trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 48.0,
  "global_step": 1200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 16.0,
      "learning_rate": 8e-05,
      "loss": 8.9546,
      "step": 400
    },
    {
      "epoch": 16.0,
      "eval_loss": 3.0705316066741943,
      "eval_runtime": 27.2617,
      "eval_samples_per_second": 7.336,
      "eval_wer": 0.99836867862969,
      "step": 400
    },
    {
      "epoch": 32.0,
      "learning_rate": 6.129032258064517e-05,
      "loss": 1.7778,
      "step": 800
    },
    {
      "epoch": 32.0,
      "eval_loss": 0.604729413986206,
      "eval_runtime": 27.2625,
      "eval_samples_per_second": 7.336,
      "eval_wer": 0.7781402936378466,
      "step": 800
    },
    {
      "epoch": 48.0,
      "learning_rate": 9.67741935483871e-06,
      "loss": 0.2996,
      "step": 1200
    },
    {
      "epoch": 48.0,
      "eval_loss": 0.49922919273376465,
      "eval_runtime": 28.4037,
      "eval_samples_per_second": 7.041,
      "eval_wer": 0.6753670473083198,
      "step": 1200
    }
  ],
  "max_steps": 1275,
  "num_train_epochs": 51,
  "total_flos": 3.8339975189263565e+18,
  "trial_name": null,
  "trial_params": null
}