trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.79,
      "eval_loss": 7.543250560760498,
      "eval_runtime": 10.8753,
      "eval_samples_per_second": 18.206,
      "eval_steps_per_second": 2.299,
      "eval_wer": 1.0,
      "step": 100
    },
    {
      "epoch": 3.57,
      "eval_loss": 4.197996616363525,
      "eval_runtime": 10.0632,
      "eval_samples_per_second": 19.676,
      "eval_steps_per_second": 2.484,
      "eval_wer": 1.0,
      "step": 200
    },
    {
      "epoch": 5.0,
      "step": 280,
      "total_flos": 3.62506981246848e+16,
      "train_loss": 4.910877336774553,
      "train_runtime": 117.8697,
      "train_samples_per_second": 9.375,
      "train_steps_per_second": 2.376
    }
  ],
  "max_steps": 280,
  "num_train_epochs": 5,
  "total_flos": 3.62506981246848e+16,
  "trial_name": null,
  "trial_params": null
}