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": 1.3813,
      "step": 275
    },
    {
      "epoch": 1.0,
      "eval_exact_match": 82.18543046357615,
      "eval_f1": 89.77735911529885,
      "eval_runtime": 11.8006,
      "eval_samples_per_second": 900.716,
      "eval_steps_per_second": 2.881,
      "step": 275
    },
    {
      "epoch": 2.0,
      "learning_rate": 3e-05,
      "loss": 0.7629,
      "step": 550
    },
    {
      "epoch": 2.0,
      "eval_exact_match": 84.00189214758751,
      "eval_f1": 90.68769715646958,
      "eval_runtime": 9.4812,
      "eval_samples_per_second": 1121.062,
      "eval_steps_per_second": 3.586,
      "step": 550
    }
  ],
  "max_steps": 5500,
  "num_train_epochs": 20,
  "total_flos": 174158642151424.0,
  "trial_name": null,
  "trial_params": null
}