trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9957805907173,
  "global_step": 354,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.84,
      "eval_loss": 4.014804363250732,
      "eval_runtime": 71.3844,
      "eval_samples_per_second": 24.949,
      "eval_steps_per_second": 3.124,
      "eval_wer": 1.0,
      "step": 100
    },
    {
      "epoch": 1.69,
      "eval_loss": 3.1738100051879883,
      "eval_runtime": 71.3843,
      "eval_samples_per_second": 24.949,
      "eval_steps_per_second": 3.124,
      "eval_wer": 1.0,
      "step": 200
    },
    {
      "epoch": 2.54,
      "eval_loss": 2.5980231761932373,
      "eval_runtime": 71.3516,
      "eval_samples_per_second": 24.961,
      "eval_steps_per_second": 3.125,
      "eval_wer": 1.0236054517352817,
      "step": 300
    },
    {
      "epoch": 3.0,
      "step": 354,
      "total_flos": 1.5058464259703962e+18,
      "train_loss": 5.89979406669315,
      "train_runtime": 748.1359,
      "train_samples_per_second": 15.19,
      "train_steps_per_second": 0.473
    }
  ],
  "max_steps": 354,
  "num_train_epochs": 3,
  "total_flos": 1.5058464259703962e+18,
  "trial_name": null,
  "trial_params": null
}