Hugging Face's logo

almanach
/
gaperon-quality-classifier

Text Classification
Transformers
ONNX
Safetensors
French
English
xlm-roberta
gaperon
quality-classifier
document-quality
data-curation
Model card Files
xet
Community
wissamantoun commited on
Commit
a385183
·
verified ·
1 Parent(s): 79e5c66

Upload folder using huggingface_hub

Browse files
Files changed (24) hide show
  1. .gitattributes +4 -0
  2. all_results.json +20 -0
  3. config.json +52 -0
  4. confusion_matrix.txt +52 -0
  5. confusion_matrix_norm.txt +52 -0
  6. eval_results.json +9 -0
  7. mgx/model-512-512-fp16.mxr +3 -0
  8. model.safetensors +3 -0
  9. onnx/config.json +54 -0
  10. onnx/model.onnx +3 -0
  11. onnx/model.onnx.data +3 -0
  12. onnx/ort_config.json +37 -0
  13. onnx/sentencepiece.bpe.model +3 -0
  14. onnx/special_tokens_map.json +51 -0
  15. onnx/tokenizer.json +3 -0
  16. onnx/tokenizer_config.json +54 -0
  17. predict_results_quality.txt +0 -0
  18. sentencepiece.bpe.model +3 -0
  19. special_tokens_map.json +15 -0
  20. test_results.json +7 -0
  21. tokenizer.json +3 -0
  22. tokenizer_config.json +54 -0
  23. train_results.json +9 -0
  24. trainer_state.json +233 -0
.gitattributes CHANGED
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ mgx/model-512-512-fp16.mxr filter=lfs diff=lfs merge=lfs -text
37
+ onnx/model.onnx.data filter=lfs diff=lfs merge=lfs -text
38
+ onnx/tokenizer.json filter=lfs diff=lfs merge=lfs -text
39
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
all_results.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "eval_loss": 0.5344868302345276,
4
+ "eval_quality_f1": 0.7511665338396387,
5
+ "eval_runtime": 1.1096,
6
+ "eval_samples": 1000,
7
+ "eval_samples_per_second": 901.238,
8
+ "eval_steps_per_second": 1.802,
9
+ "predict_loss": 0.501552164554596,
10
+ "predict_quality_f1": 0.7696919433036512,
11
+ "predict_runtime": 16.0557,
12
+ "predict_samples_per_second": 622.832,
13
+ "predict_steps_per_second": 1.246,
14
+ "total_flos": 6.010120884036649e+17,
15
+ "train_loss": 0.4924640074425701,
16
+ "train_runtime": 8914.4901,
17
+ "train_samples": 285529,
18
+ "train_samples_per_second": 256.238,
19
+ "train_steps_per_second": 1.002
20
+ }
config.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/lus/work/CT10/c1615138/SHARED/scratch/gaperon-data/models/facebook--xlm-v-base",
3
+ "architectures": [
4
+ "MultiTaskClsModel"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "eos_token_id": 2,
10
+ "finetuning_task": "text-classification",
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.1,
13
+ "hidden_size": 768,
14
+ "id2label_dict": {
15
+ "quality": {
16
+ "0": "high",
17
+ "1": "low",
18
+ "2": "medium"
19
+ }
20
+ },
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 3072,
23
+ "label2id_dict": {
24
+ "quality": {
25
+ "high": 0,
26
+ "low": 1,
27
+ "medium": 2
28
+ }
29
+ },
30
+ "labels_list": [
31
+ [
32
+ "high",
33
+ "low",
34
+ "medium"
35
+ ]
36
+ ],
37
+ "layer_norm_eps": 1e-05,
38
+ "max_position_embeddings": 514,
39
+ "model_type": "xlm-roberta",
40
+ "num_attention_heads": 12,
41
+ "num_hidden_layers": 12,
42
+ "pad_token_id": 1,
43
+ "position_embedding_type": "absolute",
44
+ "problem_types": [
45
+ "single_label_classification"
46
+ ],
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.46.2",
49
+ "type_vocab_size": 1,
50
+ "use_cache": true,
51
+ "vocab_size": 901629
52
+ }
confusion_matrix.txt ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Confusion matrix for quality:
2
+ t/p high low medium
3
+ high 171.0 3.0 72.0
4
+ low 6.0 79.0 56.0
5
+ medium 81.0 26.0 506.0
6
+ Confusion matrix for quality:
7
+ t/p high low medium
8
+ high 175.0 2.0 69.0
9
+ low 6.0 67.0 68.0
10
+ medium 71.0 15.0 527.0
11
+ Confusion matrix for quality:
12
+ t/p high low medium
13
+ high 187.0 3.0 56.0
14
+ low 6.0 84.0 51.0
15
+ medium 93.0 24.0 496.0
16
+ Confusion matrix for quality:
17
+ t/p high low medium
18
+ high 181.0 3.0 62.0
19
+ low 6.0 86.0 49.0
20
+ medium 67.0 24.0 522.0
21
+ Confusion matrix for quality:
22
+ t/p high low medium
23
+ high 167.0 4.0 75.0
24
+ low 6.0 92.0 43.0
25
+ medium 48.0 29.0 536.0
26
+ Confusion matrix for quality:
27
+ t/p high low medium
28
+ high 183.0 4.0 59.0
29
+ low 6.0 82.0 53.0
30
+ medium 72.0 23.0 518.0
31
+ Confusion matrix for quality:
32
+ t/p high low medium
33
+ high 181.0 4.0 61.0
34
+ low 6.0 88.0 47.0
35
+ medium 65.0 24.0 524.0
36
+ Confusion matrix for quality:
37
+ t/p high low medium
38
+ high 180.0 4.0 62.0
39
+ low 6.0 87.0 48.0
40
+ medium 68.0 24.0 521.0
41
+ Final Evaluation confusion matrix:
42
+ Confusion matrix for quality:
43
+ t/p high low medium
44
+ high 181.0 4.0 61.0
45
+ low 6.0 88.0 47.0
46
+ medium 65.0 24.0 524.0
47
+ Prediction confusion matrix:
48
+ Confusion matrix for quality:
49
+ t/p high low medium
50
+ high 1930.0 32.0 531.0
51
+ low 77.0 922.0 463.0
52
+ medium 623.0 203.0 5219.0
confusion_matrix_norm.txt ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Confusion matrix for quality:
2
+ t/p high low medium
3
+ high 0.2 0.0 0.1
4
+ low 0.0 0.1 0.1
5
+ medium 0.1 0.0 0.5
6
+ Confusion matrix for quality:
7
+ t/p high low medium
8
+ high 0.2 0.0 0.1
9
+ low 0.0 0.1 0.1
10
+ medium 0.1 0.0 0.5
11
+ Confusion matrix for quality:
12
+ t/p high low medium
13
+ high 0.2 0.0 0.1
14
+ low 0.0 0.1 0.1
15
+ medium 0.1 0.0 0.5
16
+ Confusion matrix for quality:
17
+ t/p high low medium
18
+ high 0.2 0.0 0.1
19
+ low 0.0 0.1 0.0
20
+ medium 0.1 0.0 0.5
21
+ Confusion matrix for quality:
22
+ t/p high low medium
23
+ high 0.2 0.0 0.1
24
+ low 0.0 0.1 0.0
25
+ medium 0.0 0.0 0.5
26
+ Confusion matrix for quality:
27
+ t/p high low medium
28
+ high 0.2 0.0 0.1
29
+ low 0.0 0.1 0.1
30
+ medium 0.1 0.0 0.5
31
+ Confusion matrix for quality:
32
+ t/p high low medium
33
+ high 0.2 0.0 0.1
34
+ low 0.0 0.1 0.0
35
+ medium 0.1 0.0 0.5
36
+ Confusion matrix for quality:
37
+ t/p high low medium
38
+ high 0.2 0.0 0.1
39
+ low 0.0 0.1 0.0
40
+ medium 0.1 0.0 0.5
41
+ Final Evaluation normalized confusion matrix:
42
+ Confusion matrix for quality:
43
+ t/p high low medium
44
+ high 0.2 0.0 0.1
45
+ low 0.0 0.1 0.0
46
+ medium 0.1 0.0 0.5
47
+ Prediction normalized confusion matrix:
48
+ Confusion matrix for quality:
49
+ t/p high low medium
50
+ high 0.2 0.0 0.1
51
+ low 0.0 0.1 0.0
52
+ medium 0.1 0.0 0.5
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "eval_loss": 0.5344868302345276,
4
+ "eval_quality_f1": 0.7511665338396387,
5
+ "eval_runtime": 1.1096,
6
+ "eval_samples": 1000,
7
+ "eval_samples_per_second": 901.238,
8
+ "eval_steps_per_second": 1.802
9
+ }
mgx/model-512-512-fp16.mxr ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0022bfb5cc335417b2d3a9563b32a5f185400af0af22d17affbf970a85717fc0
3
+ size 2026898336
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ccdef4508b10148906fae0b3508a78a5b6416051a14297acc49e23e4ed6ee006
3
+ size 3114006492
onnx/config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/scratch/gaperon-data/models/quality_cls/xlm-v/llama-only-filtered-single--512--25--8e-06--64--linear--0.0/onnx",
3
+ "architectures": [
4
+ "MultiTaskClsModel"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "attn_implementation": null,
8
+ "bos_token_id": 0,
9
+ "classifier_dropout": null,
10
+ "eos_token_id": 2,
11
+ "finetuning_task": "text-classification",
12
+ "hidden_act": "gelu",
13
+ "hidden_dropout_prob": 0.1,
14
+ "hidden_size": 768,
15
+ "id2label_dict": {
16
+ "quality": {
17
+ "0": "high",
18
+ "1": "low",
19
+ "2": "medium"
20
+ }
21
+ },
22
+ "initializer_range": 0.02,
23
+ "intermediate_size": 3072,
24
+ "label2id_dict": {
25
+ "quality": {
26
+ "high": 0,
27
+ "low": 1,
28
+ "medium": 2
29
+ }
30
+ },
31
+ "labels_list": [
32
+ [
33
+ "high",
34
+ "low",
35
+ "medium"
36
+ ]
37
+ ],
38
+ "layer_norm_eps": 1e-05,
39
+ "max_position_embeddings": 514,
40
+ "model_type": "multitaskcls",
41
+ "num_attention_heads": 12,
42
+ "num_hidden_layers": 12,
43
+ "num_tasks": 1,
44
+ "pad_token_id": 1,
45
+ "position_embedding_type": "absolute",
46
+ "problem_types": [
47
+ "single_label_classification"
48
+ ],
49
+ "torch_dtype": "float32",
50
+ "transformers_version": "4.44.2",
51
+ "type_vocab_size": 1,
52
+ "use_cache": true,
53
+ "vocab_size": 901629
54
+ }
onnx/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a3b9482b3ae966b0b1668335f7f1559f5f48adddfc951bf8a6ed48271854093
3
+ size 181983
onnx/model.onnx.data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cac89c2e951864309b6bfef02c937db682bb585fe71751f95f809c22bbc2f023
3
+ size 3113986064
onnx/ort_config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "one_external_file": true,
3
+ "opset": null,
4
+ "optimization": {
5
+ "disable_attention": null,
6
+ "disable_attention_fusion": false,
7
+ "disable_bias_gelu": null,
8
+ "disable_bias_gelu_fusion": false,
9
+ "disable_bias_skip_layer_norm": null,
10
+ "disable_bias_skip_layer_norm_fusion": false,
11
+ "disable_embed_layer_norm": true,
12
+ "disable_embed_layer_norm_fusion": true,
13
+ "disable_gelu": null,
14
+ "disable_gelu_fusion": false,
15
+ "disable_group_norm_fusion": true,
16
+ "disable_layer_norm": null,
17
+ "disable_layer_norm_fusion": false,
18
+ "disable_packed_kv": true,
19
+ "disable_rotary_embeddings": false,
20
+ "disable_shape_inference": true,
21
+ "disable_skip_layer_norm": null,
22
+ "disable_skip_layer_norm_fusion": false,
23
+ "enable_gelu_approximation": false,
24
+ "enable_gemm_fast_gelu_fusion": false,
25
+ "enable_transformers_specific_optimizations": false,
26
+ "fp16": false,
27
+ "no_attention_mask": false,
28
+ "optimization_level": 1,
29
+ "optimize_for_gpu": false,
30
+ "optimize_with_onnxruntime_only": null,
31
+ "use_mask_index": false,
32
+ "use_multi_head_attention": false,
33
+ "use_raw_attention_mask": false
34
+ },
35
+ "quantization": {},
36
+ "use_external_data_format": true
37
+ }
onnx/sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dbee9f6b1c2cba29b335e70d6088eea943c7d5ae55ac7dd17174760bf758e309
3
+ size 18241665
onnx/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
onnx/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b05f53e32927389de39388f3049315b4169bf5ac73a0cb1b97ad14384170204
3
+ size 61333617
onnx/tokenizer_config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "901628": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "mask_token": "<mask>",
49
+ "model_max_length": 1000000000000000019884624838656,
50
+ "pad_token": "<pad>",
51
+ "sep_token": "</s>",
52
+ "tokenizer_class": "XLMRobertaTokenizer",
53
+ "unk_token": "<unk>"
54
+ }
predict_results_quality.txt ADDED
The diff for this file is too large to render. See raw diff
 
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dbee9f6b1c2cba29b335e70d6088eea943c7d5ae55ac7dd17174760bf758e309
3
+ size 18241665
special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
test_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "predict_loss": 0.501552164554596,
3
+ "predict_quality_f1": 0.7696919433036512,
4
+ "predict_runtime": 16.0557,
5
+ "predict_samples_per_second": 622.832,
6
+ "predict_steps_per_second": 1.246
7
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b05f53e32927389de39388f3049315b4169bf5ac73a0cb1b97ad14384170204
3
+ size 61333617
tokenizer_config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "901628": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "mask_token": "<mask>",
49
+ "model_max_length": 1000000000000000019884624838656,
50
+ "pad_token": "<pad>",
51
+ "sep_token": "</s>",
52
+ "tokenizer_class": "XLMRobertaTokenizer",
53
+ "unk_token": "<unk>"
54
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "total_flos": 6.010120884036649e+17,
4
+ "train_loss": 0.4924640074425701,
5
+ "train_runtime": 8914.4901,
6
+ "train_samples": 285529,
7
+ "train_samples_per_second": 256.238,
8
+ "train_steps_per_second": 1.002
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,233 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.7511665338396387,
3
+ "best_model_checkpoint": "/lus/work/CT10/c1615138/SHARED/scratch/gaperon-data/models/quality_cls/xlm-v/llama-only-filtered-single--512--25--8e-06--64--linear--0.0/checkpoint-7812",
4
+ "epoch": 8.0,
5
+ "eval_steps": 500,
6
+ "global_step": 8928,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.44802867383512546,
13
+ "grad_norm": 8.850239753723145,
14
+ "learning_rate": 7.551971326164874e-06,
15
+ "loss": 0.7332,
16
+ "step": 500
17
+ },
18
+ {
19
+ "epoch": 0.8960573476702509,
20
+ "grad_norm": 3.0268373489379883,
21
+ "learning_rate": 7.103942652329749e-06,
22
+ "loss": 0.562,
23
+ "step": 1000
24
+ },
25
+ {
26
+ "epoch": 1.0,
27
+ "eval_loss": 0.582636296749115,
28
+ "eval_quality_f1": 0.7082190985655687,
29
+ "eval_runtime": 2.4363,
30
+ "eval_samples_per_second": 410.466,
31
+ "eval_steps_per_second": 0.821,
32
+ "step": 1116
33
+ },
34
+ {
35
+ "epoch": 1.3440860215053765,
36
+ "grad_norm": 4.1538987159729,
37
+ "learning_rate": 6.655913978494624e-06,
38
+ "loss": 0.5385,
39
+ "step": 1500
40
+ },
41
+ {
42
+ "epoch": 1.7921146953405018,
43
+ "grad_norm": 2.82087779045105,
44
+ "learning_rate": 6.2078853046594985e-06,
45
+ "loss": 0.5244,
46
+ "step": 2000
47
+ },
48
+ {
49
+ "epoch": 2.0,
50
+ "eval_loss": 0.5524051785469055,
51
+ "eval_quality_f1": 0.7079129220265732,
52
+ "eval_runtime": 1.1303,
53
+ "eval_samples_per_second": 884.757,
54
+ "eval_steps_per_second": 1.77,
55
+ "step": 2232
56
+ },
57
+ {
58
+ "epoch": 2.240143369175627,
59
+ "grad_norm": 5.126065254211426,
60
+ "learning_rate": 5.759856630824372e-06,
61
+ "loss": 0.5122,
62
+ "step": 2500
63
+ },
64
+ {
65
+ "epoch": 2.688172043010753,
66
+ "grad_norm": 4.198544025421143,
67
+ "learning_rate": 5.311827956989247e-06,
68
+ "loss": 0.5031,
69
+ "step": 3000
70
+ },
71
+ {
72
+ "epoch": 3.0,
73
+ "eval_loss": 0.5608200430870056,
74
+ "eval_quality_f1": 0.7284878863826232,
75
+ "eval_runtime": 1.1236,
76
+ "eval_samples_per_second": 890.01,
77
+ "eval_steps_per_second": 1.78,
78
+ "step": 3348
79
+ },
80
+ {
81
+ "epoch": 3.1362007168458783,
82
+ "grad_norm": 2.6578023433685303,
83
+ "learning_rate": 4.863799283154122e-06,
84
+ "loss": 0.4919,
85
+ "step": 3500
86
+ },
87
+ {
88
+ "epoch": 3.5842293906810037,
89
+ "grad_norm": 4.944795608520508,
90
+ "learning_rate": 4.4157706093189966e-06,
91
+ "loss": 0.4817,
92
+ "step": 4000
93
+ },
94
+ {
95
+ "epoch": 4.0,
96
+ "eval_loss": 0.5346963405609131,
97
+ "eval_quality_f1": 0.7463488580781336,
98
+ "eval_runtime": 3.8034,
99
+ "eval_samples_per_second": 262.92,
100
+ "eval_steps_per_second": 0.526,
101
+ "step": 4464
102
+ },
103
+ {
104
+ "epoch": 4.032258064516129,
105
+ "grad_norm": 5.10132360458374,
106
+ "learning_rate": 3.9677419354838704e-06,
107
+ "loss": 0.4817,
108
+ "step": 4500
109
+ },
110
+ {
111
+ "epoch": 4.480286738351254,
112
+ "grad_norm": 5.194272994995117,
113
+ "learning_rate": 3.5197132616487456e-06,
114
+ "loss": 0.4682,
115
+ "step": 5000
116
+ },
117
+ {
118
+ "epoch": 4.92831541218638,
119
+ "grad_norm": 3.7467427253723145,
120
+ "learning_rate": 3.07168458781362e-06,
121
+ "loss": 0.4634,
122
+ "step": 5500
123
+ },
124
+ {
125
+ "epoch": 5.0,
126
+ "eval_loss": 0.5277291536331177,
127
+ "eval_quality_f1": 0.7510086276061395,
128
+ "eval_runtime": 1.5031,
129
+ "eval_samples_per_second": 665.304,
130
+ "eval_steps_per_second": 1.331,
131
+ "step": 5580
132
+ },
133
+ {
134
+ "epoch": 5.376344086021505,
135
+ "grad_norm": 5.132294654846191,
136
+ "learning_rate": 2.6236559139784946e-06,
137
+ "loss": 0.4563,
138
+ "step": 6000
139
+ },
140
+ {
141
+ "epoch": 5.824372759856631,
142
+ "grad_norm": 5.5977020263671875,
143
+ "learning_rate": 2.175627240143369e-06,
144
+ "loss": 0.449,
145
+ "step": 6500
146
+ },
147
+ {
148
+ "epoch": 6.0,
149
+ "eval_loss": 0.5459088087081909,
150
+ "eval_quality_f1": 0.7371203028874914,
151
+ "eval_runtime": 1.2411,
152
+ "eval_samples_per_second": 805.766,
153
+ "eval_steps_per_second": 1.612,
154
+ "step": 6696
155
+ },
156
+ {
157
+ "epoch": 6.272401433691757,
158
+ "grad_norm": 5.542992115020752,
159
+ "learning_rate": 1.7275985663082438e-06,
160
+ "loss": 0.4466,
161
+ "step": 7000
162
+ },
163
+ {
164
+ "epoch": 6.720430107526882,
165
+ "grad_norm": 4.132849216461182,
166
+ "learning_rate": 1.2795698924731181e-06,
167
+ "loss": 0.4422,
168
+ "step": 7500
169
+ },
170
+ {
171
+ "epoch": 7.0,
172
+ "eval_loss": 0.5344868302345276,
173
+ "eval_quality_f1": 0.7511665338396387,
174
+ "eval_runtime": 1.1245,
175
+ "eval_samples_per_second": 889.293,
176
+ "eval_steps_per_second": 1.779,
177
+ "step": 7812
178
+ },
179
+ {
180
+ "epoch": 7.168458781362007,
181
+ "grad_norm": 3.255526304244995,
182
+ "learning_rate": 8.315412186379928e-07,
183
+ "loss": 0.4382,
184
+ "step": 8000
185
+ },
186
+ {
187
+ "epoch": 7.616487455197133,
188
+ "grad_norm": 6.981889247894287,
189
+ "learning_rate": 3.8351254480286737e-07,
190
+ "loss": 0.433,
191
+ "step": 8500
192
+ },
193
+ {
194
+ "epoch": 8.0,
195
+ "eval_loss": 0.5395817160606384,
196
+ "eval_quality_f1": 0.7457693595927116,
197
+ "eval_runtime": 9.4684,
198
+ "eval_samples_per_second": 105.614,
199
+ "eval_steps_per_second": 0.211,
200
+ "step": 8928
201
+ },
202
+ {
203
+ "epoch": 8.0,
204
+ "step": 8928,
205
+ "total_flos": 6.010120884036649e+17,
206
+ "train_loss": 0.4924640074425701,
207
+ "train_runtime": 8914.4901,
208
+ "train_samples_per_second": 256.238,
209
+ "train_steps_per_second": 1.002
210
+ }
211
+ ],
212
+ "logging_steps": 500,
213
+ "max_steps": 8928,
214
+ "num_input_tokens_seen": 0,
215
+ "num_train_epochs": 8,
216
+ "save_steps": 500,
217
+ "stateful_callbacks": {
218
+ "TrainerControl": {
219
+ "args": {
220
+ "should_epoch_stop": false,
221
+ "should_evaluate": false,
222
+ "should_log": false,
223
+ "should_save": true,
224
+ "should_training_stop": true
225
+ },
226
+ "attributes": {}
227
+ }
228
+ },
229
+ "total_flos": 6.010120884036649e+17,
230
+ "train_batch_size": 256,
231
+ "trial_name": null,
232
+ "trial_params": null
233
+ }