trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 16,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 0,
      "loss": 3.5254,
      "step": 1
    },
    {
      "epoch": 1.0,
      "learning_rate": 5e-05,
      "loss": 3.2842,
      "step": 16
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.0373780487804878,
      "eval_loss": 3.044921875,
      "eval_runtime": 2.254,
      "eval_samples_per_second": 44.366,
      "eval_steps_per_second": 0.887,
      "step": 16
    },
    {
      "epoch": 1.0,
      "step": 16,
      "total_flos": 933007196160.0,
      "train_loss": 3.2992401123046875,
      "train_runtime": 24.3935,
      "train_samples_per_second": 40.994,
      "train_steps_per_second": 0.656
    }
  ],
  "max_steps": 16,
  "num_train_epochs": 1,
  "total_flos": 933007196160.0,
  "trial_name": null,
  "trial_params": null
}