trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 41.666666666666664,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 20.83,
      "learning_rate": 0.0001512,
      "loss": 1.9115,
      "step": 500
    },
    {
      "epoch": 20.83,
      "eval_loss": 1.540008544921875,
      "eval_runtime": 5.0397,
      "eval_samples_per_second": 28.177,
      "eval_steps_per_second": 3.572,
      "eval_wer": 0.7279860505666957,
      "step": 500
    },
    {
      "epoch": 41.67,
      "learning_rate": 1.2e-06,
      "loss": 0.1155,
      "step": 1000
    },
    {
      "epoch": 41.67,
      "eval_loss": 1.5613987445831299,
      "eval_runtime": 5.0481,
      "eval_samples_per_second": 28.129,
      "eval_steps_per_second": 3.566,
      "eval_wer": 0.6765475152571927,
      "step": 1000
    },
    {
      "epoch": 41.67,
      "step": 1000,
      "total_flos": 2.318457048084749e+18,
      "train_loss": 1.0135024032592774,
      "train_runtime": 1622.9478,
      "train_samples_per_second": 9.859,
      "train_steps_per_second": 0.616
    }
  ],
  "max_steps": 1000,
  "num_train_epochs": 42,
  "total_flos": 2.318457048084749e+18,
  "trial_name": null,
  "trial_params": null
}