trainer_state.json

{
  "best_metric": 1.6104975938796997,
  "best_model_checkpoint": "./outputdir/checkpoint-10",
  "epoch": 1.0,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 0.0,
      "loss": 1.6036,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.6104975938796997,
      "eval_runtime": 2007.0049,
      "eval_samples_per_second": 1.466,
      "eval_steps_per_second": 1.466,
      "step": 10
    }
  ],
  "max_steps": 10,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 2.988576972000461e+16,
  "trial_name": null,
  "trial_params": null
}