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": 1.3978,
      "step": 275
    },
    {
      "epoch": 1.0,
      "eval_exact_match": 82.16650898770104,
      "eval_f1": 89.74449498371443,
      "eval_runtime": 10.8365,
      "eval_samples_per_second": 980.855,
      "eval_steps_per_second": 3.138,
      "step": 275
    },
    {
      "epoch": 2.0,
      "learning_rate": 3e-05,
      "loss": 0.7654,
      "step": 550
    },
    {
      "epoch": 2.0,
      "eval_exact_match": 83.79375591296122,
      "eval_f1": 90.73612532740755,
      "eval_runtime": 9.414,
      "eval_samples_per_second": 1129.065,
      "eval_steps_per_second": 3.612,
      "step": 550
    }
  ],
  "max_steps": 5500,
  "num_train_epochs": 20,
  "total_flos": 174158642151424.0,
  "trial_name": null,
  "trial_params": null
}