trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 41.666666666666664,
  "eval_steps": 250,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 20.833333333333332,
      "grad_norm": 0.165288046002388,
      "learning_rate": 4.92e-06,
      "loss": 0.3226,
      "step": 250
    },
    {
      "epoch": 20.833333333333332,
      "eval_loss": 1.0812960863113403,
      "eval_runtime": 20.5975,
      "eval_samples_per_second": 4.127,
      "eval_steps_per_second": 0.146,
      "eval_wer": 0.7583774250440917,
      "step": 250
    },
    {
      "epoch": 41.666666666666664,
      "grad_norm": 0.01815676875412464,
      "learning_rate": 9.920000000000002e-06,
      "loss": 0.0021,
      "step": 500
    },
    {
      "epoch": 41.666666666666664,
      "eval_loss": 1.2795981168746948,
      "eval_runtime": 20.8995,
      "eval_samples_per_second": 4.067,
      "eval_steps_per_second": 0.144,
      "eval_wer": 0.763668430335097,
      "step": 500
    },
    {
      "epoch": 41.666666666666664,
      "step": 500,
      "total_flos": 4.617366645252293e+18,
      "train_loss": 0.16236594557762146,
      "train_runtime": 2555.1718,
      "train_samples_per_second": 6.262,
      "train_steps_per_second": 0.196
    }
  ],
  "logging_steps": 250,
  "max_steps": 500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 42,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.617366645252293e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}