trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "global_step": 550,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 3e-05,
      "loss": 1.5332,
      "step": 275
    },
    {
      "epoch": 1.0,
      "eval_exact_match": 81.12582781456953,
      "eval_f1": 88.8038314930431,
      "eval_runtime": 11.6663,
      "eval_samples_per_second": 910.825,
      "eval_steps_per_second": 2.914,
      "step": 275
    },
    {
      "epoch": 2.0,
      "learning_rate": 3e-05,
      "loss": 0.8065,
      "step": 550
    },
    {
      "epoch": 2.0,
      "eval_exact_match": 83.03689687795648,
      "eval_f1": 90.07424587925397,
      "eval_runtime": 9.3779,
      "eval_samples_per_second": 1133.087,
      "eval_steps_per_second": 3.626,
      "step": 550
    }
  ],
  "max_steps": 5500,
  "num_train_epochs": 20,
  "total_flos": 174079118147584.0,
  "trial_name": null,
  "trial_params": null
}