trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 48.78048780487805,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 12.2,
      "learning_rate": 0.0004200542005420054,
      "loss": 2.2288,
      "step": 1000
    },
    {
      "epoch": 12.2,
      "eval_loss": 0.4119809865951538,
      "eval_runtime": 39.8264,
      "eval_samples_per_second": 7.307,
      "eval_wer": 0.4587706146926537,
      "step": 1000
    },
    {
      "epoch": 24.39,
      "learning_rate": 0.0002845528455284553,
      "loss": 0.1307,
      "step": 2000
    },
    {
      "epoch": 24.39,
      "eval_loss": 0.46079644560813904,
      "eval_runtime": 39.8036,
      "eval_samples_per_second": 7.311,
      "eval_wer": 0.3800599700149925,
      "step": 2000
    },
    {
      "epoch": 36.59,
      "learning_rate": 0.00014905149051490515,
      "loss": 0.0645,
      "step": 3000
    },
    {
      "epoch": 36.59,
      "eval_loss": 0.4456443190574646,
      "eval_runtime": 41.0007,
      "eval_samples_per_second": 7.097,
      "eval_wer": 0.35007496251874065,
      "step": 3000
    },
    {
      "epoch": 48.78,
      "learning_rate": 1.3550135501355015e-05,
      "loss": 0.042,
      "step": 4000
    },
    {
      "epoch": 48.78,
      "eval_loss": 0.46975016593933105,
      "eval_runtime": 40.134,
      "eval_samples_per_second": 7.251,
      "eval_wer": 0.3365817091454273,
      "step": 4000
    }
  ],
  "max_steps": 4100,
  "num_train_epochs": 50,
  "total_flos": 1.961718571271133e+19,
  "trial_name": null,
  "trial_params": null
}