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": 1.8697,
      "step": 275
    },
    {
      "epoch": 1.0,
      "eval_exact_match": 51.29998834091174,
      "eval_f1": 60.31539060889624,
      "eval_runtime": 3.6798,
      "eval_samples_per_second": 2330.843,
      "eval_steps_per_second": 7.337,
      "step": 275
    },
    {
      "epoch": 2.0,
      "learning_rate": 3e-05,
      "loss": 1.0286,
      "step": 550
    },
    {
      "epoch": 2.0,
      "eval_exact_match": 52.03451090124752,
      "eval_f1": 60.730837242513196,
      "eval_runtime": 3.0428,
      "eval_samples_per_second": 2818.822,
      "eval_steps_per_second": 8.874,
      "step": 550
    }
  ],
  "max_steps": 2750,
  "num_train_epochs": 10,
  "total_flos": 66106662322176.0,
  "trial_name": null,
  "trial_params": null
}