trainer_state.json

{
  "best_global_step": 4000,
  "best_metric": 61.04808211777417,
  "best_model_checkpoint": "./whisper-base-en/checkpoint-4000",
  "epoch": 40.0,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 5.0,
      "grad_norm": 8.189814567565918,
      "learning_rate": 9.9e-06,
      "loss": 1.1053,
      "step": 500
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.7731186151504517,
      "eval_runtime": 124.1426,
      "eval_samples_per_second": 1.603,
      "eval_steps_per_second": 1.603,
      "eval_wer": 69.74608319827121,
      "step": 500
    },
    {
      "epoch": 10.0,
      "grad_norm": 6.20778751373291,
      "learning_rate": 9.658620689655173e-06,
      "loss": 0.1669,
      "step": 1000
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.890454113483429,
      "eval_runtime": 117.5449,
      "eval_samples_per_second": 1.693,
      "eval_steps_per_second": 1.693,
      "eval_wer": 65.09994597514857,
      "step": 1000
    },
    {
      "epoch": 15.0,
      "grad_norm": 2.8285183906555176,
      "learning_rate": 9.313793103448276e-06,
      "loss": 0.0208,
      "step": 1500
    },
    {
      "epoch": 15.0,
      "eval_loss": 1.1541786193847656,
      "eval_runtime": 117.6046,
      "eval_samples_per_second": 1.692,
      "eval_steps_per_second": 1.692,
      "eval_wer": 62.452728254997304,
      "step": 1500
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.5594582557678223,
      "learning_rate": 8.96896551724138e-06,
      "loss": 0.0045,
      "step": 2000
    },
    {
      "epoch": 20.0,
      "eval_loss": 1.191721796989441,
      "eval_runtime": 118.3942,
      "eval_samples_per_second": 1.681,
      "eval_steps_per_second": 1.681,
      "eval_wer": 62.452728254997304,
      "step": 2000
    },
    {
      "epoch": 25.0,
      "grad_norm": 0.05242369323968887,
      "learning_rate": 8.624137931034482e-06,
      "loss": 0.0019,
      "step": 2500
    },
    {
      "epoch": 25.0,
      "eval_loss": 1.2775946855545044,
      "eval_runtime": 117.3796,
      "eval_samples_per_second": 1.695,
      "eval_steps_per_second": 1.695,
      "eval_wer": 61.85845488924906,
      "step": 2500
    },
    {
      "epoch": 30.0,
      "grad_norm": 10.651979446411133,
      "learning_rate": 8.279310344827587e-06,
      "loss": 0.0029,
      "step": 3000
    },
    {
      "epoch": 30.0,
      "eval_loss": 1.2370944023132324,
      "eval_runtime": 119.3897,
      "eval_samples_per_second": 1.667,
      "eval_steps_per_second": 1.667,
      "eval_wer": 65.53214478660183,
      "step": 3000
    },
    {
      "epoch": 35.0,
      "grad_norm": 1.879603385925293,
      "learning_rate": 7.934482758620691e-06,
      "loss": 0.0052,
      "step": 3500
    },
    {
      "epoch": 35.0,
      "eval_loss": 1.2851910591125488,
      "eval_runtime": 118.7346,
      "eval_samples_per_second": 1.676,
      "eval_steps_per_second": 1.676,
      "eval_wer": 62.02052944354403,
      "step": 3500
    },
    {
      "epoch": 40.0,
      "grad_norm": 2.0357964038848877,
      "learning_rate": 7.589655172413793e-06,
      "loss": 0.0014,
      "step": 4000
    },
    {
      "epoch": 40.0,
      "eval_loss": 1.3512685298919678,
      "eval_runtime": 119.2661,
      "eval_samples_per_second": 1.669,
      "eval_steps_per_second": 1.669,
      "eval_wer": 61.04808211777417,
      "step": 4000
    }
  ],
  "logging_steps": 500,
  "max_steps": 15000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 150,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.1484453543936e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}