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": 1.8713,
      "step": 275
    },
    {
      "epoch": 1.0,
      "eval_exact_match": 47.89553456919669,
      "eval_f1": 57.54505882019725,
      "eval_runtime": 3.9244,
      "eval_samples_per_second": 2185.544,
      "eval_steps_per_second": 6.88,
      "step": 275
    },
    {
      "epoch": 2.0,
      "learning_rate": 3e-05,
      "loss": 1.0257,
      "step": 550
    },
    {
      "epoch": 2.0,
      "eval_exact_match": 49.096420659904396,
      "eval_f1": 57.873644765414554,
      "eval_runtime": 3.1541,
      "eval_samples_per_second": 2719.347,
      "eval_steps_per_second": 8.56,
      "step": 550
    }
  ],
  "max_steps": 2750,
  "num_train_epochs": 10,
  "total_flos": 66106662322176.0,
  "trial_name": null,
  "trial_params": null
}