trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "global_step": 548,
  "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": 1.3156,
      "step": 274
    },
    {
      "epoch": 1.0,
      "eval_exact_match": 73.04418794450274,
      "eval_f1": 76.04190236033945,
      "eval_runtime": 9.556,
      "eval_samples_per_second": 897.555,
      "eval_steps_per_second": 2.825,
      "step": 274
    },
    {
      "epoch": 2.0,
      "learning_rate": 3e-05,
      "loss": 0.6645,
      "step": 548
    },
    {
      "epoch": 2.0,
      "eval_exact_match": 72.82266526757607,
      "eval_f1": 76.49070114778583,
      "eval_runtime": 7.8341,
      "eval_samples_per_second": 1094.825,
      "eval_steps_per_second": 3.446,
      "step": 548
    }
  ],
  "max_steps": 2740,
  "num_train_epochs": 10,
  "total_flos": 173999594143744.0,
  "trial_name": null,
  "trial_params": null
}