trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 199.66666666666666,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 199.67,
      "eval_loss": 5.05859375,
      "eval_runtime": 0.1079,
      "eval_samples_per_second": 64.877,
      "eval_steps_per_second": 9.268,
      "step": 200
    },
    {
      "epoch": 199.67,
      "step": 200,
      "total_flos": 1984251671085056.0,
      "train_loss": 1.8223367309570313,
      "train_runtime": 378.2401,
      "train_samples_per_second": 10.047,
      "train_steps_per_second": 0.529
    }
  ],
  "max_steps": 200,
  "num_train_epochs": 200,
  "total_flos": 1984251671085056.0,
  "trial_name": null,
  "trial_params": null
}