trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.997601918465228,
  "global_step": 520,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.96,
      "learning_rate": 6.533333333333334e-06,
      "loss": 11.216,
      "step": 100
    },
    {
      "epoch": 1.92,
      "learning_rate": 1.3200000000000002e-05,
      "loss": 3.4158,
      "step": 200
    },
    {
      "epoch": 2.88,
      "learning_rate": 1.9866666666666667e-05,
      "loss": 2.4363,
      "step": 300
    },
    {
      "epoch": 3.84,
      "learning_rate": 1.1090909090909092e-05,
      "loss": 1.6274,
      "step": 400
    },
    {
      "epoch": 4.81,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 1.3883,
      "step": 500
    },
    {
      "epoch": 4.81,
      "eval_loss": 0.6559479236602783,
      "eval_runtime": 61.7219,
      "eval_samples_per_second": 18.016,
      "eval_steps_per_second": 4.504,
      "eval_wer": 0.7843946188340807,
      "step": 500
    },
    {
      "epoch": 5.0,
      "step": 520,
      "total_flos": 7.17353684308089e+18,
      "train_loss": 3.913738654210017,
      "train_runtime": 2210.4533,
      "train_samples_per_second": 7.546,
      "train_steps_per_second": 0.235
    }
  ],
  "max_steps": 520,
  "num_train_epochs": 5,
  "total_flos": 7.17353684308089e+18,
  "trial_name": null,
  "trial_params": null
}