trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.010559662090813094,
  "eval_steps": 10,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 1.0548523206751055e-06,
      "loss": 4.6245,
      "step": 10
    },
    {
      "epoch": 0.01,
      "eval_accuracy": 0.00943894389438944,
      "eval_loss": 4.621488094329834,
      "eval_runtime": 140.7774,
      "eval_samples_per_second": 107.617,
      "eval_steps_per_second": 6.727,
      "step": 10
    }
  ],
  "logging_steps": 10,
  "max_steps": 4735,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 4.9638879166464e+16,
  "trial_name": null,
  "trial_params": null
}