trainer_state.json

{
  "best_metric": 0.5285645127296448,
  "best_model_checkpoint": "./test-top-5-acc/checkpoint-40",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 40,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.825,
      "eval_loss": 0.5285645127296448,
      "eval_runtime": 2.4158,
      "eval_samples_per_second": 33.115,
      "eval_steps_per_second": 2.07,
      "step": 40
    },
    {
      "epoch": 1.0,
      "step": 40,
      "total_flos": 1.4897708531712e+17,
      "train_loss": 0.9101000785827636,
      "train_runtime": 81.6992,
      "train_samples_per_second": 7.834,
      "train_steps_per_second": 0.49
    }
  ],
  "logging_steps": 100,
  "max_steps": 40,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.4897708531712e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}