trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.054988216810683,
  "eval_steps": 1000,
  "global_step": 800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.26,
      "learning_rate": 0.0001,
      "loss": 0.1991,
      "step": 100
    },
    {
      "epoch": 2.51,
      "learning_rate": 0.0001,
      "loss": 0.076,
      "step": 200
    },
    {
      "epoch": 3.77,
      "learning_rate": 0.0001,
      "loss": 0.061,
      "step": 300
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0001,
      "loss": 0.0522,
      "step": 400
    },
    {
      "epoch": 6.28,
      "learning_rate": 0.0001,
      "loss": 0.0471,
      "step": 500
    },
    {
      "epoch": 7.54,
      "learning_rate": 0.0001,
      "loss": 0.044,
      "step": 600
    },
    {
      "epoch": 8.8,
      "learning_rate": 0.0001,
      "loss": 0.0411,
      "step": 700
    },
    {
      "epoch": 10.05,
      "learning_rate": 0.0001,
      "loss": 0.0383,
      "step": 800
    }
  ],
  "logging_steps": 100,
  "max_steps": 4500,
  "num_train_epochs": 57,
  "save_steps": 200,
  "total_flos": 2.710103747932979e+18,
  "trial_name": null,
  "trial_params": null
}