                 
{'loss': 2.1549, 'grad_norm': 0.0, 'learning_rate': 0.11675062233047365, 'epoch': 0.46}
+
 46%|████▋     | 241/520 [15:05<18:22,  3.95s/it]
 47%|████▋     | 242/520 [15:09<18:15,  3.94s/it]
                                                 
{'loss': 2.0452, 'grad_norm': 0.0, 'learning_rate': 0.11613577606328068, 'epoch': 0.47}
+
 47%|████▋     | 242/520 [15:09<18:15,  3.94s/it]
 47%|████▋     | 243/520 [15:13<18:04,  3.91s/it]
                                                 
{'loss': 2.0951, 'grad_norm': 0.0, 'learning_rate': 0.11552030285443515, 'epoch': 0.47}
+
 47%|████▋     | 243/520 [15:13<18:04,  3.91s/it]
 47%|████▋     | 244/520 [15:16<17:55,  3.90s/it]
                                                 
{'loss': 2.0286, 'grad_norm': 0.0, 'learning_rate': 0.11490422661761744, 'epoch': 0.47}
+
 47%|████▋     | 244/520 [15:16<17:55,  3.90s/it]
 47%|████▋     | 245/520 [15:20<17:48,  3.89s/it]
                                                 
{'loss': 2.1138, 'grad_norm': 0.0, 'learning_rate': 0.11428757128993801, 'epoch': 0.47}
+
 47%|████▋     | 245/520 [15:20<17:48,  3.89s/it]
 47%|████▋     | 246/520 [15:24<17:46,  3.89s/it]
                                                 
{'loss': 1.9508, 'grad_norm': 0.0, 'learning_rate': 0.11367036083100734, 'epoch': 0.47}
+
 47%|████▋     | 246/520 [15:24<17:46,  3.89s/it]
 48%|████▊     | 247/520 [15:28<17:40,  3.88s/it]
                                                 
{'loss': 2.1498, 'grad_norm': 0.0, 'learning_rate': 0.11305261922200519, 'epoch': 0.47}
+
 48%|████▊     | 247/520 [15:28<17:40,  3.88s/it]
 48%|████▊     | 248/520 [15:32<17:33,  3.87s/it]
                                                 
{'loss': 2.1007, 'grad_norm': 0.0, 'learning_rate': 0.11243437046474854, 'epoch': 0.48}
+
 48%|████▊     | 248/520 [15:32<17:33,  3.87s/it]
 48%|████▊     | 249/520 [15:36<17:27,  3.87s/it]
                                                 
{'loss': 2.118, 'grad_norm': 0.0, 'learning_rate': 0.1118156385807593, 'epoch': 0.48}
+
 48%|████▊     | 249/520 [15:36<17:27,  3.87s/it]
 48%|████▊     | 250/520 [15:40<17:25,  3.87s/it]
                                                 
{'loss': 2.2819, 'grad_norm': 0.0, 'learning_rate': 0.11119644761033079, 'epoch': 0.48}
+
 48%|████▊     | 250/520 [15:40<17:25,  3.87s/it]
 48%|████▊     | 251/520 [15:43<17:15,  3.85s/it]
                                                 
{'loss': 2.1132, 'grad_norm': 0.0, 'learning_rate': 0.1105768216115938, 'epoch': 0.48}
+
 48%|████▊     | 251/520 [15:43<17:15,  3.85s/it]
 48%|████▊     | 252/520 [15:47<17:11,  3.85s/it]
                                                 
{'loss': 1.8759, 'grad_norm': 0.0, 'learning_rate': 0.10995678465958168, 'epoch': 0.48}
+
 48%|████▊     | 252/520 [15:47<17:11,  3.85s/it]
 49%|████▊     | 253/520 [15:51<17:02,  3.83s/it]
                                                 
{'loss': 2.2258, 'grad_norm': 0.0, 'learning_rate': 0.10933636084529506, 'epoch': 0.49}
+
 49%|████▊     | 253/520 [15:51<17:02,  3.83s/it]
 49%|████▉     | 254/520 [15:55<16:56,  3.82s/it]
                                                 
{'loss': 2.059, 'grad_norm': 0.0, 'learning_rate': 0.10871557427476584, 'epoch': 0.49}
+
 49%|████▉     | 254/520 [15:55<16:56,  3.82s/it]
 49%|████▉     | 255/520 [15:59<16:48,  3.80s/it]
                                                 
{'loss': 2.1846, 'grad_norm': 0.0, 'learning_rate': 0.10809444906812034, 'epoch': 0.49}
+
 49%|████▉     | 255/520 [15:59<16:48,  3.80s/it]
 49%|████▉     | 256/520 [16:02<16:35,  3.77s/it]
                                                 
{'loss': 2.1549, 'grad_norm': 0.0, 'learning_rate': 0.10747300935864244, 'epoch': 0.49}
+
 49%|████▉     | 256/520 [16:02<16:35,  3.77s/it]
 49%|████▉     | 257/520 [16:06<16:18,  3.72s/it]
                                                 
{'loss': 2.1188, 'grad_norm': 0.0, 'learning_rate': 0.10685127929183567, 'epoch': 0.49}
+
 49%|████▉     | 257/520 [16:06<16:18,  3.72s/it]
 50%|████▉     | 258/520 [16:10<16:07,  3.69s/it]
                                                 
{'loss': 2.0202, 'grad_norm': 0.0, 'learning_rate': 0.10622928302448523, 'epoch': 0.5}
+
 50%|████▉     | 258/520 [16:10<16:07,  3.69s/it]
 50%|████▉     | 259/520 [16:13<15:56,  3.67s/it]
                                                 
{'loss': 2.2765, 'grad_norm': 0.0, 'learning_rate': 0.10560704472371918, 'epoch': 0.5}
+
 50%|████▉     | 259/520 [16:13<15:56,  3.67s/it]
 50%|█████     | 260/520 [16:17<15:51,  3.66s/it]
                                                 
{'loss': 1.8251, 'grad_norm': 0.0, 'learning_rate': 0.10498458856606972, 'epoch': 0.5}
+
 50%|█████     | 260/520 [16:17<15:51,  3.66s/it]
 50%|█████     | 261/520 [16:20<15:45,  3.65s/it]
                                                 
{'loss': 1.9131, 'grad_norm': 0.0, 'learning_rate': 0.10436193873653361, 'epoch': 0.5}
+
 50%|█████     | 261/520 [16:20<15:45,  3.65s/it]
 50%|█████     | 262/520 [16:24<15:39,  3.64s/it]
                                                 
{'loss': 2.1102, 'grad_norm': 0.0, 'learning_rate': 0.10373911942763259, 'epoch': 0.5}
+
 50%|█████     | 262/520 [16:24<15:39,  3.64s/it]
 51%|█████     | 263/520 [16:28<15:35,  3.64s/it]
                                                 
{'loss': 1.9227, 'grad_norm': 0.0, 'learning_rate': 0.10311615483847332, 'epoch': 0.51}
+
 51%|█████     | 263/520 [16:28<15:35,  3.64s/it]
 51%|█████     | 264/520 [16:31<15:31,  3.64s/it]
                                                 
{'loss': 2.1143, 'grad_norm': 0.0, 'learning_rate': 0.1024930691738073, 'epoch': 0.51}
+
 51%|█████     | 264/520 [16:31<15:31,  3.64s/it]
 51%|█████     | 265/520 [16:35<15:31,  3.65s/it]
                                                 
{'loss': 2.276, 'grad_norm': 0.0, 'learning_rate': 0.10186988664309023, 'epoch': 0.51}
+
 51%|█████     | 265/520 [16:35<15:31,  3.65s/it]
 51%|█████     | 266/520 [16:39<15:25,  3.64s/it]
                                                 
{'loss': 1.9522, 'grad_norm': 0.0, 'learning_rate': 0.10124663145954152, 'epoch': 0.51}
+
 51%|█████     | 266/520 [16:39<15:25,  3.64s/it]
 51%|█████▏    | 267/520 [16:42<15:24,  3.65s/it]
                                                 
{'loss': 2.0106, 'grad_norm': 0.0, 'learning_rate': 0.10062332783920336, 'epoch': 0.51}
+
 51%|█████▏    | 267/520 [16:42<15:24,  3.65s/it]
 52%|█████▏    | 268/520 [16:46<15:28,  3.68s/it]
                                                 
{'loss': 2.0169, 'grad_norm': 0.0, 'learning_rate': 0.1, 'epoch': 0.52}
+
 52%|█████▏    | 268/520 [16:46<15:28,  3.68s/it]
 52%|█████▏    | 269/520 [16:50<15:21,  3.67s/it]
                                                 
{'loss': 2.2081, 'grad_norm': 0.0, 'learning_rate': 0.09937667216079665, 'epoch': 0.52}
+
 52%|█████▏    | 269/520 [16:50<15:21,  3.67s/it]
 52%|█████▏    | 270/520 [16:53<15:15,  3.66s/it]
                                                 
{'loss': 1.872, 'grad_norm': 0.0, 'learning_rate': 0.0987533685404585, 'epoch': 0.52}
+
 52%|█████▏    | 270/520 [16:53<15:15,  3.66s/it]
 52%|█████▏    | 271/520 [16:57<15:11,  3.66s/it]
                                                 
{'loss': 2.1221, 'grad_norm': 0.0, 'learning_rate': 0.0981301133569098, 'epoch': 0.52}
+
 52%|█████▏    | 271/520 [16:57<15:11,  3.66s/it]
 52%|█████▏    | 272/520 [17:01<15:07,  3.66s/it]
                                                 
{'loss': 1.8539, 'grad_norm': 0.0, 'learning_rate': 0.09750693082619273, 'epoch': 0.52}
+
 52%|█████▏    | 272/520 [17:01<15:07,  3.66s/it]
 52%|█████▎    | 273/520 [17:04<15:06,  3.67s/it]
                                                 
{'loss': 1.8384, 'grad_norm': 0.0, 'learning_rate': 0.0968838451615267, 'epoch': 0.53}
+
 52%|█████▎    | 273/520 [17:04<15:06,  3.67s/it]
 53%|█████▎    | 274/520 [17:08<15:04,  3.68s/it]
                                                 
{'loss': 2.1584, 'grad_norm': 0.0, 'learning_rate': 0.09626088057236745, 'epoch': 0.53}
+
 53%|█████▎    | 274/520 [17:08<15:04,  3.68s/it]
 53%|█████▎    | 275/520 [17:12<14:58,  3.67s/it]
                                                 
{'loss': 2.1887, 'grad_norm': 0.0, 'learning_rate': 0.09563806126346641, 'epoch': 0.53}
+
 53%|█████▎    | 275/520 [17:12<14:58,  3.67s/it]
 53%|█████▎    | 276/520 [17:15<14:56,  3.67s/it]
                                                 
{'loss': 2.1666, 'grad_norm': 0.0, 'learning_rate': 0.09501541143393027, 'epoch': 0.53}
+
 53%|█████▎    | 276/520 [17:15<14:56,  3.67s/it]
 53%|█████▎    | 277/520 [17:19<14:51,  3.67s/it]
                                                 
{'loss': 1.8919, 'grad_norm': 0.0, 'learning_rate': 0.09439295527628082, 'epoch': 0.53}
+
 53%|█████▎    | 277/520 [17:19<14:51,  3.67s/it]
 53%|█████▎    | 278/520 [17:23<14:48,  3.67s/it]
                                                 
{'loss': 2.0114, 'grad_norm': 0.0, 'learning_rate': 0.0937707169755148, 'epoch': 0.53}
+
 53%|█████▎    | 278/520 [17:23<14:48,  3.67s/it]
 54%|█████▎    | 279/520 [17:26<14:45,  3.67s/it]
                                                 
{'loss': 2.005, 'grad_norm': 0.0, 'learning_rate': 0.09314872070816434, 'epoch': 0.54}
+
 54%|█████▎    | 279/520 [17:26<14:45,  3.67s/it]
 54%|█████▍    | 280/520 [17:30<14:56,  3.73s/it]
                                                 
{'loss': 2.1648, 'grad_norm': 0.0, 'learning_rate': 0.09252699064135758, 'epoch': 0.54}
+
 54%|█████▍    | 280/520 [17:30<14:56,  3.73s/it]
 54%|█████▍    | 281/520 [17:34<15:01,  3.77s/it]
                                                 
{'loss': 2.2198, 'grad_norm': 0.0, 'learning_rate': 0.09190555093187967, 'epoch': 0.54}
+
 54%|█████▍    | 281/520 [17:34<15:01,  3.77s/it]
 54%|█████▍    | 282/520 [17:38<15:03,  3.80s/it]
                                                 
{'loss': 2.0525, 'grad_norm': 0.0, 'learning_rate': 0.09128442572523418, 'epoch': 0.54}
+
 54%|█████▍    | 282/520 [17:38<15:03,  3.80s/it]
 54%|█████▍    | 283/520 [17:42<15:04,  3.82s/it]
                                                 
{'loss': 2.189, 'grad_norm': 0.0, 'learning_rate': 0.09066363915470495, 'epoch': 0.54}
+
 54%|█████▍    | 283/520 [17:42<15:04,  3.82s/it]
 55%|█████▍    | 284/520 [17:46<15:05,  3.83s/it]
                                                 
{'loss': 1.9636, 'grad_norm': 0.0, 'learning_rate': 0.09004321534041836, 'epoch': 0.55}
+
 55%|█████▍    | 284/520 [17:46<15:05,  3.83s/it]
 55%|█████▍    | 285/520 [17:50<15:02,  3.84s/it]
                                                 
{'loss': 2.0822, 'grad_norm': 0.0, 'learning_rate': 0.08942317838840624, 'epoch': 0.55}
+
 55%|█████▍    | 285/520 [17:50<15:02,  3.84s/it]
 55%|█████▌    | 286/520 [17:53<15:00,  3.85s/it]
                                                 
{'loss': 2.1565, 'grad_norm': 0.0, 'learning_rate': 0.08880355238966922, 'epoch': 0.55}
+
 55%|█████▌    | 286/520 [17:53<15:00,  3.85s/it]
 55%|█████▌    | 287/520 [17:57<14:48,  3.81s/it]
                                                 
{'loss': 2.2603, 'grad_norm': 0.0, 'learning_rate': 0.08818436141924073, 'epoch': 0.55}
+
 55%|█████▌    | 287/520 [17:57<14:48,  3.81s/it]
 55%|█████▌    | 288/520 [18:01<14:32,  3.76s/it]
                                                 
{'loss': 2.0485, 'grad_norm': 0.0, 'learning_rate': 0.08756562953525152, 'epoch': 0.55}
+
 55%|█████▌    | 288/520 [18:01<14:32,  3.76s/it]
 56%|█████▌    | 289/520 [18:04<14:20,  3.72s/it]
                                                 
{'loss': 2.2023, 'grad_norm': 0.0, 'learning_rate': 0.08694738077799487, 'epoch': 0.56}
+
 56%|█████▌    | 289/520 [18:04<14:20,  3.72s/it]
 56%|█████▌    | 290/520 [18:08<14:11,  3.70s/it]
                                                 
{'loss': 1.914, 'grad_norm': 0.0, 'learning_rate': 0.08632963916899268, 'epoch': 0.56}
+
 56%|█████▌    | 290/520 [18:08<14:11,  3.70s/it]
 56%|█████▌    | 291/520 [18:12<14:17,  3.74s/it]
                                                 
{'loss': 2.0878, 'grad_norm': 0.0, 'learning_rate': 0.08571242871006202, 'epoch': 0.56}
+
 56%|█████▌    | 291/520 [18:12<14:17,  3.74s/it]
 56%|█████▌    | 292/520 [18:16<14:21,  3.78s/it]
                                                 
{'loss': 2.1345, 'grad_norm': 0.0, 'learning_rate': 0.08509577338238256, 'epoch': 0.56}
+
 56%|█████▌    | 292/520 [18:16<14:21,  3.78s/it]
 56%|█████▋    | 293/520 [18:20<14:22,  3.80s/it]
                                                 
{'loss': 2.1425, 'grad_norm': 0.0, 'learning_rate': 0.08447969714556484, 'epoch': 0.56}
+
 56%|█████▋    | 293/520 [18:20<14:22,  3.80s/it]
 57%|█████▋    | 294/520 [18:24<14:24,  3.83s/it]
                                                 
{'loss': 2.2303, 'grad_norm': 0.0, 'learning_rate': 0.08386422393671933, 'epoch': 0.57}
+
 57%|█████▋    | 294/520 [18:24<14:24,  3.83s/it]
 57%|█████▋    | 295/520 [18:27<14:24,  3.84s/it]
                                                 
{'loss': 1.832, 'grad_norm': 0.0, 'learning_rate': 0.08324937766952638, 'epoch': 0.57}
+
 57%|█████▋    | 295/520 [18:27<14:24,  3.84s/it]
 57%|█████▋    | 296/520 [18:31<14:22,  3.85s/it]
                                                 
{'loss': 2.0787, 'grad_norm': 0.0, 'learning_rate': 0.08263518223330697, 'epoch': 0.57}
+
 57%|█████▋    | 296/520 [18:31<14:22,  3.85s/it]
 57%|█████▋    | 297/520 [18:35<14:19,  3.86s/it]
                                                 
{'loss': 2.1771, 'grad_norm': 0.0, 'learning_rate': 0.08202166149209474, 'epoch': 0.57}
+
 57%|█████▋    | 297/520 [18:35<14:19,  3.86s/it]
 57%|█████▋    | 298/520 [18:39<14:15,  3.85s/it]
                                                 
{'loss': 1.9442, 'grad_norm': 0.0, 'learning_rate': 0.08140883928370855, 'epoch': 0.57}
+
 57%|█████▋    | 298/520 [18:39<14:15,  3.85s/it]
 57%|█████▊    | 299/520 [18:43<13:58,  3.79s/it]
                                                 
{'loss': 1.8794, 'grad_norm': 0.0, 'learning_rate': 0.0807967394188264, 'epoch': 0.57}
+
 57%|█████▊    | 299/520 [18:43<13:58,  3.79s/it]
 58%|█████▊    | 300/520 [18:46<13:45,  3.75s/it]
                                                 
{'loss': 2.1153, 'grad_norm': 0.0, 'learning_rate': 0.08018538568006027, 'epoch': 0.58}
+
 58%|█████▊    | 300/520 [18:46<13:45,  3.75s/it]
 58%|█████▊    | 301/520 [18:50<13:36,  3.73s/it]
                                                 
{'loss': 2.0707, 'grad_norm': 0.0, 'learning_rate': 0.07957480182103199, 'epoch': 0.58}
+
 58%|█████▊    | 301/520 [18:50<13:36,  3.73s/it]
 58%|█████▊    | 302/520 [18:54<13:28,  3.71s/it]
                                                 
{'loss': 1.8532, 'grad_norm': 0.0, 'learning_rate': 0.07896501156545044, 'epoch': 0.58}
+
 58%|█████▊    | 302/520 [18:54<13:28,  3.71s/it]
 58%|█████▊    | 303/520 [18:57<13:23,  3.70s/it]
                                                 
{'loss': 2.1844, 'grad_norm': 0.0, 'learning_rate': 0.07835603860618973, 'epoch': 0.58}
+
 58%|█████▊    | 303/520 [18:57<13:23,  3.70s/it]
 58%|█████▊    | 304/520 [19:01<13:22,  3.71s/it]
                                                 
{'loss': 2.081, 'grad_norm': 0.0, 'learning_rate': 0.07774790660436857, 'epoch': 0.58}
+
 58%|█████▊    | 304/520 [19:01<13:22,  3.71s/it]
 59%|█████▊    | 305/520 [19:05<13:15,  3.70s/it]
                                                 
{'loss': 2.1651, 'grad_norm': 0.0, 'learning_rate': 0.07714063918843106, 'epoch': 0.59}
+
 59%|█████▊    | 305/520 [19:05<13:15,  3.70s/it]
 59%|█████▉    | 306/520 [19:08<13:11,  3.70s/it]
                                                 
{'loss': 2.1107, 'grad_norm': 0.0, 'learning_rate': 0.0765342599532285, 'epoch': 0.59}
+
 59%|█████▉    | 306/520 [19:08<13:11,  3.70s/it]
 59%|█████▉    | 307/520 [19:13<13:29,  3.80s/it]
                                                 
{'loss': 2.0262, 'grad_norm': 0.0, 'learning_rate': 0.07592879245910272, 'epoch': 0.59}
+
 59%|█████▉    | 307/520 [19:13<13:29,  3.80s/it]
 59%|█████▉    | 308/520 [19:16<13:18,  3.77s/it]
                                                 
{'loss': 2.0165, 'grad_norm': 0.0, 'learning_rate': 0.07532426023097064, 'epoch': 0.59}
+
 59%|█████▉    | 308/520 [19:16<13:18,  3.77s/it]
 59%|█████▉    | 309/520 [19:20<13:08,  3.74s/it]
                                                 
{'loss': 1.9269, 'grad_norm': 0.0, 'learning_rate': 0.07472068675741024, 'epoch': 0.59}
+
 59%|█████▉    | 309/520 [19:20<13:08,  3.74s/it]
 60%|█████▉    | 310/520 [19:24<13:01,  3.72s/it]
                                                 
{'loss': 1.9954, 'grad_norm': 0.0, 'learning_rate': 0.07411809548974792, 'epoch': 0.6}
+
 60%|█████▉    | 310/520 [19:24<13:01,  3.72s/it]
 60%|█████▉    | 311/520 [19:27<12:53,  3.70s/it]
                                                 
{'loss': 2.065, 'grad_norm': 0.0, 'learning_rate': 0.07351650984114727, 'epoch': 0.6}
+
 60%|█████▉    | 311/520 [19:27<12:53,  3.70s/it]
 60%|██████    | 312/520 [19:31<12:50,  3.70s/it]
                                                 
{'loss': 2.1635, 'grad_norm': 0.0, 'learning_rate': 0.0729159531856995, 'epoch': 0.6}
+
 60%|██████    | 312/520 [19:31<12:50,  3.70s/it]
 60%|██████    | 313/520 [19:35<12:44,  3.69s/it]
                                                 
{'loss': 1.8959, 'grad_norm': 0.0, 'learning_rate': 0.07231644885751508, 'epoch': 0.6}
+
 60%|██████    | 313/520 [19:35<12:44,  3.69s/it]
 60%|██████    | 314/520 [19:39<13:07,  3.82s/it]
                                                 
{'loss': 2.0684, 'grad_norm': 0.0, 'learning_rate': 0.07171802014981725, 'epoch': 0.6}
+
 60%|██████    | 314/520 [19:39<13:07,  3.82s/it]
 61%|██████    | 315/520 [19:42<12:53,  3.77s/it]
                                                 
{'loss': 2.0994, 'grad_norm': 0.0, 'learning_rate': 0.07112069031403703, 'epoch': 0.61}
+
 61%|██████    | 315/520 [19:42<12:53,  3.77s/it]
 61%|██████    | 316/520 [19:47<13:14,  3.89s/it]
                                                 
{'loss': 2.1863, 'grad_norm': 0.0, 'learning_rate': 0.07052448255890957, 'epoch': 0.61}
+
 61%|██████    | 316/520 [19:47<13:14,  3.89s/it]
 61%|██████    | 317/520 [19:50<12:55,  3.82s/it]
                                                 
{'loss': 1.9533, 'grad_norm': 0.0, 'learning_rate': 0.0699294200495727, 'epoch': 0.61}
+
 61%|██████    | 317/520 [19:50<12:55,  3.82s/it]
 61%|██████    | 318/520 [19:54<12:41,  3.77s/it]
                                                 
{'loss': 2.2686, 'grad_norm': 0.0, 'learning_rate': 0.06933552590666658, 'epoch': 0.61}
+
 61%|██████    | 318/520 [19:54<12:41,  3.77s/it]
 61%|██████▏   | 319/520 [19:58<12:51,  3.84s/it]
                                                 
{'loss': 1.8863, 'grad_norm': 0.0, 'learning_rate': 0.06874282320543557, 'epoch': 0.61}
+
 61%|██████▏   | 319/520 [19:58<12:51,  3.84s/it]
 62%|██████▏   | 320/520 [20:02<12:37,  3.79s/it]
                                                 
{'loss': 2.0865, 'grad_norm': 0.0, 'learning_rate': 0.06815133497483157, 'epoch': 0.62}
+
 62%|██████▏   | 320/520 [20:02<12:37,  3.79s/it]
 62%|██████▏   | 321/520 [20:05<12:28,  3.76s/it]
                                                 
{'loss': 2.0712, 'grad_norm': 0.0, 'learning_rate': 0.06756108419661931, 'epoch': 0.62}
+
 62%|██████▏   | 321/520 [20:05<12:28,  3.76s/it]
 62%|██████▏   | 322/520 [20:09<12:18,  3.73s/it]
                                                 
{'loss': 1.8969, 'grad_norm': 0.0, 'learning_rate': 0.06697209380448332, 'epoch': 0.62}
+
 62%|██████▏   | 322/520 [20:09<12:18,  3.73s/it]
 62%|██████▏   | 323/520 [20:12<12:08,  3.70s/it]
                                                 
{'loss': 2.0202, 'grad_norm': 0.0, 'learning_rate': 0.06638438668313694, 'epoch': 0.62}
+
 62%|██████▏   | 323/520 [20:12<12:08,  3.70s/it]
 62%|██████▏   | 324/520 [20:16<12:00,  3.68s/it]
                                                 
{'loss': 2.0551, 'grad_norm': 0.0, 'learning_rate': 0.06579798566743314, 'epoch': 0.62}
+
 62%|██████▏   | 324/520 [20:16<12:00,  3.68s/it]
 62%|██████▎   | 325/520 [20:20<11:54,  3.66s/it]
                                                 
{'loss': 2.1566, 'grad_norm': 0.0, 'learning_rate': 0.06521291354147728, 'epoch': 0.62}
+
 62%|██████▎   | 325/520 [20:20<11:54,  3.66s/it]
 63%|██████▎   | 326/520 [20:23<11:50,  3.66s/it]
                                                 
{'loss': 2.1909, 'grad_norm': 0.0, 'learning_rate': 0.06462919303774187, 'epoch': 0.63}
+
 63%|██████▎   | 326/520 [20:23<11:50,  3.66s/it]
 63%|██████▎   | 327/520 [20:27<11:47,  3.67s/it]
                                                 
{'loss': 2.061, 'grad_norm': 0.0, 'learning_rate': 0.06404684683618325, 'epoch': 0.63}
+
 63%|██████▎   | 327/520 [20:27<11:47,  3.67s/it]
 63%|██████▎   | 328/520 [20:31<11:44,  3.67s/it]
                                                 
{'loss': 2.1111, 'grad_norm': 0.0, 'learning_rate': 0.0634658975633605, 'epoch': 0.63}
+
 63%|██████▎   | 328/520 [20:31<11:44,  3.67s/it]
 63%|██████▎   | 329/520 [20:35<11:50,  3.72s/it]
                                                 
{'loss': 1.9436, 'grad_norm': 0.0, 'learning_rate': 0.06288636779155621, 'epoch': 0.63}
+
 63%|██████▎   | 329/520 [20:35<11:50,  3.72s/it]
 63%|██████▎   | 330/520 [20:38<11:44,  3.71s/it]
                                                 
{'loss': 2.1281, 'grad_norm': 0.0, 'learning_rate': 0.06230828003789948, 'epoch': 0.63}
+
 63%|██████▎   | 330/520 [20:38<11:44,  3.71s/it]
 64%|██████▎   | 331/520 [20:42<11:37,  3.69s/it]
                                                 
{'loss': 2.1551, 'grad_norm': 0.0, 'learning_rate': 0.06173165676349103, 'epoch': 0.64}
+
 64%|██████▎   | 331/520 [20:42<11:37,  3.69s/it]
 64%|██████▍   | 332/520 [20:46<11:34,  3.70s/it]
                                                 
{'loss': 1.8491, 'grad_norm': 0.0, 'learning_rate': 0.06115652037253053, 'epoch': 0.64}
+
 64%|██████▍   | 332/520 [20:46<11:34,  3.70s/it]
 64%|██████▍   | 333/520 [20:49<11:27,  3.67s/it]
                                                 
{'loss': 2.1184, 'grad_norm': 0.0, 'learning_rate': 0.06058289321144608, 'epoch': 0.64}
+
 64%|██████▍   | 333/520 [20:49<11:27,  3.67s/it]
 64%|██████▍   | 334/520 [20:53<11:22,  3.67s/it]
                                                 
{'loss': 2.1157, 'grad_norm': 0.0, 'learning_rate': 0.06001079756802592, 'epoch': 0.64}
+
 64%|██████▍   | 334/520 [20:53<11:22,  3.67s/it]
 64%|██████▍   | 335/520 [20:57<11:19,  3.67s/it]
                                                 
{'loss': 2.013, 'grad_norm': 0.0, 'learning_rate': 0.059440255670552514, 'epoch': 0.64}
+
 64%|██████▍   | 335/520 [20:57<11:19,  3.67s/it]
 65%|██████▍   | 336/520 [21:00<11:15,  3.67s/it]
                                                 
{'loss': 2.1874, 'grad_norm': 0.0, 'learning_rate': 0.05887128968693887, 'epoch': 0.65}
+
 65%|██████▍   | 336/520 [21:00<11:15,  3.67s/it]
 65%|██████▍   | 337/520 [21:04<11:10,  3.67s/it]
                                                 
{'loss': 2.2477, 'grad_norm': 0.0, 'learning_rate': 0.058303921723867225, 'epoch': 0.65}
+
 65%|██████▍   | 337/520 [21:04<11:10,  3.67s/it]
 65%|██████▌   | 338/520 [21:08<11:08,  3.67s/it]
                                                 
{'loss': 2.1774, 'grad_norm': 0.0, 'learning_rate': 0.05773817382593008, 'epoch': 0.65}
+
 65%|██████▌   | 338/520 [21:08<11:08,  3.67s/it]
 65%|██████▌   | 339/520 [21:11<11:04,  3.67s/it]
                                                 
{'loss': 2.126, 'grad_norm': 0.0, 'learning_rate': 0.057174067974773715, 'epoch': 0.65}
+
 65%|██████▌   | 339/520 [21:11<11:04,  3.67s/it]
 65%|██████▌   | 340/520 [21:15<11:01,  3.68s/it]
                                                 
{'loss': 2.0845, 'grad_norm': 0.0, 'learning_rate': 0.056611626088244195, 'epoch': 0.65}
+
 65%|██████▌   | 340/520 [21:15<11:01,  3.68s/it]
 66%|██████▌   | 341/520 [21:19<10:56,  3.67s/it]
                                                 
{'loss': 2.094, 'grad_norm': 0.0, 'learning_rate': 0.056050870019535494, 'epoch': 0.66}
+
 66%|██████▌   | 341/520 [21:19<10:56,  3.67s/it]
 66%|██████▌   | 342/520 [21:22<10:50,  3.65s/it]
                                                 
{'loss': 2.0199, 'grad_norm': 0.0, 'learning_rate': 0.05549182155634076, 'epoch': 0.66}
+
 66%|██████▌   | 342/520 [21:22<10:50,  3.65s/it]
 66%|██████▌   | 343/520 [21:26<10:47,  3.66s/it]
                                                 
{'loss': 1.7182, 'grad_norm': 0.0, 'learning_rate': 0.054934502420005464, 'epoch': 0.66}
+
 66%|██████▌   | 343/520 [21:26<10:47,  3.66s/it]
 66%|██████▌   | 344/520 [21:30<10:44,  3.66s/it]
                                                 
{'loss': 2.1244, 'grad_norm': 0.0, 'learning_rate': 0.0543789342646837, 'epoch': 0.66}
+
 66%|██████▌   | 344/520 [21:30<10:44,  3.66s/it]
 66%|██████▋   | 345/520 [21:33<10:40,  3.66s/it]
                                                 
{'loss': 2.2588, 'grad_norm': 0.0, 'learning_rate': 0.05382513867649663, 'epoch': 0.66}
+
 66%|██████▋   | 345/520 [21:33<10:40,  3.66s/it]
 67%|██████▋   | 346/520 [21:37<10:37,  3.66s/it]
                                                 
{'loss': 1.859, 'grad_norm': 0.0, 'learning_rate': 0.0532731371726938, 'epoch': 0.67}
+
 67%|██████▋   | 346/520 [21:37<10:37,  3.66s/it]
 67%|██████▋   | 347/520 [21:41<10:33,  3.66s/it]
                                                 
{'loss': 1.9277, 'grad_norm': 0.0, 'learning_rate': 0.05272295120081732, 'epoch': 0.67}
+
 67%|██████▋   | 347/520 [21:41<10:33,  3.66s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2778 > 2048). Running this sequence through the model will result in indexing errors
+
 67%|██████▋   | 348/520 [21:44<10:28,  3.65s/it]
                                                 
{'loss': 2.405, 'grad_norm': 0.0, 'learning_rate': 0.05217460213786821, 'epoch': 0.67}
+
 67%|██████▋   | 348/520 [21:44<10:28,  3.65s/it]
 67%|██████▋   | 349/520 [21:48<10:24,  3.65s/it]
                                                 
{'loss': 2.2236, 'grad_norm': 0.0, 'learning_rate': 0.051628111289476024, 'epoch': 0.67}
+
 67%|██████▋   | 349/520 [21:48<10:24,  3.65s/it]
 67%|██████▋   | 350/520 [21:51<10:19,  3.65s/it]
                                                 
{'loss': 2.1184, 'grad_norm': 0.0, 'learning_rate': 0.051083499889071106, 'epoch': 0.67}
+
 67%|██████▋   | 350/520 [21:51<10:19,  3.65s/it]
 68%|██████▊   | 351/520 [21:55<10:17,  3.66s/it]
                                                 
{'loss': 2.0414, 'grad_norm': 0.0, 'learning_rate': 0.05054078909705926, 'epoch': 0.68}
+
 68%|██████▊   | 351/520 [21:55<10:17,  3.66s/it]
 68%|██████▊   | 352/520 [21:59<10:15,  3.66s/it]
                                                 
{'loss': 2.0824, 'grad_norm': 0.0, 'learning_rate': 0.050000000000000024, 'epoch': 0.68}
+
 68%|██████▊   | 352/520 [21:59<10:15,  3.66s/it]
 68%|██████▊   | 353/520 [22:02<10:12,  3.67s/it]
                                                 
{'loss': 1.8115, 'grad_norm': 0.0, 'learning_rate': 0.04946115360978696, 'epoch': 0.68}
+
 68%|██████▊   | 353/520 [22:03<10:12,  3.67s/it]
 68%|██████▊   | 354/520 [22:06<10:09,  3.67s/it]
                                                 
{'loss': 1.8797, 'grad_norm': 0.0, 'learning_rate': 0.048924270862831465, 'epoch': 0.68}
+
 68%|██████▊   | 354/520 [22:06<10:09,  3.67s/it]
 68%|██████▊   | 355/520 [22:10<10:03,  3.66s/it]
                                                 
{'loss': 2.0561, 'grad_norm': 0.0, 'learning_rate': 0.04838937261924933, 'epoch': 0.68}
+
 68%|██████▊   | 355/520 [22:10<10:03,  3.66s/it]
 68%|██████▊   | 356/520 [22:13<09:57,  3.65s/it]
                                                 
{'loss': 2.2531, 'grad_norm': 0.0, 'learning_rate': 0.0478564796620502, 'epoch': 0.68}
+
 68%|██████▊   | 356/520 [22:13<09:57,  3.65s/it]
 69%|██████▊   | 357/520 [22:17<09:53,  3.64s/it]
                                                 
{'loss': 2.0294, 'grad_norm': 0.0, 'learning_rate': 0.04732561269632993, 'epoch': 0.69}
+
 69%|██████▊   | 357/520 [22:17<09:53,  3.64s/it]
 69%|██████▉   | 358/520 [22:21<09:50,  3.65s/it]
                                                 
{'loss': 2.0531, 'grad_norm': 0.0, 'learning_rate': 0.04679679234846636, 'epoch': 0.69}
+
 69%|██████▉   | 358/520 [22:21<09:50,  3.65s/it]
 69%|██████▉   | 359/520 [22:24<09:49,  3.66s/it]
                                                 
{'loss': 2.011, 'grad_norm': 0.0, 'learning_rate': 0.046270039165317606, 'epoch': 0.69}
+
 69%|██████▉   | 359/520 [22:24<09:49,  3.66s/it]
 69%|██████▉   | 360/520 [22:28<09:46,  3.67s/it]
                                                 
{'loss': 1.9948, 'grad_norm': 0.0, 'learning_rate': 0.04574537361342407, 'epoch': 0.69}
+
 69%|██████▉   | 360/520 [22:28<09:46,  3.67s/it]
 69%|██████▉   | 361/520 [22:32<09:41,  3.66s/it]
                                                 
{'loss': 1.7504, 'grad_norm': 0.0, 'learning_rate': 0.04522281607821288, 'epoch': 0.69}
+
 69%|██████▉   | 361/520 [22:32<09:41,  3.66s/it]
 70%|██████▉   | 362/520 [22:35<09:38,  3.66s/it]
                                                 
{'loss': 2.2058, 'grad_norm': 0.0, 'learning_rate': 0.04470238686320606, 'epoch': 0.7}
+
 70%|██████▉   | 362/520 [22:35<09:38,  3.66s/it]
 70%|██████▉   | 363/520 [22:39<09:35,  3.66s/it]
                                                 
{'loss': 2.0753, 'grad_norm': 0.0, 'learning_rate': 0.044184106189231624, 'epoch': 0.7}
+
 70%|██████▉   | 363/520 [22:39<09:35,  3.66s/it]
 70%|███████   | 364/520 [22:43<09:31,  3.66s/it]
                                                 
{'loss': 1.9811, 'grad_norm': 0.0, 'learning_rate': 0.043667994193637795, 'epoch': 0.7}
+
 70%|███████   | 364/520 [22:43<09:31,  3.66s/it]
 70%|███████   | 365/520 [22:46<09:28,  3.67s/it]
                                                 
{'loss': 2.1137, 'grad_norm': 0.0, 'learning_rate': 0.043154070929510784, 'epoch': 0.7}
+
 70%|███████   | 365/520 [22:46<09:28,  3.67s/it]
 70%|███████   | 366/520 [22:50<09:25,  3.67s/it]
                                                 
{'loss': 2.1027, 'grad_norm': 0.0, 'learning_rate': 0.04264235636489542, 'epoch': 0.7}
+
 70%|███████   | 366/520 [22:50<09:25,  3.67s/it]
 71%|███████   | 367/520 [22:54<09:22,  3.68s/it]
                                                 
{'loss': 2.1701, 'grad_norm': 0.0, 'learning_rate': 0.04213287038201943, 'epoch': 0.71}
+
 71%|███████   | 367/520 [22:54<09:22,  3.68s/it]
 71%|███████   | 368/520 [22:57<09:20,  3.69s/it]
                                                 
{'loss': 2.1175, 'grad_norm': 0.0, 'learning_rate': 0.04162563277652104, 'epoch': 0.71}
+
 71%|███████   | 368/520 [22:57<09:20,  3.69s/it]
 71%|███████   | 369/520 [23:01<09:15,  3.68s/it]
                                                 
{'loss': 1.7789, 'grad_norm': 0.0, 'learning_rate': 0.04112066325667954, 'epoch': 0.71}
+
 71%|███████   | 369/520 [23:01<09:15,  3.68s/it]
 71%|███████   | 370/520 [23:05<09:14,  3.70s/it]
                                                 
{'loss': 2.0015, 'grad_norm': 0.0, 'learning_rate': 0.04061798144264986, 'epoch': 0.71}
+
 71%|███████   | 370/520 [23:05<09:14,  3.70s/it]
 71%|███████▏  | 371/520 [23:09<09:08,  3.68s/it]
                                                 
{'loss': 2.1704, 'grad_norm': 0.0, 'learning_rate': 0.04011760686569998, 'epoch': 0.71}
+
 71%|███████▏  | 371/520 [23:09<09:08,  3.68s/it]
 72%|███████▏  | 372/520 [23:12<09:04,  3.68s/it]
                                                 
{'loss': 1.8294, 'grad_norm': 0.0, 'learning_rate': 0.03961955896745224, 'epoch': 0.72}
+
 72%|███████▏  | 372/520 [23:12<09:04,  3.68s/it]
 72%|███████▏  | 373/520 [23:16<08:59,  3.67s/it]
                                                 
{'loss': 2.0181, 'grad_norm': 0.0, 'learning_rate': 0.03912385709912794, 'epoch': 0.72}
+
 72%|███████▏  | 373/520 [23:16<08:59,  3.67s/it]
 72%|███████▏  | 374/520 [23:20<08:55,  3.67s/it]
                                                 
{'loss': 2.1018, 'grad_norm': 0.0, 'learning_rate': 0.038630520520795276, 'epoch': 0.72}
+
 72%|███████▏  | 374/520 [23:20<08:55,  3.67s/it]
 72%|███████▏  | 375/520 [23:23<08:51,  3.67s/it]
                                                 
{'loss': 2.1132, 'grad_norm': 0.0, 'learning_rate': 0.03813956840062119, 'epoch': 0.72}
+
 72%|███████▏  | 375/520 [23:23<08:51,  3.67s/it]
 72%|███████▏  | 376/520 [23:27<08:50,  3.68s/it]
                                                 
{'loss': 2.0573, 'grad_norm': 0.0, 'learning_rate': 0.037651019814126656, 'epoch': 0.72}
+
 72%|███████▏  | 376/520 [23:27<08:50,  3.68s/it]
 72%|███████▎  | 377/520 [23:31<08:46,  3.68s/it]
                                                 
{'loss': 2.0899, 'grad_norm': 0.0, 'learning_rate': 0.037164893743445275, 'epoch': 0.72}
+
 72%|███████▎  | 377/520 [23:31<08:46,  3.68s/it]
 73%|███████▎  | 378/520 [23:34<08:41,  3.67s/it]
                                                 
{'loss': 2.0289, 'grad_norm': 0.0, 'learning_rate': 0.03668120907658603, 'epoch': 0.73}
+
 73%|███████▎  | 378/520 [23:34<08:41,  3.67s/it]
 73%|███████▎  | 379/520 [23:38<08:36,  3.67s/it]
                                                 
{'loss': 1.9774, 'grad_norm': 0.0, 'learning_rate': 0.036199984606699154, 'epoch': 0.73}
+
 73%|███████▎  | 379/520 [23:38<08:36,  3.67s/it]
 73%|███████▎  | 380/520 [23:42<08:35,  3.68s/it]
                                                 
{'loss': 1.8319, 'grad_norm': 0.0, 'learning_rate': 0.035721239031346066, 'epoch': 0.73}
+
 73%|███████▎  | 380/520 [23:42<08:35,  3.68s/it]
 73%|███████▎  | 381/520 [23:45<08:38,  3.73s/it]
                                                 
{'loss': 2.0371, 'grad_norm': 0.0, 'learning_rate': 0.03524499095177297, 'epoch': 0.73}
+
 73%|███████▎  | 381/520 [23:45<08:38,  3.73s/it]
 73%|███████▎  | 382/520 [23:49<08:40,  3.77s/it]
                                                 
{'loss': 1.9153, 'grad_norm': 0.0, 'learning_rate': 0.03477125887218792, 'epoch': 0.73}
+
 73%|███████▎  | 382/520 [23:49<08:40,  3.77s/it]
 74%|███████▎  | 383/520 [23:53<08:38,  3.79s/it]
                                                 
{'loss': 2.2443, 'grad_norm': 0.0, 'learning_rate': 0.03430006119904196, 'epoch': 0.74}
+
 74%|███████▎  | 383/520 [23:53<08:38,  3.79s/it]
 74%|███████▍  | 384/520 [23:57<08:36,  3.80s/it]
                                                 
{'loss': 1.6572, 'grad_norm': 0.0, 'learning_rate': 0.033831416240314084, 'epoch': 0.74}
+
 74%|███████▍  | 384/520 [23:57<08:36,  3.80s/it]
 74%|███████▍  | 385/520 [24:01<08:33,  3.80s/it]
                                                 
{'loss': 1.9484, 'grad_norm': 0.0, 'learning_rate': 0.03336534220479961, 'epoch': 0.74}
+
 74%|███████▍  | 385/520 [24:01<08:33,  3.80s/it]
 74%|███████▍  | 386/520 [24:05<08:31,  3.81s/it]
                                                 
{'loss': 2.0001, 'grad_norm': 0.0, 'learning_rate': 0.032901857201403005, 'epoch': 0.74}
+
 74%|███████▍  | 386/520 [24:05<08:31,  3.81s/it]
 74%|███████▍  | 387/520 [24:08<08:28,  3.83s/it]
                                                 
{'loss': 1.7967, 'grad_norm': 0.0, 'learning_rate': 0.032440979238433976, 'epoch': 0.74}
+
 74%|███████▍  | 387/520 [24:08<08:28,  3.83s/it]
 75%|███████▍  | 388/520 [24:12<08:25,  3.83s/it]
                                                 
{'loss': 2.1252, 'grad_norm': 0.0, 'learning_rate': 0.03198272622290804, 'epoch': 0.75}
+
 75%|███████▍  | 388/520 [24:12<08:25,  3.83s/it]
 75%|███████▍  | 389/520 [24:16<08:22,  3.84s/it]
                                                 
{'loss': 2.2819, 'grad_norm': 0.0, 'learning_rate': 0.03152711595985065, 'epoch': 0.75}
+
 75%|███████▍  | 389/520 [24:16<08:22,  3.84s/it]
 75%|███████▌  | 390/520 [24:20<08:19,  3.84s/it]
                                                 
{'loss': 2.0794, 'grad_norm': 0.0, 'learning_rate': 0.031074166151605298, 'epoch': 0.75}
+
 75%|███████▌  | 390/520 [24:20<08:19,  3.84s/it]
 75%|███████▌  | 391/520 [24:24<08:17,  3.86s/it]
                                                 
{'loss': 2.0751, 'grad_norm': 0.0, 'learning_rate': 0.030623894397145836, 'epoch': 0.75}
+
 75%|███████▌  | 391/520 [24:24<08:17,  3.86s/it]
 75%|███████▌  | 392/520 [24:28<08:10,  3.83s/it]
                                                 
{'loss': 2.0834, 'grad_norm': 0.0, 'learning_rate': 0.03017631819139273, 'epoch': 0.75}
+
 75%|███████▌  | 392/520 [24:28<08:10,  3.83s/it]
 76%|███████▌  | 393/520 [24:31<08:01,  3.79s/it]
                                                 
{'loss': 1.6935, 'grad_norm': 0.0, 'learning_rate': 0.029731454924533086, 'epoch': 0.76}
+
 76%|███████▌  | 393/520 [24:31<08:01,  3.79s/it]
 76%|███████▌  | 394/520 [24:35<07:51,  3.74s/it]
                                                 
{'loss': 2.1218, 'grad_norm': 0.0, 'learning_rate': 0.029289321881345254, 'epoch': 0.76}
+
 76%|███████▌  | 394/520 [24:35<07:51,  3.74s/it]
 76%|███████▌  | 395/520 [24:39<07:44,  3.71s/it]
                                                 
{'loss': 2.1493, 'grad_norm': 0.0, 'learning_rate': 0.028849936240527008, 'epoch': 0.76}
+
 76%|███████▌  | 395/520 [24:39<07:44,  3.71s/it]
 76%|███████▌  | 396/520 [24:42<07:36,  3.68s/it]
                                                 
{'loss': 2.0961, 'grad_norm': 0.0, 'learning_rate': 0.028413315074028157, 'epoch': 0.76}
+
 76%|███████▌  | 396/520 [24:42<07:36,  3.68s/it]
 76%|███████▋  | 397/520 [24:46<07:32,  3.68s/it]
                                                 
{'loss': 2.0472, 'grad_norm': 0.0, 'learning_rate': 0.027979475346387363, 'epoch': 0.76}
+
 76%|███████▋  | 397/520 [24:46<07:32,  3.68s/it]
 77%|███████▋  | 398/520 [24:50<07:26,  3.66s/it]
                                                 
{'loss': 2.2098, 'grad_norm': 0.0, 'learning_rate': 0.027548433914072735, 'epoch': 0.77}
+
 77%|███████▋  | 398/520 [24:50<07:26,  3.66s/it]
 77%|███████▋  | 399/520 [24:53<07:22,  3.66s/it]
                                                 
{'loss': 1.8453, 'grad_norm': 0.0, 'learning_rate': 0.027120207524827168, 'epoch': 0.77}
+
 77%|███████▋  | 399/520 [24:53<07:22,  3.66s/it]
 77%|███████▋  | 400/520 [24:57<07:20,  3.67s/it]
                                                 
{'loss': 1.8965, 'grad_norm': 0.0, 'learning_rate': 0.02669481281701739, 'epoch': 0.77}
+
 77%|███████▋  | 400/520 [24:57<07:20,  3.67s/it]
 77%|███████▋  | 401/520 [25:01<07:16,  3.67s/it]
                                                 
{'loss': 2.0165, 'grad_norm': 0.0, 'learning_rate': 0.026272266318987603, 'epoch': 0.77}
+
 77%|███████▋  | 401/520 [25:01<07:16,  3.67s/it]
 77%|███████▋  | 402/520 [25:04<07:12,  3.67s/it]
                                                 
{'loss': 2.1351, 'grad_norm': 0.0, 'learning_rate': 0.02585258444841733, 'epoch': 0.77}
+
 77%|███████▋  | 402/520 [25:04<07:12,  3.67s/it]
 78%|███████▊  | 403/520 [25:08<07:09,  3.67s/it]
                                                 
{'loss': 2.1219, 'grad_norm': 0.0, 'learning_rate': 0.025435783511683442, 'epoch': 0.78}
+
 78%|███████▊  | 403/520 [25:08<07:09,  3.67s/it]
 78%|███████▊  | 404/520 [25:12<07:04,  3.66s/it]
                                                 
{'loss': 2.2818, 'grad_norm': 0.0, 'learning_rate': 0.02502187970322657, 'epoch': 0.78}
+
 78%|███████▊  | 404/520 [25:12<07:04,  3.66s/it]
 78%|███████▊  | 405/520 [25:15<07:01,  3.66s/it]
                                                 
{'loss': 1.875, 'grad_norm': 0.0, 'learning_rate': 0.02461088910492202, 'epoch': 0.78}
+
 78%|███████▊  | 405/520 [25:15<07:01,  3.66s/it]
 78%|███████▊  | 406/520 [25:19<06:58,  3.67s/it]
                                                 
{'loss': 2.1806, 'grad_norm': 0.0, 'learning_rate': 0.02420282768545469, 'epoch': 0.78}
+
 78%|███████▊  | 406/520 [25:19<06:58,  3.67s/it]
 78%|███████▊  | 407/520 [25:23<06:54,  3.67s/it]
                                                 
{'loss': 2.0986, 'grad_norm': 0.0, 'learning_rate': 0.02379771129969892, 'epoch': 0.78}
+
 78%|███████▊  | 407/520 [25:23<06:54,  3.67s/it]
 78%|███████▊  | 408/520 [25:26<06:50,  3.66s/it]
                                                 
{'loss': 2.1517, 'grad_norm': 0.0, 'learning_rate': 0.023395555688102213, 'epoch': 0.78}
+
 78%|███████▊  | 408/520 [25:26<06:50,  3.66s/it]
 79%|███████▊  | 409/520 [25:30<06:46,  3.66s/it]
                                                 
{'loss': 2.2385, 'grad_norm': 0.0, 'learning_rate': 0.02299637647607372, 'epoch': 0.79}
+
 79%|███████▊  | 409/520 [25:30<06:46,  3.66s/it]
 79%|███████▉  | 410/520 [25:34<06:42,  3.65s/it]
                                                 
{'loss': 2.1727, 'grad_norm': 0.0, 'learning_rate': 0.022600189173377264, 'epoch': 0.79}
+
 79%|███████▉  | 410/520 [25:34<06:42,  3.65s/it]
 79%|███████▉  | 411/520 [25:37<06:38,  3.66s/it]
                                                 
{'loss': 2.195, 'grad_norm': 0.0, 'learning_rate': 0.022207009173528525, 'epoch': 0.79}
+
 79%|███████▉  | 411/520 [25:37<06:38,  3.66s/it]
 79%|███████▉  | 412/520 [25:41<06:34,  3.65s/it]
                                                 
{'loss': 2.0965, 'grad_norm': 0.0, 'learning_rate': 0.02181685175319702, 'epoch': 0.79}
+
 79%|███████▉  | 412/520 [25:41<06:34,  3.65s/it]
 79%|███████▉  | 413/520 [25:44<06:30,  3.65s/it]
                                                 
{'loss': 1.916, 'grad_norm': 0.0, 'learning_rate': 0.021429732071612653, 'epoch': 0.79}
+
 79%|███████▉  | 413/520 [25:44<06:30,  3.65s/it]
 80%|███████▉  | 414/520 [25:48<06:27,  3.66s/it]
                                                 
{'loss': 1.757, 'grad_norm': 0.0, 'learning_rate': 0.02104566516997647, 'epoch': 0.8}
+
 80%|███████▉  | 414/520 [25:48<06:27,  3.66s/it]
 80%|███████▉  | 415/520 [25:52<06:24,  3.66s/it]
                                                 
{'loss': 2.0894, 'grad_norm': 0.0, 'learning_rate': 0.020664665970876496, 'epoch': 0.8}
+
 80%|███████▉  | 415/520 [25:52<06:24,  3.66s/it]
 80%|████████  | 416/520 [25:55<06:20,  3.66s/it]
                                                 
{'loss': 2.3404, 'grad_norm': 0.0, 'learning_rate': 0.020286749277707784, 'epoch': 0.8}
+
 80%|████████  | 416/520 [25:55<06:20,  3.66s/it]
 80%|████████  | 417/520 [25:59<06:16,  3.66s/it]
                                                 
{'loss': 2.0376, 'grad_norm': 0.0, 'learning_rate': 0.019911929774097215, 'epoch': 0.8}
+
 80%|████████  | 417/520 [25:59<06:16,  3.66s/it]
 80%|████████  | 418/520 [26:03<06:12,  3.65s/it]
                                                 
{'loss': 1.9876, 'grad_norm': 0.0, 'learning_rate': 0.019540222023333165, 'epoch': 0.8}
+
 80%|████████  | 418/520 [26:03<06:12,  3.65s/it]
 81%|████████  | 419/520 [26:06<06:08,  3.65s/it]
                                                 
{'loss': 2.2701, 'grad_norm': 0.0, 'learning_rate': 0.01917164046779948, 'epoch': 0.81}
+
 81%|████████  | 419/520 [26:06<06:08,  3.65s/it]
 81%|████████  | 420/520 [26:10<06:05,  3.66s/it]
                                                 
{'loss': 2.1783, 'grad_norm': 0.0, 'learning_rate': 0.018806199428414352, 'epoch': 0.81}
+
 81%|████████  | 420/520 [26:10<06:05,  3.66s/it]
 81%|████████  | 421/520 [26:14<06:02,  3.66s/it]
                                                 
{'loss': 2.3788, 'grad_norm': 0.0, 'learning_rate': 0.018443913104073985, 'epoch': 0.81}
+
 81%|████████  | 421/520 [26:14<06:02,  3.66s/it]
 81%|████████  | 422/520 [26:17<05:59,  3.67s/it]
                                                 
{'loss': 2.1751, 'grad_norm': 0.0, 'learning_rate': 0.01808479557110081, 'epoch': 0.81}
+
 81%|████████  | 422/520 [26:17<05:59,  3.67s/it]
 81%|████████▏ | 423/520 [26:21<05:59,  3.70s/it]
                                                 
{'loss': 2.3239, 'grad_norm': 0.0, 'learning_rate': 0.017728860782696667, 'epoch': 0.81}
+
 81%|████████▏ | 423/520 [26:21<05:59,  3.70s/it]
 82%|████████▏ | 424/520 [26:25<05:59,  3.74s/it]
                                                 
{'loss': 1.8431, 'grad_norm': 0.0, 'learning_rate': 0.017376122568400532, 'epoch': 0.82}
+
 82%|████████▏ | 424/520 [26:25<05:59,  3.74s/it]
 82%|████████▏ | 425/520 [26:29<05:56,  3.75s/it]
                                                 
{'loss': 2.0338, 'grad_norm': 0.0, 'learning_rate': 0.017026594633551252, 'epoch': 0.82}
+
 82%|████████▏ | 425/520 [26:29<05:56,  3.75s/it]
 82%|████████▏ | 426/520 [26:33<05:53,  3.77s/it]
                                                 
{'loss': 2.2809, 'grad_norm': 0.0, 'learning_rate': 0.01668029055875512, 'epoch': 0.82}
+
 82%|████████▏ | 426/520 [26:33<05:53,  3.77s/it]
 82%|████████▏ | 427/520 [26:36<05:51,  3.78s/it]
                                                 
{'loss': 1.9615, 'grad_norm': 0.0, 'learning_rate': 0.016337223799358026, 'epoch': 0.82}
+
 82%|████████▏ | 427/520 [26:36<05:51,  3.78s/it]
 82%|████████▏ | 428/520 [26:40<05:43,  3.73s/it]
                                                 
{'loss': 2.179, 'grad_norm': 0.0, 'learning_rate': 0.01599740768492286, 'epoch': 0.82}
+
 82%|████████▏ | 428/520 [26:40<05:43,  3.73s/it]
 82%|████████▎ | 429/520 [26:44<05:36,  3.70s/it]
                                                 
{'loss': 2.1882, 'grad_norm': 0.0, 'learning_rate': 0.015660855418711452, 'epoch': 0.82}
+
 82%|████████▎ | 429/520 [26:44<05:36,  3.70s/it]Token indices sequence length is longer than the specified maximum sequence length for this model (2076 > 2048). Running this sequence through the model will result in indexing errors
+
 83%|████████▎ | 430/520 [26:47<05:31,  3.68s/it]
                                                 
{'loss': 2.0206, 'grad_norm': 0.0, 'learning_rate': 0.015327580077171589, 'epoch': 0.83}
+
 83%|████████▎ | 430/520 [26:47<05:31,  3.68s/it]
 83%|████████▎ | 431/520 [26:51<05:26,  3.67s/it]
                                                 
{'loss': 1.8737, 'grad_norm': 0.0, 'learning_rate': 0.014997594609429088, 'epoch': 0.83}
+
 83%|████████▎ | 431/520 [26:51<05:26,  3.67s/it]
 83%|████████▎ | 432/520 [26:55<05:22,  3.67s/it]
                                                 
{'loss': 2.0893, 'grad_norm': 0.0, 'learning_rate': 0.01467091183678444, 'epoch': 0.83}
+
 83%|████████▎ | 432/520 [26:55<05:22,  3.67s/it]
 83%|████████▎ | 433/520 [26:58<05:19,  3.67s/it]
                                                 
{'loss': 2.1446, 'grad_norm': 0.0, 'learning_rate': 0.014347544452214867, 'epoch': 0.83}
+
 83%|████████▎ | 433/520 [26:58<05:19,  3.67s/it]
 83%|████████▎ | 434/520 [27:02<05:15,  3.67s/it]
                                                 
{'loss': 2.162, 'grad_norm': 0.0, 'learning_rate': 0.014027505019880971, 'epoch': 0.83}
+
 83%|████████▎ | 434/520 [27:02<05:15,  3.67s/it]
 84%|████████▎ | 435/520 [27:06<05:11,  3.66s/it]
                                                 
{'loss': 2.1714, 'grad_norm': 0.0, 'learning_rate': 0.013710805974638696, 'epoch': 0.84}
+
 84%|████████▎ | 435/520 [27:06<05:11,  3.66s/it]
 84%|████████▍ | 436/520 [27:09<05:08,  3.67s/it]
                                                 
{'loss': 2.1073, 'grad_norm': 0.0, 'learning_rate': 0.01339745962155613, 'epoch': 0.84}
+
 84%|████████▍ | 436/520 [27:09<05:08,  3.67s/it]
 84%|████████▍ | 437/520 [27:13<05:04,  3.67s/it]
                                                 
{'loss': 2.1399, 'grad_norm': 0.0, 'learning_rate': 0.01308747813543536, 'epoch': 0.84}
+
 84%|████████▍ | 437/520 [27:13<05:04,  3.67s/it]
 84%|████████▍ | 438/520 [27:17<05:01,  3.67s/it]
                                                 
{'loss': 2.1034, 'grad_norm': 0.0, 'learning_rate': 0.012780873560339467, 'epoch': 0.84}
+
 84%|████████▍ | 438/520 [27:17<05:01,  3.67s/it]
 84%|████████▍ | 439/520 [27:20<04:57,  3.67s/it]
                                                 
{'loss': 1.7664, 'grad_norm': 0.0, 'learning_rate': 0.012477657809124632, 'epoch': 0.84}
+
 84%|████████▍ | 439/520 [27:20<04:57,  3.67s/it]
 85%|████████▍ | 440/520 [27:24<04:52,  3.66s/it]
                                                 
{'loss': 2.0058, 'grad_norm': 0.0, 'learning_rate': 0.012177842662977134, 'epoch': 0.85}
+
 85%|████████▍ | 440/520 [27:24<04:52,  3.66s/it]
 85%|████████▍ | 441/520 [27:28<04:50,  3.68s/it]
                                                 
{'loss': 1.8248, 'grad_norm': 0.0, 'learning_rate': 0.01188143977095576, 'epoch': 0.85}
+
 85%|████████▍ | 441/520 [27:28<04:50,  3.68s/it]
 85%|████████▌ | 442/520 [27:31<04:46,  3.67s/it]
                                                 
{'loss': 2.3179, 'grad_norm': 0.0, 'learning_rate': 0.011588460649539035, 'epoch': 0.85}
+
 85%|████████▌ | 442/520 [27:31<04:46,  3.67s/it]
 85%|████████▌ | 443/520 [27:35<04:42,  3.67s/it]
                                                 
{'loss': 2.0141, 'grad_norm': 0.0, 'learning_rate': 0.011298916682177829, 'epoch': 0.85}
+
 85%|████████▌ | 443/520 [27:35<04:42,  3.67s/it]
 85%|████████▌ | 444/520 [27:39<04:39,  3.67s/it]
                                                 
{'loss': 1.9937, 'grad_norm': 0.0, 'learning_rate': 0.011012819118853146, 'epoch': 0.85}
+
 85%|████████▌ | 444/520 [27:39<04:39,  3.67s/it]
 86%|████████▌ | 445/520 [27:42<04:34,  3.66s/it]
                                                 
{'loss': 1.9637, 'grad_norm': 0.0, 'learning_rate': 0.01073017907563887, 'epoch': 0.86}
+
 86%|████████▌ | 445/520 [27:42<04:34,  3.66s/it]
 86%|████████▌ | 446/520 [27:46<04:30,  3.66s/it]
                                                 
{'loss': 1.8401, 'grad_norm': 0.0, 'learning_rate': 0.010451007534269908, 'epoch': 0.86}
+
 86%|████████▌ | 446/520 [27:46<04:30,  3.66s/it]
 86%|████████▌ | 447/520 [27:50<04:27,  3.66s/it]
                                                 
{'loss': 2.1475, 'grad_norm': 0.0, 'learning_rate': 0.010175315341715598, 'epoch': 0.86}
+
 86%|████████▌ | 447/520 [27:50<04:27,  3.66s/it]
 86%|████████▌ | 448/520 [27:53<04:22,  3.65s/it]
                                                 
{'loss': 2.0884, 'grad_norm': 0.0, 'learning_rate': 0.009903113209758098, 'epoch': 0.86}
+
 86%|████████▌ | 448/520 [27:53<04:22,  3.65s/it]
 86%|████████▋ | 449/520 [27:57<04:20,  3.67s/it]
                                                 
{'loss': 1.9783, 'grad_norm': 0.0, 'learning_rate': 0.009634411714576352, 'epoch': 0.86}
+
 86%|████████▋ | 449/520 [27:57<04:20,  3.67s/it]
 87%|████████▋ | 450/520 [28:01<04:15,  3.66s/it]
                                                 
{'loss': 2.1244, 'grad_norm': 0.0, 'learning_rate': 0.009369221296335007, 'epoch': 0.87}
+
 87%|████████▋ | 450/520 [28:01<04:15,  3.66s/it]
 87%|████████▋ | 451/520 [28:04<04:12,  3.65s/it]
                                                 
{'loss': 2.1608, 'grad_norm': 0.0, 'learning_rate': 0.009107552258778906, 'epoch': 0.87}
+
 87%|████████▋ | 451/520 [28:04<04:12,  3.65s/it]
 87%|████████▋ | 452/520 [28:08<04:08,  3.66s/it]
                                                 
{'loss': 1.8367, 'grad_norm': 0.0, 'learning_rate': 0.008849414768832687, 'epoch': 0.87}
+
 87%|████████▋ | 452/520 [28:08<04:08,  3.66s/it]
 87%|████████▋ | 453/520 [28:12<04:05,  3.66s/it]
                                                 
{'loss': 1.9767, 'grad_norm': 0.0, 'learning_rate': 0.008594818856205699, 'epoch': 0.87}
+
 87%|████████▋ | 453/520 [28:12<04:05,  3.66s/it]
 87%|████████▋ | 454/520 [28:15<04:02,  3.68s/it]
                                                 
{'loss': 2.0911, 'grad_norm': 0.0, 'learning_rate': 0.00834377441300238, 'epoch': 0.87}
+
 87%|████████▋ | 454/520 [28:15<04:02,  3.68s/it]
 88%|████████▊ | 455/520 [28:19<03:59,  3.69s/it]
                                                 
{'loss': 2.0563, 'grad_norm': 0.0, 'learning_rate': 0.008096291193337934, 'epoch': 0.88}
+
 88%|████████▊ | 455/520 [28:19<03:59,  3.69s/it]
 88%|████████▊ | 456/520 [28:23<03:55,  3.68s/it]
                                                 
{'loss': 2.0794, 'grad_norm': 0.0, 'learning_rate': 0.007852378812959226, 'epoch': 0.88}
+
 88%|████████▊ | 456/520 [28:23<03:55,  3.68s/it]
 88%|████████▊ | 457/520 [28:26<03:51,  3.67s/it]
                                                 
{'loss': 1.7164, 'grad_norm': 0.0, 'learning_rate': 0.007612046748871327, 'epoch': 0.88}
+
 88%|████████▊ | 457/520 [28:26<03:51,  3.67s/it]
 88%|████████▊ | 458/520 [28:30<03:47,  3.66s/it]
                                                 
{'loss': 2.2316, 'grad_norm': 0.0, 'learning_rate': 0.007375304338969136, 'epoch': 0.88}
+
 88%|████████▊ | 458/520 [28:30<03:47,  3.66s/it]
 88%|████████▊ | 459/520 [28:34<03:43,  3.66s/it]
                                                 
{'loss': 2.0823, 'grad_norm': 0.0, 'learning_rate': 0.007142160781674645, 'epoch': 0.88}
+
 88%|████████▊ | 459/520 [28:34<03:43,  3.66s/it]
 88%|████████▊ | 460/520 [28:37<03:39,  3.65s/it]
                                                 
{'loss': 2.1604, 'grad_norm': 0.0, 'learning_rate': 0.006912625135579587, 'epoch': 0.88}
+
 88%|████████▊ | 460/520 [28:37<03:39,  3.65s/it]
 89%|████████▊ | 461/520 [28:41<03:35,  3.66s/it]
                                                 
{'loss': 1.499, 'grad_norm': 0.0, 'learning_rate': 0.0066867063190933496, 'epoch': 0.89}
+
 89%|████████▊ | 461/520 [28:41<03:35,  3.66s/it]
 89%|████████▉ | 462/520 [28:45<03:31,  3.64s/it]
                                                 
{'loss': 1.9033, 'grad_norm': 0.0, 'learning_rate': 0.006464413110096601, 'epoch': 0.89}
+
 89%|████████▉ | 462/520 [28:45<03:31,  3.64s/it]
 89%|████████▉ | 463/520 [28:48<03:27,  3.64s/it]
                                                 
{'loss': 2.3309, 'grad_norm': 0.0, 'learning_rate': 0.006245754145600091, 'epoch': 0.89}
+
 89%|████████▉ | 463/520 [28:48<03:27,  3.64s/it]
 89%|████████▉ | 464/520 [28:52<03:24,  3.65s/it]
                                                 
{'loss': 2.1128, 'grad_norm': 0.0, 'learning_rate': 0.006030737921409169, 'epoch': 0.89}
+
 89%|████████▉ | 464/520 [28:52<03:24,  3.65s/it]
 89%|████████▉ | 465/520 [28:55<03:20,  3.65s/it]
                                                 
{'loss': 2.1251, 'grad_norm': 0.0, 'learning_rate': 0.005819372791793654, 'epoch': 0.89}
+
 89%|████████▉ | 465/520 [28:55<03:20,  3.65s/it]
 90%|████████▉ | 466/520 [28:59<03:16,  3.65s/it]
                                                 
{'loss': 1.9657, 'grad_norm': 0.0, 'learning_rate': 0.005611666969163243, 'epoch': 0.9}
+
 90%|████████▉ | 466/520 [28:59<03:16,  3.65s/it]
 90%|████████▉ | 467/520 [29:03<03:13,  3.65s/it]
                                                 
{'loss': 1.8888, 'grad_norm': 0.0, 'learning_rate': 0.005407628523748398, 'epoch': 0.9}
+
 90%|████████▉ | 467/520 [29:03<03:13,  3.65s/it]
 90%|█████████ | 468/520 [29:06<03:09,  3.65s/it]
                                                 
{'loss': 2.2701, 'grad_norm': 0.0, 'learning_rate': 0.00520726538328683, 'epoch': 0.9}
+
 90%|█████████ | 468/520 [29:06<03:09,  3.65s/it]
 90%|█████████ | 469/520 [29:10<03:06,  3.65s/it]
                                                 
{'loss': 2.1097, 'grad_norm': 0.0, 'learning_rate': 0.005010585332715401, 'epoch': 0.9}
+
 90%|█████████ | 469/520 [29:10<03:06,  3.65s/it]
 90%|█████████ | 470/520 [29:14<03:02,  3.66s/it]
                                                 
{'loss': 2.0168, 'grad_norm': 0.0, 'learning_rate': 0.004817596013867765, 'epoch': 0.9}
+
 90%|█████████ | 470/520 [29:14<03:02,  3.66s/it]
 91%|█████████ | 471/520 [29:17<02:59,  3.66s/it]
                                                 
{'loss': 2.2295, 'grad_norm': 0.0, 'learning_rate': 0.004628304925177318, 'epoch': 0.91}
+
 91%|█████████ | 471/520 [29:17<02:59,  3.66s/it]
 91%|█████████ | 472/520 [29:21<02:56,  3.68s/it]
                                                 
{'loss': 2.1844, 'grad_norm': 0.0, 'learning_rate': 0.004442719421385921, 'epoch': 0.91}
+
 91%|█████████ | 472/520 [29:21<02:56,  3.68s/it]
 91%|█████████ | 473/520 [29:25<02:52,  3.67s/it]
                                                 
{'loss': 2.2132, 'grad_norm': 0.0, 'learning_rate': 0.004260846713258193, 'epoch': 0.91}
+
 91%|█████████ | 473/520 [29:25<02:52,  3.67s/it]
 91%|█████████ | 474/520 [29:28<02:48,  3.67s/it]
                                                 
{'loss': 1.9064, 'grad_norm': 0.0, 'learning_rate': 0.004082693867301224, 'epoch': 0.91}
+
 91%|█████████ | 474/520 [29:28<02:48,  3.67s/it]
 91%|█████████▏| 475/520 [29:32<02:45,  3.67s/it]
                                                 
{'loss': 1.8533, 'grad_norm': 0.0, 'learning_rate': 0.003908267805490051, 'epoch': 0.91}
+
 91%|█████████▏| 475/520 [29:32<02:45,  3.67s/it]
 92%|█████████▏| 476/520 [29:36<02:41,  3.68s/it]
                                                 
{'loss': 2.165, 'grad_norm': 0.0, 'learning_rate': 0.003737575304998797, 'epoch': 0.92}
+
 92%|█████████▏| 476/520 [29:36<02:41,  3.68s/it]
 92%|█████████▏| 477/520 [29:39<02:37,  3.67s/it]
                                                 
{'loss': 2.156, 'grad_norm': 0.0, 'learning_rate': 0.003570622997937234, 'epoch': 0.92}
+
 92%|█████████▏| 477/520 [29:39<02:37,  3.67s/it]
 92%|█████████▏| 478/520 [29:43<02:34,  3.68s/it]
                                                 
{'loss': 2.069, 'grad_norm': 0.0, 'learning_rate': 0.00340741737109318, 'epoch': 0.92}
+
 92%|█████████▏| 478/520 [29:43<02:34,  3.68s/it]
 92%|█████████▏| 479/520 [29:47<02:30,  3.68s/it]
                                                 
{'loss': 1.9471, 'grad_norm': 0.0, 'learning_rate': 0.003247964765680389, 'epoch': 0.92}
+
 92%|█████████▏| 479/520 [29:47<02:30,  3.68s/it]
 92%|█████████▏| 480/520 [29:51<02:29,  3.73s/it]
                                                 
{'loss': 1.9655, 'grad_norm': 0.0, 'learning_rate': 0.0030922713770922153, 'epoch': 0.92}
+
 92%|█████████▏| 480/520 [29:51<02:29,  3.73s/it]
 92%|█████████▎| 481/520 [29:55<02:27,  3.78s/it]
                                                 
{'loss': 1.8537, 'grad_norm': 0.0, 'learning_rate': 0.0029403432546609046, 'epoch': 0.93}
+
 92%|█████████▎| 481/520 [29:55<02:27,  3.78s/it]
 93%|█████████▎| 482/520 [29:58<02:22,  3.75s/it]
                                                 
{'loss': 1.8943, 'grad_norm': 0.0, 'learning_rate': 0.0027921863014225504, 'epoch': 0.93}
+
 93%|█████████▎| 482/520 [29:58<02:22,  3.75s/it]
 93%|█████████▎| 483/520 [30:02<02:18,  3.74s/it]
                                                 
{'loss': 2.0377, 'grad_norm': 0.0, 'learning_rate': 0.002647806273887665, 'epoch': 0.93}
+
 93%|█████████▎| 483/520 [30:02<02:18,  3.74s/it]
 93%|█████████▎| 484/520 [30:06<02:14,  3.72s/it]
                                                 
{'loss': 2.1901, 'grad_norm': 0.0, 'learning_rate': 0.0025072087818176383, 'epoch': 0.93}
+
 93%|█████████▎| 484/520 [30:06<02:14,  3.72s/it]
 93%|█████████▎| 485/520 [30:09<02:09,  3.71s/it]
                                                 
{'loss': 2.0144, 'grad_norm': 0.0, 'learning_rate': 0.002370399288006664, 'epoch': 0.93}
+
 93%|█████████▎| 485/520 [30:09<02:09,  3.71s/it]
 93%|█████████▎| 486/520 [30:13<02:05,  3.70s/it]
                                                 
{'loss': 2.0347, 'grad_norm': 0.0, 'learning_rate': 0.0022373831080695463, 'epoch': 0.93}
+
 93%|█████████▎| 486/520 [30:13<02:05,  3.70s/it]
 94%|█████████▎| 487/520 [30:17<02:01,  3.69s/it]
                                                 
{'loss': 2.0192, 'grad_norm': 0.0, 'learning_rate': 0.0021081654102351635, 'epoch': 0.94}
+
 94%|█████████▎| 487/520 [30:17<02:01,  3.69s/it]
 94%|█████████▍| 488/520 [30:20<01:57,  3.68s/it]
                                                 
{'loss': 2.1552, 'grad_norm': 0.0, 'learning_rate': 0.0019827512151456175, 'epoch': 0.94}
+
 94%|█████████▍| 488/520 [30:20<01:57,  3.68s/it]
 94%|█████████▍| 489/520 [30:24<01:54,  3.68s/it]
                                                 
{'loss': 1.7671, 'grad_norm': 0.0, 'learning_rate': 0.0018611453956612345, 'epoch': 0.94}
+
 94%|█████████▍| 489/520 [30:24<01:54,  3.68s/it]
 94%|█████████▍| 490/520 [30:28<01:50,  3.69s/it]
                                                 
{'loss': 2.1193, 'grad_norm': 0.0, 'learning_rate': 0.0017433526766711727, 'epoch': 0.94}
+
 94%|█████████▍| 490/520 [30:28<01:50,  3.69s/it]
 94%|█████████▍| 491/520 [30:31<01:47,  3.70s/it]
                                                 
{'loss': 2.1849, 'grad_norm': 0.0, 'learning_rate': 0.0016293776349098677, 'epoch': 0.94}
+
 94%|█████████▍| 491/520 [30:31<01:47,  3.70s/it]
 95%|█████████▍| 492/520 [30:35<01:43,  3.69s/it]
                                                 
{'loss': 2.1631, 'grad_norm': 0.0, 'learning_rate': 0.0015192246987791981, 'epoch': 0.95}
+
 95%|█████████▍| 492/520 [30:35<01:43,  3.69s/it]
 95%|█████████▍| 493/520 [30:39<01:39,  3.68s/it]
                                                 
{'loss': 1.9782, 'grad_norm': 0.0, 'learning_rate': 0.0014128981481764114, 'epoch': 0.95}
+
 95%|█████████▍| 493/520 [30:39<01:39,  3.68s/it]
 95%|█████████▌| 494/520 [30:42<01:35,  3.67s/it]
                                                 
{'loss': 2.001, 'grad_norm': 0.0, 'learning_rate': 0.0013104021143278911, 'epoch': 0.95}
+
 95%|█████████▌| 494/520 [30:42<01:35,  3.67s/it]
 95%|█████████▌| 495/520 [30:46<01:32,  3.69s/it]
                                                 
{'loss': 2.0557, 'grad_norm': 0.0, 'learning_rate': 0.0012117405796285285, 'epoch': 0.95}
+
 95%|█████████▌| 495/520 [30:46<01:32,  3.69s/it]
 95%|█████████▌| 496/520 [30:50<01:28,  3.68s/it]
                                                 
{'loss': 2.1532, 'grad_norm': 0.0, 'learning_rate': 0.0011169173774871477, 'epoch': 0.95}
+
 95%|█████████▌| 496/520 [30:50<01:28,  3.68s/it]
 96%|█████████▌| 497/520 [30:54<01:24,  3.68s/it]
                                                 
{'loss': 1.8146, 'grad_norm': 0.0, 'learning_rate': 0.0010259361921774012, 'epoch': 0.96}
+
 96%|█████████▌| 497/520 [30:54<01:24,  3.68s/it]
 96%|█████████▌| 498/520 [30:57<01:20,  3.67s/it]
                                                 
{'loss': 2.109, 'grad_norm': 0.0, 'learning_rate': 0.000938800558694719, 'epoch': 0.96}
+
 96%|█████████▌| 498/520 [30:57<01:20,  3.67s/it]
 96%|█████████▌| 499/520 [31:01<01:16,  3.66s/it]
                                                 
{'loss': 2.0238, 'grad_norm': 0.0, 'learning_rate': 0.0008555138626189618, 'epoch': 0.96}
+
 96%|█████████▌| 499/520 [31:01<01:16,  3.66s/it]
 96%|█████████▌| 500/520 [31:04<01:13,  3.66s/it]
                                                 
{'loss': 2.3196, 'grad_norm': 0.0, 'learning_rate': 0.0007760793399827937, 'epoch': 0.96}
+
 96%|█████████▌| 500/520 [31:04<01:13,  3.66s/it]
 96%|█████████▋| 501/520 [31:08<01:09,  3.66s/it]
                                                 
{'loss': 2.0128, 'grad_norm': 0.0, 'learning_rate': 0.000700500077146038, 'epoch': 0.96}
+
 96%|█████████▋| 501/520 [31:08<01:09,  3.66s/it]
 97%|█████████▋| 502/520 [31:12<01:06,  3.70s/it]
                                                 
{'loss': 2.0904, 'grad_norm': 0.0, 'learning_rate': 0.0006287790106757397, 'epoch': 0.97}
+
 97%|█████████▋| 502/520 [31:12<01:06,  3.70s/it]
 97%|█████████▋| 503/520 [31:16<01:02,  3.70s/it]
                                                 
{'loss': 1.888, 'grad_norm': 0.0, 'learning_rate': 0.0005609189272320237, 'epoch': 0.97}
+
 97%|█████████▋| 503/520 [31:16<01:02,  3.70s/it]
 97%|█████████▋| 504/520 [31:19<00:59,  3.70s/it]
                                                 
{'loss': 2.3021, 'grad_norm': 0.0, 'learning_rate': 0.000496922463459859, 'epoch': 0.97}
+
 97%|█████████▋| 504/520 [31:19<00:59,  3.70s/it]
 97%|█████████▋| 505/520 [31:23<00:55,  3.73s/it]
                                                 
{'loss': 2.1435, 'grad_norm': 0.0, 'learning_rate': 0.0004367921058866187, 'epoch': 0.97}
+
 97%|█████████▋| 505/520 [31:23<00:55,  3.73s/it]
 97%|█████████▋| 506/520 [31:27<00:53,  3.79s/it]
                                                 
{'loss': 2.1081, 'grad_norm': 0.0, 'learning_rate': 0.0003805301908254455, 'epoch': 0.97}
+
 97%|█████████▋| 506/520 [31:27<00:53,  3.79s/it]
 98%|█████████▊| 507/520 [31:31<00:49,  3.79s/it]
                                                 
{'loss': 1.9223, 'grad_norm': 0.0, 'learning_rate': 0.0003281389042844918, 'epoch': 0.97}
+
 98%|█████████▊| 507/520 [31:31<00:49,  3.79s/it]
 98%|█████████▊| 508/520 [31:34<00:44,  3.75s/it]
                                                 
{'loss': 2.1602, 'grad_norm': 0.0, 'learning_rate': 0.00027962028188198705, 'epoch': 0.98}
+
 98%|█████████▊| 508/520 [31:34<00:44,  3.75s/it]
 98%|█████████▊| 509/520 [31:38<00:41,  3.73s/it]
                                                 
{'loss': 2.0725, 'grad_norm': 0.0, 'learning_rate': 0.00023497620876711256, 'epoch': 0.98}
+
 98%|█████████▊| 509/520 [31:38<00:41,  3.73s/it]
 98%|█████████▊| 510/520 [31:42<00:37,  3.71s/it]
                                                 
{'loss': 2.1279, 'grad_norm': 0.0, 'learning_rate': 0.00019420841954681523, 'epoch': 0.98}
+
 98%|█████████▊| 510/520 [31:42<00:37,  3.71s/it]
 98%|█████████▊| 511/520 [31:46<00:33,  3.73s/it]
                                                 
{'loss': 2.0474, 'grad_norm': 0.0, 'learning_rate': 0.00015731849821833956, 'epoch': 0.98}
+
 98%|█████████▊| 511/520 [31:46<00:33,  3.73s/it]
 98%|█████████▊| 512/520 [31:49<00:29,  3.75s/it]
                                                 
{'loss': 2.0024, 'grad_norm': 0.0, 'learning_rate': 0.00012430787810776555, 'epoch': 0.98}
+
 98%|█████████▊| 512/520 [31:49<00:29,  3.75s/it]
 99%|█████████▊| 513/520 [31:53<00:26,  3.72s/it]
                                                 
{'loss': 2.1995, 'grad_norm': 0.0, 'learning_rate': 9.517784181422018e-05, 'epoch': 0.99}
+
 99%|█████████▊| 513/520 [31:53<00:26,  3.72s/it]
 99%|█████████▉| 514/520 [31:57<00:22,  3.70s/it]
                                                 
{'loss': 2.0174, 'grad_norm': 0.0, 'learning_rate': 6.992952116013917e-05, 'epoch': 0.99}
+
 99%|█████████▉| 514/520 [31:57<00:22,  3.70s/it]
 99%|█████████▉| 515/520 [32:00<00:18,  3.69s/it]
                                                 
{'loss': 2.2217, 'grad_norm': 0.0, 'learning_rate': 4.856389714723575e-05, 'epoch': 0.99}
+
 99%|█████████▉| 515/520 [32:00<00:18,  3.69s/it]
 99%|█████████▉| 516/520 [32:04<00:14,  3.70s/it]
                                                 
{'loss': 2.1999, 'grad_norm': 0.0, 'learning_rate': 3.108179991837545e-05, 'epoch': 0.99}
+
 99%|█████████▉| 516/520 [32:04<00:14,  3.70s/it]
 99%|█████████▉| 517/520 [32:08<00:11,  3.67s/it]
                                                 
{'loss': 1.8719, 'grad_norm': 0.0, 'learning_rate': 1.7483908725357544e-05, 'epoch': 0.99}
+
 99%|█████████▉| 517/520 [32:08<00:11,  3.67s/it]
100%|█████████▉| 518/520 [32:11<00:07,  3.65s/it]
                                                 
{'loss': 2.1615, 'grad_norm': 0.0, 'learning_rate': 7.770751902513862e-06, 'epoch': 1.0}
+
100%|█████████▉| 518/520 [32:11<00:07,  3.65s/it]
100%|█████████▉| 519/520 [32:15<00:03,  3.65s/it]
                                                 
{'loss': 1.9508, 'grad_norm': 0.0, 'learning_rate': 1.9427068461808086e-06, 'epoch': 1.0}
+
100%|█████████▉| 519/520 [32:15<00:03,  3.65s/it]
100%|██████████| 520/520 [32:19<00:00,  3.92s/it]
                                                 
{'loss': 1.7443, 'grad_norm': 0.0, 'learning_rate': 0.0, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:20<00:00,  3.92s/it]
                                                 
{'train_runtime': 1940.0192, 'train_samples_per_second': 34.293, 'train_steps_per_second': 0.268, 'train_loss': 2.066992656771953, 'epoch': 1.0}
+
100%|██████████| 520/520 [32:20<00:00,  3.92s/it]
100%|██████████| 520/520 [32:20<00:00,  3.73s/it]
+[2025-10-16 19:00:36,426] [INFO] [launch.py:348:main] Process 2589806 exits successfully.
+[2025-10-16 19:00:36,427] [INFO] [launch.py:348:main] Process 2589804 exits successfully.
+[2025-10-16 19:00:36,427] [INFO] [launch.py:348:main] Process 2589808 exits successfully.
+[2025-10-16 19:00:36,428] [INFO] [launch.py:348:main] Process 2589807 exits successfully.
+[2025-10-16 19:00:37,429] [INFO] [launch.py:348:main] Process 2589802 exits successfully.
+[2025-10-16 19:00:37,430] [INFO] [launch.py:348:main] Process 2589803 exits successfully.
+[2025-10-16 19:00:37,430] [INFO] [launch.py:348:main] Process 2589805 exits successfully.
+[2025-10-16 19:00:40,434] [INFO] [launch.py:348:main] Process 2589801 exits successfully.
+==== EXPERIMENT COMPLETED: qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_0.9_2e-1_connector-7.0_0.9_2e-1_ablation_20251016_180316.log
+Timestamp: 2025-10-16 19:00:43
+=====================================
diff --git a/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation_20251016_190043.log b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation_20251016_190043.log
new file mode 100644
index 0000000000000000000000000000000000000000..c6109ce1e5877605c173ef278a660c564c88065f
--- /dev/null
+++ b/logs_oct13/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation_20251016_190043.log
@@ -0,0 +1,1800 @@
+==== STARTING EXPERIMENT: qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation ====
+Log File: qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation_20251016_190043.log
+Timestamp: 2025-10-16 19:00:43
+=====================================
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 19:00:45,737] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:48,659] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-10-16 19:00:48,661] [INFO] [runner.py:568:main] cmd = /opt/conda/envs/tinyllava/bin/python3.10 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None tinyllava/train/train.py --deepspeed ./scripts/zero3.json --data_path /s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json --image_folder /s3-code/ywang29/datasets/tinyllava --is_multimodal True --conv_version qwen2_base --model_name_or_path Qwen/Qwen2.5-0.5B --vision_tower google/siglip-so400m-patch14-384 --vision_tower2  --connector_type mlp2x_gelu --mm_vision_select_layer -2 --image_aspect_ratio square --attn_implementation flash_attention_2 --bf16 True --training_recipe common --tune_type_llm full --tune_type_vision_tower frozen --tune_vision_tower_from_layer 0 --tune_type_connector full --group_by_modality_length True --pretrained_model_path /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain --output_dir /nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 4 --evaluation_strategy no --learning_rate 2e-1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type cosine --logging_steps 1 --tf32 False --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 8 --lazy_preprocess True --report_to tensorboard --tokenizer_use_fast False --run_name tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune --subnet_mode_text both --subnet_type_text None --mask_type_text soft --init_mean_text 7.0 --temperature_attn_text 1.1 --temperature_mlp_text 1.1 --backward_type_text normal --masked_layers_text all --subnet_mode_vision both --subnet_type_vision None --mask_type_vision soft --init_mean_vision 7.0 --temperature_attn_vision 1.1 --temperature_mlp_vision 1.1 --backward_type_vision normal --masked_layers_vision all --subnet_type_connector global --mask_type_connector soft --init_mean_connector 7.0 --temperature_connector 1.1 --backward_type_connector normal --mm_projector_lr 2e-1 --seed 42 --mask_model llm-connector --save_strategy steps --save_steps 50000 --save_total_limit 1 --train_data_ratio 0.1
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 19:00:51,236] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:52,295] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.21.5
+[2025-10-16 19:00:52,295] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth
+[2025-10-16 19:00:52,296] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-10-16 19:00:52,296] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-10-16 19:00:52,296] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-10-16 19:00:52,296] [INFO] [launch.py:163:main] dist_world_size=8
+[2025-10-16 19:00:52,296] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-10-16 19:00:52,298] [INFO] [launch.py:253:main] process 2622013 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=0', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,300] [INFO] [launch.py:253:main] process 2622014 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=1', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,302] [INFO] [launch.py:253:main] process 2622015 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=2', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,304] [INFO] [launch.py:253:main] process 2622016 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=3', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,306] [INFO] [launch.py:253:main] process 2622017 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=4', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,307] [INFO] [launch.py:253:main] process 2622018 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=5', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,309] [INFO] [launch.py:253:main] process 2622019 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=6', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+[2025-10-16 19:00:52,311] [INFO] [launch.py:253:main] process 2622020 spawned with command: ['/opt/conda/envs/tinyllava/bin/python3.10', '-u', 'tinyllava/train/train.py', '--local_rank=7', '--deepspeed', './scripts/zero3.json', '--data_path', '/s3-code/ywang29/datasets/tinyllava/text_files/llava_v1_5_mix665k.json', '--image_folder', '/s3-code/ywang29/datasets/tinyllava', '--is_multimodal', 'True', '--conv_version', 'qwen2_base', '--model_name_or_path', 'Qwen/Qwen2.5-0.5B', '--vision_tower', 'google/siglip-so400m-patch14-384', '--vision_tower2', '', '--connector_type', 'mlp2x_gelu', '--mm_vision_select_layer', '-2', '--image_aspect_ratio', 'square', '--attn_implementation', 'flash_attention_2', '--bf16', 'True', '--training_recipe', 'common', '--tune_type_llm', 'full', '--tune_type_vision_tower', 'frozen', '--tune_vision_tower_from_layer', '0', '--tune_type_connector', 'full', '--group_by_modality_length', 'True', '--pretrained_model_path', '/nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain', '--output_dir', '/nfs/ywang29/TinyLLaVA/checkpoints/qwen2.5-0_5b_base_masktune_42_llm-connector_text-7.0_1.1_2e-1_connector-7.0_1.1_2e-1_ablation', '--num_train_epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '4', '--evaluation_strategy', 'no', '--learning_rate', '2e-1', '--weight_decay', '0.', '--warmup_ratio', '0.03', '--lr_scheduler_type', 'cosine', '--logging_steps', '1', '--tf32', 'False', '--model_max_length', '2048', '--gradient_checkpointing', 'True', '--dataloader_num_workers', '8', '--lazy_preprocess', 'True', '--report_to', 'tensorboard', '--tokenizer_use_fast', 'False', '--run_name', 'tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-qwen2.5-0_5b_base-masktune', '--subnet_mode_text', 'both', '--subnet_type_text', 'None', '--mask_type_text', 'soft', '--init_mean_text', '7.0', '--temperature_attn_text', '1.1', '--temperature_mlp_text', '1.1', '--backward_type_text', 'normal', '--masked_layers_text', 'all', '--subnet_mode_vision', 'both', '--subnet_type_vision', 'None', '--mask_type_vision', 'soft', '--init_mean_vision', '7.0', '--temperature_attn_vision', '1.1', '--temperature_mlp_vision', '1.1', '--backward_type_vision', 'normal', '--masked_layers_vision', 'all', '--subnet_type_connector', 'global', '--mask_type_connector', 'soft', '--init_mean_connector', '7.0', '--temperature_connector', '1.1', '--backward_type_connector', 'normal', '--mm_projector_lr', '2e-1', '--seed', '42', '--mask_model', 'llm-connector', '--save_strategy', 'steps', '--save_steps', '50000', '--save_total_limit', '1', '--train_data_ratio', '0.1']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/cuda/__init__.py:51: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
+  import pynvml  # type: ignore[import]
+[2025-10-16 19:00:58,968] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,188] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,372] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,404] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,405] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,440] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,446] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,499] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,499] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-10-16 19:00:59,584] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,807] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,808] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,845] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,845] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,895] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,895] [INFO] [comm.py:637:init_distributed] cdb=None
+[2025-10-16 19:00:59,895] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+{'llm': {'model_name_or_path': 'Qwen/Qwen2.5-0.5B', 'cache_dir': None, 'attn_implementation': 'flash_attention_2', 'subnet_mode': 'both', 'subnet_type': 'None', 'sparsity_attn': None, 'sparsity_mlp': None, 'threshold_attn': None, 'threshold_mlp': None, 'temperature_attn': 1.1, 'temperature_mlp': 1.1, 'masked_layers': 'all', 'mask_type': 'soft', 'backward_type': 'normal'}, 'vision_tower': {'model_name_or_path': 'google/siglip-so400m-patch14-384'}, 'connector': {'connector_type': 'mlp2x_gelu', 'subnet_type': 'global', 'threshold': None, 'sparsity': None, 'temperature': 1.1, 'mask_type': 'soft', 'backward_type': 'normal'}}
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+Apply masks for the following modules:  ['llm', 'connector']
+Apply masks for the following modules:  ['llm', 'connector']
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/huggingface_hub/file_download.py:945: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
+  warnings.warn(
+TinyLlavaConfig {
+  "backward_type_connector": "normal",
+  "cache_dir": null,
+  "connector_type": "mlp2x_gelu",
+  "hidden_size": 896,
+  "ignore_index": -100,
+  "image_aspect_ratio": "square",
+  "image_token_index": -200,
+  "llm_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "mask_model": [
+    "llm",
+    "connector"
+  ],
+  "mask_type_connector": "soft",
+  "model_type": "tinyllava",
+  "num_queries": 128,
+  "num_resampler_layers": 3,
+  "pad_token": null,
+  "resampler_hidden_size": 768,
+  "sparsity_connector": null,
+  "subnet_type_connector": "global",
+  "temperature_connector": 1.1,
+  "text_config": {
+    "_name_or_path": "Qwen/Qwen2.5-0.5B",
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "backward_type": "normal",
+    "bos_token_id": 151643,
+    "eos_token_id": 151643,
+    "hidden_size": 896,
+    "intermediate_size": 4864,
+    "mask_type": "soft",
+    "masked_layers": "all",
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "model_type": "qwen2",
+    "num_attention_heads": 14,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "subnet_mode": "both",
+    "subnet_type": "None",
+    "temperature_attn": 1.1,
+    "temperature_mlp": 1.1,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_mrope": false,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "threshold_connector": null,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "tokenizer_padding_side": "right",
+  "tokenizer_use_fast": false,
+  "transformers_version": "4.40.1",
+  "tune_type_connector": "frozen",
+  "tune_type_llm": "frozen",
+  "tune_type_vision_tower": "frozen",
+  "tune_vision_tower_from_layer": -1,
+  "use_cache": false,
+  "vision_config": {
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_name_or_path": "google/siglip-so400m-patch14-384",
+    "model_name_or_path2": "",
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "patch",
+  "vision_hidden_size": 1152,
+  "vision_model_name_or_path": "google/siglip-so400m-patch14-384",
+  "vision_model_name_or_path2": "",
+  "vocab_size": 151936
+}
+
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2622013:2622013 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622013:2622013 [0] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622013:2622013 [0] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622013:2622013 [0] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622013:2622013 [0] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622013:2622013 [0] NCCL INFO cudaDriverVersion 12040
+NCCL version 2.21.5+cuda12.1
+ywang29-vrdb-test2-worker-0:2622015:2622015 [2] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622015:2622015 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622015:2622015 [2] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622015:2622015 [2] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622015:2622015 [2] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622015:2622015 [2] NCCL INFO NET/Plugin: Using internal network plugin.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Using network Socket
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2622014:2622014 [1] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622014:2622014 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622014:2622014 [1] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622014:2622014 [1] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622014:2622014 [1] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622014:2622014 [1] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2622020:2622020 [7] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622020:2622020 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622020:2622020 [7] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622020:2622020 [7] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622020:2622020 [7] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622020:2622020 [7] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622017:2622017 [4] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622017:2622017 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622017:2622017 [4] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622017:2622017 [4] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622017:2622017 [4] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622017:2622017 [4] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Using network Socket
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2622016:2622016 [3] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622016:2622016 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622016:2622016 [3] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622016:2622016 [3] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622016:2622016 [3] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622016:2622016 [3] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Using network Socket
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2622019:2622019 [6] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622019:2622019 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622019:2622019 [6] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622019:2622019 [6] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622019:2622019 [6] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622019:2622019 [6] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Using network Socket
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
+  return self.fget.__get__(instance, owner)()
+You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
+ywang29-vrdb-test2-worker-0:2622018:2622018 [5] NCCL INFO cudaDriverVersion 12040
+ywang29-vrdb-test2-worker-0:2622018:2622018 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622018:2622018 [5] NCCL INFO Bootstrap : Using eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622018:2622018 [5] NCCL INFO NET/Plugin: No plugin found (libnccl-net.so)
+ywang29-vrdb-test2-worker-0:2622018:2622018 [5] NCCL INFO NET/Plugin: Plugin load returned 2 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-net.so
+ywang29-vrdb-test2-worker-0:2622018:2622018 [5] NCCL INFO NET/Plugin: Using internal network plugin.
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO NET/IB : No device found.
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO NET/Socket : Using [0]eth0:10.200.152.48<0>
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO ncclCommInitRank comm 0x55a299b42340 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO ncclCommInitRank comm 0x55e594abd980 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO ncclCommInitRank comm 0x563427ea1a60 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO ncclCommInitRank comm 0x564ac45148e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO ncclCommInitRank comm 0x55cb7c2618c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO ncclCommInitRank comm 0x5601890bd100 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO ncclCommInitRank comm 0x55fa6fbd3730 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO ncclCommInitRank comm 0x55d0b3c95680 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xbf3ea31acac42f24 - Init START
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO comm 0x5601890bd100 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO comm 0x563427ea1a60 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO comm 0x55d0b3c95680 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO comm 0x564ac45148e0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO comm 0x55fa6fbd3730 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO comm 0x55a299b42340 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO comm 0x55e594abd980 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO comm 0x55cb7c2618c0 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622018:2623654 [5] NCCL INFO ncclCommInitRank comm 0x563427ea1a60 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622020:2623617 [7] NCCL INFO ncclCommInitRank comm 0x55cb7c2618c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO TUNER/Plugin: Plugin load returned 11 : libnccl-net.so: cannot open shared object file: No such file or directory : when loading libnccl-tuner.so
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622013:2623614 [0] NCCL INFO ncclCommInitRank comm 0x55e594abd980 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622014:2623616 [1] NCCL INFO ncclCommInitRank comm 0x55a299b42340 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622016:2623619 [3] NCCL INFO ncclCommInitRank comm 0x55d0b3c95680 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO TUNER/Plugin: Using internal tuner plugin.
+ywang29-vrdb-test2-worker-0:2622015:2623615 [2] NCCL INFO ncclCommInitRank comm 0x564ac45148e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622019:2623636 [6] NCCL INFO ncclCommInitRank comm 0x5601890bd100 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622017:2623618 [4] NCCL INFO ncclCommInitRank comm 0x55fa6fbd3730 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0xbf3ea31acac42f24 - Init COMPLETE
+[2025-10-16 19:01:40,186] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 459, num_elems = 0.99B
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint at /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model and are newly initialized: ['model.layers.0.mlp.down_proj.scores', 'model.layers.0.mlp.gate_proj.scores', 'model.layers.0.mlp.up_proj.scores', 'model.layers.0.self_attn.k_proj.scores', 'model.layers.0.self_attn.o_proj.scores', 'model.layers.0.self_attn.q_proj.scores', 'model.layers.0.self_attn.v_proj.scores', 'model.layers.1.mlp.down_proj.scores', 'model.layers.1.mlp.gate_proj.scores', 'model.layers.1.mlp.up_proj.scores', 'model.layers.1.self_attn.k_proj.scores', 'model.layers.1.self_attn.o_proj.scores', 'model.layers.1.self_attn.q_proj.scores', 'model.layers.1.self_attn.v_proj.scores', 'model.layers.10.mlp.down_proj.scores', 'model.layers.10.mlp.gate_proj.scores', 'model.layers.10.mlp.up_proj.scores', 'model.layers.10.self_attn.k_proj.scores', 'model.layers.10.self_attn.o_proj.scores', 'model.layers.10.self_attn.q_proj.scores', 'model.layers.10.self_attn.v_proj.scores', 'model.layers.11.mlp.down_proj.scores', 'model.layers.11.mlp.gate_proj.scores', 'model.layers.11.mlp.up_proj.scores', 'model.layers.11.self_attn.k_proj.scores', 'model.layers.11.self_attn.o_proj.scores', 'model.layers.11.self_attn.q_proj.scores', 'model.layers.11.self_attn.v_proj.scores', 'model.layers.12.mlp.down_proj.scores', 'model.layers.12.mlp.gate_proj.scores', 'model.layers.12.mlp.up_proj.scores', 'model.layers.12.self_attn.k_proj.scores', 'model.layers.12.self_attn.o_proj.scores', 'model.layers.12.self_attn.q_proj.scores', 'model.layers.12.self_attn.v_proj.scores', 'model.layers.13.mlp.down_proj.scores', 'model.layers.13.mlp.gate_proj.scores', 'model.layers.13.mlp.up_proj.scores', 'model.layers.13.self_attn.k_proj.scores', 'model.layers.13.self_attn.o_proj.scores', 'model.layers.13.self_attn.q_proj.scores', 'model.layers.13.self_attn.v_proj.scores', 'model.layers.14.mlp.down_proj.scores', 'model.layers.14.mlp.gate_proj.scores', 'model.layers.14.mlp.up_proj.scores', 'model.layers.14.self_attn.k_proj.scores', 'model.layers.14.self_attn.o_proj.scores', 'model.layers.14.self_attn.q_proj.scores', 'model.layers.14.self_attn.v_proj.scores', 'model.layers.15.mlp.down_proj.scores', 'model.layers.15.mlp.gate_proj.scores', 'model.layers.15.mlp.up_proj.scores', 'model.layers.15.self_attn.k_proj.scores', 'model.layers.15.self_attn.o_proj.scores', 'model.layers.15.self_attn.q_proj.scores', 'model.layers.15.self_attn.v_proj.scores', 'model.layers.16.mlp.down_proj.scores', 'model.layers.16.mlp.gate_proj.scores', 'model.layers.16.mlp.up_proj.scores', 'model.layers.16.self_attn.k_proj.scores', 'model.layers.16.self_attn.o_proj.scores', 'model.layers.16.self_attn.q_proj.scores', 'model.layers.16.self_attn.v_proj.scores', 'model.layers.17.mlp.down_proj.scores', 'model.layers.17.mlp.gate_proj.scores', 'model.layers.17.mlp.up_proj.scores', 'model.layers.17.self_attn.k_proj.scores', 'model.layers.17.self_attn.o_proj.scores', 'model.layers.17.self_attn.q_proj.scores', 'model.layers.17.self_attn.v_proj.scores', 'model.layers.18.mlp.down_proj.scores', 'model.layers.18.mlp.gate_proj.scores', 'model.layers.18.mlp.up_proj.scores', 'model.layers.18.self_attn.k_proj.scores', 'model.layers.18.self_attn.o_proj.scores', 'model.layers.18.self_attn.q_proj.scores', 'model.layers.18.self_attn.v_proj.scores', 'model.layers.19.mlp.down_proj.scores', 'model.layers.19.mlp.gate_proj.scores', 'model.layers.19.mlp.up_proj.scores', 'model.layers.19.self_attn.k_proj.scores', 'model.layers.19.self_attn.o_proj.scores', 'model.layers.19.self_attn.q_proj.scores', 'model.layers.19.self_attn.v_proj.scores', 'model.layers.2.mlp.down_proj.scores', 'model.layers.2.mlp.gate_proj.scores', 'model.layers.2.mlp.up_proj.scores', 'model.layers.2.self_attn.k_proj.scores', 'model.layers.2.self_attn.o_proj.scores', 'model.layers.2.self_attn.q_proj.scores', 'model.layers.2.self_attn.v_proj.scores', 'model.layers.20.mlp.down_proj.scores', 'model.layers.20.mlp.gate_proj.scores', 'model.layers.20.mlp.up_proj.scores', 'model.layers.20.self_attn.k_proj.scores', 'model.layers.20.self_attn.o_proj.scores', 'model.layers.20.self_attn.q_proj.scores', 'model.layers.20.self_attn.v_proj.scores', 'model.layers.21.mlp.down_proj.scores', 'model.layers.21.mlp.gate_proj.scores', 'model.layers.21.mlp.up_proj.scores', 'model.layers.21.self_attn.k_proj.scores', 'model.layers.21.self_attn.o_proj.scores', 'model.layers.21.self_attn.q_proj.scores', 'model.layers.21.self_attn.v_proj.scores', 'model.layers.22.mlp.down_proj.scores', 'model.layers.22.mlp.gate_proj.scores', 'model.layers.22.mlp.up_proj.scores', 'model.layers.22.self_attn.k_proj.scores', 'model.layers.22.self_attn.o_proj.scores', 'model.layers.22.self_attn.q_proj.scores', 'model.layers.22.self_attn.v_proj.scores', 'model.layers.23.mlp.down_proj.scores', 'model.layers.23.mlp.gate_proj.scores', 'model.layers.23.mlp.up_proj.scores', 'model.layers.23.self_attn.k_proj.scores', 'model.layers.23.self_attn.o_proj.scores', 'model.layers.23.self_attn.q_proj.scores', 'model.layers.23.self_attn.v_proj.scores', 'model.layers.3.mlp.down_proj.scores', 'model.layers.3.mlp.gate_proj.scores', 'model.layers.3.mlp.up_proj.scores', 'model.layers.3.self_attn.k_proj.scores', 'model.layers.3.self_attn.o_proj.scores', 'model.layers.3.self_attn.q_proj.scores', 'model.layers.3.self_attn.v_proj.scores', 'model.layers.4.mlp.down_proj.scores', 'model.layers.4.mlp.gate_proj.scores', 'model.layers.4.mlp.up_proj.scores', 'model.layers.4.self_attn.k_proj.scores', 'model.layers.4.self_attn.o_proj.scores', 'model.layers.4.self_attn.q_proj.scores', 'model.layers.4.self_attn.v_proj.scores', 'model.layers.5.mlp.down_proj.scores', 'model.layers.5.mlp.gate_proj.scores', 'model.layers.5.mlp.up_proj.scores', 'model.layers.5.self_attn.k_proj.scores', 'model.layers.5.self_attn.o_proj.scores', 'model.layers.5.self_attn.q_proj.scores', 'model.layers.5.self_attn.v_proj.scores', 'model.layers.6.mlp.down_proj.scores', 'model.layers.6.mlp.gate_proj.scores', 'model.layers.6.mlp.up_proj.scores', 'model.layers.6.self_attn.k_proj.scores', 'model.layers.6.self_attn.o_proj.scores', 'model.layers.6.self_attn.q_proj.scores', 'model.layers.6.self_attn.v_proj.scores', 'model.layers.7.mlp.down_proj.scores', 'model.layers.7.mlp.gate_proj.scores', 'model.layers.7.mlp.up_proj.scores', 'model.layers.7.self_attn.k_proj.scores', 'model.layers.7.self_attn.o_proj.scores', 'model.layers.7.self_attn.q_proj.scores', 'model.layers.7.self_attn.v_proj.scores', 'model.layers.8.mlp.down_proj.scores', 'model.layers.8.mlp.gate_proj.scores', 'model.layers.8.mlp.up_proj.scores', 'model.layers.8.self_attn.k_proj.scores', 'model.layers.8.self_attn.o_proj.scores', 'model.layers.8.self_attn.q_proj.scores', 'model.layers.8.self_attn.v_proj.scores', 'model.layers.9.mlp.down_proj.scores', 'model.layers.9.mlp.gate_proj.scores', 'model.layers.9.mlp.up_proj.scores', 'model.layers.9.self_attn.k_proj.scores', 'model.layers.9.self_attn.o_proj.scores', 'model.layers.9.self_attn.q_proj.scores', 'model.layers.9.self_attn.v_proj.scores']
+You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+loading language model from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/language_model
+[2025-10-16 19:01:42,047] [INFO] [partition_parameters.py:343:__exit__] finished initializing model - num_params = 907, num_elems = 1.42B
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+Loading vision tower from  /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/vision_tower
+Loading connector from /nfs/ywang29/TinyLLaVA/checkpoints/tiny-llava-Qwen2.5-0.5B-siglip-so400m-patch14-384-pretrain/connector/pytorch_model.bin...
+TinyLlavaForConditionalGeneration(
+  (language_model): Qwen2ForCausalLM(
+    (model): Qwen2Model(
+      (embed_tokens): Embedding(151936, 896)
+      (layers): ModuleList(
+        (0-23): 24 x Qwen2DecoderLayer(
+          (self_attn): Qwen2FlashAttention2(
+            (q_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+            (k_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (v_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=128, bias=True)
+            (o_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=False)
+            (rotary_emb): Qwen2RotaryEmbedding()
+          )
+          (mlp): Qwen2MLP(
+            (gate_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (up_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=4864, bias=False)
+            (down_proj): SupermaskLinearSparsity_SoftForward_Normal(in_features=4864, out_features=896, bias=False)
+            (act_fn): SiLU()
+          )
+          (input_layernorm): Qwen2RMSNorm()
+          (post_attention_layernorm): Qwen2RMSNorm()
+        )
+      )
+      (norm): Qwen2RMSNorm()
+    )
+    (lm_head): Linear(in_features=896, out_features=151936, bias=False)
+  )
+  (vision_tower): SIGLIPVisionTower(
+    (_vision_tower): SiglipVisionModel(
+      (vision_model): SiglipVisionTransformer(
+        (embeddings): SiglipVisionEmbeddings(
+          (patch_embedding): Conv2d(3, 1152, kernel_size=(14, 14), stride=(14, 14), padding=valid)
+          (position_embedding): Embedding(729, 1152)
+        )
+        (encoder): SiglipEncoder(
+          (layers): ModuleList(
+            (0-26): 27 x SiglipEncoderLayer(
+              (self_attn): SiglipAttention(
+                (k_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (v_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (q_proj): Linear(in_features=1152, out_features=1152, bias=True)
+                (out_proj): Linear(in_features=1152, out_features=1152, bias=True)
+              )
+              (layer_norm1): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+              (mlp): SiglipMLP(
+                (activation_fn): PytorchGELUTanh()
+                (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+                (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+              )
+              (layer_norm2): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+            )
+          )
+        )
+        (post_layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+        (head): SiglipMultiheadAttentionPoolingHead(
+          (attention): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=1152, out_features=1152, bias=True)
+          )
+          (layernorm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True)
+          (mlp): SiglipMLP(
+            (activation_fn): PytorchGELUTanh()
+            (fc1): Linear(in_features=1152, out_features=4304, bias=True)
+            (fc2): Linear(in_features=4304, out_features=1152, bias=True)
+          )
+        )
+      )
+    )
+  )
+  (connector): MLPConnector(
+    (_connector): Sequential(
+      (0): SupermaskLinearSparsity_SoftForward_Normal(in_features=1152, out_features=896, bias=True)
+      (1): GELU(approximate='none')
+      (2): SupermaskLinearSparsity_SoftForward_Normal(in_features=896, out_features=896, bias=True)
+    )
+  )
+)
+Pre-training init language_model.model.layers.0.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.0.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.1.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.2.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.3.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.4.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.5.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.6.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.7.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.8.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.9.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.10.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.11.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.12.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.13.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.14.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.15.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.16.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.17.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.18.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.19.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.20.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.21.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.22.mlp.down_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.q_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.k_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.v_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.self_attn.o_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.mlp.gate_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.mlp.up_proj.scores: Mean=7.000000
+Pre-training init language_model.model.layers.23.mlp.down_proj.scores: Mean=7.000000
+Pre-training init connector._connector.0.scores: Mean=7.000005
+Pre-training init connector._connector.2.scores: Mean=6.999971
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+Randomly sampled 66529 training samples (10.0% of 665298 total samples)
+2025-10-16 19:02:00,709 | INFO: Total Parameters: 1283756736, Total Trainable Parameters: 359661568
+2025-10-16 19:02:00,714 | INFO: Trainable Parameters:
+language_model.model.layers.0.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.0.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.0.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.0.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.0.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.1.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.1.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.1.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.1.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.1.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.2.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.2.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.2.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.2.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.2.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.3.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.3.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.3.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.3.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.3.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.4.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.4.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.4.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.4.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.4.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.5.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.5.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.5.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.5.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.5.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.6.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.6.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.6.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.6.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.6.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.7.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.7.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.7.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.7.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.7.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.8.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.8.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.8.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.8.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.8.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.9.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.9.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.9.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.9.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.9.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.10.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.10.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.10.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.10.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.10.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.11.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.11.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.11.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.11.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.11.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.12.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.12.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.12.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.12.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.12.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.13.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.13.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.13.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.13.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.13.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.14.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.14.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.14.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.14.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.14.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.15.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.15.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.15.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.15.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.15.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.16.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.16.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.16.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.16.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.16.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.17.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.17.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.17.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.17.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.17.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.18.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.18.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.18.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.18.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.18.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.19.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.19.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.19.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.19.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.19.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.20.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.20.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.20.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.20.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.20.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.21.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.21.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.21.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.21.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.21.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.22.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.22.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.22.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.22.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.22.mlp.down_proj.scores: 4358144 parameters
+language_model.model.layers.23.self_attn.q_proj.scores: 802816 parameters
+language_model.model.layers.23.self_attn.k_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.v_proj.scores: 114688 parameters
+language_model.model.layers.23.self_attn.o_proj.scores: 802816 parameters
+language_model.model.layers.23.mlp.gate_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.up_proj.scores: 4358144 parameters
+language_model.model.layers.23.mlp.down_proj.scores: 4358144 parameters
+connector._connector.0.scores: 1032192 parameters
+connector._connector.2.scores: 802816 parameters
+Parameter Offload: Total persistent parameters: 486464 in 403 params
+
  0%|          | 0/520 [00:00<?, ?it/s]/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
+  warnings.warn(
+/opt/conda/envs/tinyllava/lib/python3.10/site-packages/torch/utils/checkpoint.py:61: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
+  warnings.warn(
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Using non-device net plugin version 0
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Using network Socket
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO ncclCommInitRank comm 0x7f962406aee0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO ncclCommInitRank comm 0x7f139806b8d0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO ncclCommInitRank comm 0x7f9ad006a7f0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO ncclCommInitRank comm 0x7f870406b1f0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO ncclCommInitRank comm 0x7f371c06a860 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO ncclCommInitRank comm 0x7efcdc06aa90 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO ncclCommInitRank comm 0x7f217406b3f0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO ncclCommInitRank comm 0x7f6f9806b200 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x6cb3a76280ec57c9 - Init START
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO NVLS multicast support is not available on dev 3
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Setting affinity for GPU 5 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO NVLS multicast support is not available on dev 5
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO NVLS multicast support is not available on dev 2
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Setting affinity for GPU 4 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO NVLS multicast support is not available on dev 4
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Setting affinity for GPU 6 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO NVLS multicast support is not available on dev 6
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Setting affinity for GPU 7 to ffffff00,0000ffff,ff000000
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO NVLS multicast support is not available on dev 7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO NVLS multicast support is not available on dev 0
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffff0000,00ffffff
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO NVLS multicast support is not available on dev 1
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO comm 0x7f962406aee0 rank 1 nRanks 8 nNodes 1 localRanks 8 localRank 1 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO comm 0x7f9ad006a7f0 rank 2 nRanks 8 nNodes 1 localRanks 8 localRank 2 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO comm 0x7f139806b8d0 rank 0 nRanks 8 nNodes 1 localRanks 8 localRank 0 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO comm 0x7f371c06a860 rank 7 nRanks 8 nNodes 1 localRanks 8 localRank 7 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 00/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO comm 0x7f6f9806b200 rank 5 nRanks 8 nNodes 1 localRanks 8 localRank 5 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 01/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 02/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO comm 0x7efcdc06aa90 rank 4 nRanks 8 nNodes 1 localRanks 8 localRank 4 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 03/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 04/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO comm 0x7f217406b3f0 rank 6 nRanks 8 nNodes 1 localRanks 8 localRank 6 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 05/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO comm 0x7f870406b1f0 rank 3 nRanks 8 nNodes 1 localRanks 8 localRank 3 MNNVL 0
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 06/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 07/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 08/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 09/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 10/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 11/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 12/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 13/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 14/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 15/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 16/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 17/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 18/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 19/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 20/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 21/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 22/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 23/24 :    0   1   2   3   4   5   6   7
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO P2P Chunksize set to 524288
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Connected all rings
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/CUMEM/read
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO Connected all trees
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO 24 coll channels, 24 collnet channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer
+ywang29-vrdb-test2-worker-0:2622013:2628561 [0] NCCL INFO ncclCommInitRank comm 0x7f139806b8d0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 101c0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622017:2628566 [4] NCCL INFO ncclCommInitRank comm 0x7efcdc06aa90 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 901c0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622020:2628564 [7] NCCL INFO ncclCommInitRank comm 0x7f371c06a860 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId a01d0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622016:2628563 [3] NCCL INFO ncclCommInitRank comm 0x7f870406b1f0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 201d0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622014:2628562 [1] NCCL INFO ncclCommInitRank comm 0x7f962406aee0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 101d0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622018:2628567 [5] NCCL INFO ncclCommInitRank comm 0x7f6f9806b200 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId 901d0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622015:2628568 [2] NCCL INFO ncclCommInitRank comm 0x7f9ad006a7f0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 201c0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+ywang29-vrdb-test2-worker-0:2622019:2628565 [6] NCCL INFO ncclCommInitRank comm 0x7f217406b3f0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId a01c0 commId 0x6cb3a76280ec57c9 - Init COMPLETE
+
  0%|          | 1/520 [00:22<3:17:53, 22.88s/it]
                                                 
{'loss': 2.0497, 'grad_norm': 0.0, 'learning_rate': 0.0125, 'epoch': 0.0}
+
  0%|          | 1/520 [00:22<3:17:53, 22.88s/it]
  0%|          | 2/520 [00:26<1:40:54, 11.69s/it]
                                                 
{'loss': 2.06, 'grad_norm': 0.0, 'learning_rate': 0.025, 'epoch': 0.0}
+
  0%|          | 2/520 [00:26<1:40:54, 11.69s/it]
  1%|          | 3/520 [00:30<1:10:03,  8.13s/it]
                                                 
{'loss': 2.1958, 'grad_norm': 0.0, 'learning_rate': 0.037500000000000006, 'epoch': 0.01}
+
  1%|          | 3/520 [00:30<1:10:03,  8.13s/it]
  1%|          | 4/520 [00:34<55:24,  6.44s/it]  
                                               
{'loss': 2.0688, 'grad_norm': 0.0, 'learning_rate': 0.05, 'epoch': 0.01}
+
  1%|          | 4/520 [00:34<55:24,  6.44s/it]
  1%|          | 5/520 [00:38<47:28,  5.53s/it]
                                               
{'loss': 2.2403, 'grad_norm': 0.0, 'learning_rate': 0.0625, 'epoch': 0.01}
+
  1%|          | 5/520 [00:38<47:28,  5.53s/it]
  1%|          | 6/520 [00:42<42:34,  4.97s/it]
                                               
{'loss': 1.6782, 'grad_norm': 0.0, 'learning_rate': 0.07500000000000001, 'epoch': 0.01}
+
  1%|          | 6/520 [00:42<42:34,  4.97s/it]
  1%|▏         | 7/520 [00:46<39:23,  4.61s/it]
                                               
{'loss': 2.0829, 'grad_norm': 0.0, 'learning_rate': 0.08750000000000001, 'epoch': 0.01}
+
  1%|▏         | 7/520 [00:46<39:23,  4.61s/it]
  2%|▏         | 8/520 [00:50<38:59,  4.57s/it]
                                               
{'loss': 2.0585, 'grad_norm': 0.0, 'learning_rate': 0.1, 'epoch': 0.02}
+
  2%|▏         | 8/520 [00:50<38:59,  4.57s/it]
  2%|▏         | 9/520 [00:54<37:04,  4.35s/it]
                                               
{'loss': 2.1936, 'grad_norm': 0.0, 'learning_rate': 0.1125, 'epoch': 0.02}
+
  2%|▏         | 9/520 [00:54<37:04,  4.35s/it]
  2%|▏         | 10/520 [00:58<35:45,  4.21s/it]
                                                
{'loss': 2.0887, 'grad_norm': 0.0, 'learning_rate': 0.125, 'epoch': 0.02}
+
  2%|▏         | 10/520 [00:58<35:45,  4.21s/it]
  2%|▏         | 11/520 [01:02<35:04,  4.14s/it]
                                                
{'loss': 2.0637, 'grad_norm': 0.0, 'learning_rate': 0.1375, 'epoch': 0.02}
+
  2%|▏         | 11/520 [01:02<35:04,  4.14s/it]
  2%|▏         | 12/520 [01:06<34:18,  4.05s/it]
                                                
{'loss': 1.8848, 'grad_norm': 0.0, 'learning_rate': 0.15000000000000002, 'epoch': 0.02}
+
  2%|▏         | 12/520 [01:06<34:18,  4.05s/it][2025-10-16 19:03:16,766] [WARNING] [stage3.py:2069:step] 1 pytorch allocator cache flushes since last step. this happens when there is high memory pressure and is detrimental to performance. if this is happening frequently consider adjusting settings to reduce memory consumption. If you are unable to make the cache flushes go away consider adding get_accelerator().empty_cache() calls in your training loop to ensure that all ranks flush their caches at the same time
+
  2%|▎         | 13/520 [01:10<35:33,  4.21s/it]
                                                
{'loss': 2.0728, 'grad_norm': 0.0, 'learning_rate': 0.1625, 'epoch': 0.03}
+
  2%|▎         | 13/520 [01:10<35:33,  4.21s/it]
  3%|▎         | 14/520 [01:14<34:35,  4.10s/it]
                                                
{'loss': 2.1118, 'grad_norm': 0.0, 'learning_rate': 0.17500000000000002, 'epoch': 0.03}
+
  3%|▎         | 14/520 [01:14<34:35,  4.10s/it]
  3%|▎         | 15/520 [01:18<33:55,  4.03s/it]
                                                
{'loss': 1.7478, 'grad_norm': 0.0, 'learning_rate': 0.1875, 'epoch': 0.03}
+
  3%|▎         | 15/520 [01:18<33:55,  4.03s/it]
  3%|▎         | 16/520 [01:22<33:21,  3.97s/it]
                                                
{'loss': 1.8954, 'grad_norm': 0.0, 'learning_rate': 0.2, 'epoch': 0.03}
+
  3%|▎         | 16/520 [01:22<33:21,  3.97s/it]
  3%|▎         | 17/520 [01:26<32:58,  3.93s/it]
                                                
{'loss': 2.1158, 'grad_norm': 0.0, 'learning_rate': 0.1999980572931538, 'epoch': 0.03}
+
  3%|▎         | 17/520 [01:26<32:58,  3.93s/it]
  3%|▎         | 18/520 [01:30<32:41,  3.91s/it]
                                                
{'loss': 2.1718, 'grad_norm': 0.0, 'learning_rate': 0.19999222924809748, 'epoch': 0.03}
+
  3%|▎         | 18/520 [01:30<32:41,  3.91s/it]
  4%|▎         | 19/520 [01:33<32:28,  3.89s/it]
                                                
{'loss': 1.8467, 'grad_norm': 0.0, 'learning_rate': 0.19998251609127465, 'epoch': 0.04}
+
  4%|▎         | 19/520 [01:33<32:28,  3.89s/it]
  4%|▍         | 20/520 [01:37<32:15,  3.87s/it]
                                                
{'loss': 2.2091, 'grad_norm': 0.0, 'learning_rate': 0.19996891820008164, 'epoch': 0.04}
+
  4%|▍         | 20/520 [01:37<32:15,  3.87s/it]
  4%|▍         | 21/520 [01:41<32:15,  3.88s/it]
                                                
{'loss': 2.0718, 'grad_norm': 0.0, 'learning_rate': 0.19995143610285276, 'epoch': 0.04}
+
  4%|▍         | 21/520 [01:41<32:15,  3.88s/it]
  4%|▍         | 22/520 [01:45<32:11,  3.88s/it]
                                                
{'loss': 2.0488, 'grad_norm': 0.0, 'learning_rate': 0.19993007047883987, 'epoch': 0.04}
+
  4%|▍         | 22/520 [01:45<32:11,  3.88s/it]
  4%|▍         | 23/520 [01:49<32:09,  3.88s/it]
                                                
{'loss': 2.0811, 'grad_norm': 0.0, 'learning_rate': 0.1999048221581858, 'epoch': 0.04}
+
  4%|▍         | 23/520 [01:49<32:09,  3.88s/it]
\ No newline at end of file