trainer_state.json

{
  "best_metric": 61.04375329467581,
  "best_model_checkpoint": "./finam2/checkpoint-500",
  "epoch": 7.246376811594203,
  "eval_steps": 250,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.4492753623188406,
      "grad_norm": 94431.8984375,
      "learning_rate": 2.5e-05,
      "loss": 0.1608,
      "step": 100
    },
    {
      "epoch": 2.898550724637681,
      "grad_norm": 82763.5,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.0404,
      "step": 200
    },
    {
      "epoch": 3.6231884057971016,
      "eval_loss": 0.016866113990545273,
      "eval_runtime": 160.3369,
      "eval_samples_per_second": 1.279,
      "eval_steps_per_second": 0.162,
      "eval_wer": 87.55930416447022,
      "step": 250
    },
    {
      "epoch": 4.3478260869565215,
      "grad_norm": 53642.96875,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 0.0116,
      "step": 300
    },
    {
      "epoch": 5.797101449275362,
      "grad_norm": 53491.53125,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0059,
      "step": 400
    },
    {
      "epoch": 7.246376811594203,
      "grad_norm": 24437.5546875,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.0035,
      "step": 500
    },
    {
      "epoch": 7.246376811594203,
      "eval_loss": 0.0023117074742913246,
      "eval_runtime": 186.2811,
      "eval_samples_per_second": 1.1,
      "eval_steps_per_second": 0.14,
      "eval_wer": 61.04375329467581,
      "step": 500
    }
  ],
  "logging_steps": 100,
  "max_steps": 1000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 250,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.60322572050432e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}