{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 1085,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 4.9953917050691245e-05,
      "loss": 11.8912,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_cer": 0.07741935483870968,
      "eval_loss": 0.23858200013637543,
      "eval_runtime": 140.9135,
      "eval_samples_per_second": 0.979,
      "eval_steps_per_second": 0.128,
      "step": 217
    },
    {
      "epoch": 2.0,
      "eval_cer": 0.04516129032258064,
      "eval_loss": 0.15434867143630981,
      "eval_runtime": 37.8243,
      "eval_samples_per_second": 3.648,
      "eval_steps_per_second": 0.476,
      "step": 434
    },
    {
      "epoch": 2.3,
      "learning_rate": 2.6958525345622122e-05,
      "loss": 0.4293,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_cer": 0.01935483870967742,
      "eval_loss": 0.10877370089292526,
      "eval_runtime": 35.4493,
      "eval_samples_per_second": 3.893,
      "eval_steps_per_second": 0.508,
      "step": 651
    },
    {
      "epoch": 4.0,
      "eval_cer": 0.03225806451612903,
      "eval_loss": 0.12325212359428406,
      "eval_runtime": 35.8699,
      "eval_samples_per_second": 3.847,
      "eval_steps_per_second": 0.502,
      "step": 868
    },
    {
      "epoch": 4.61,
      "learning_rate": 3.9170506912442395e-06,
      "loss": 0.0775,
      "step": 1000
    },
    {
      "epoch": 5.0,
      "eval_cer": 0.016129032258064516,
      "eval_loss": 0.10414960980415344,
      "eval_runtime": 37.8862,
      "eval_samples_per_second": 3.642,
      "eval_steps_per_second": 0.475,
      "step": 1085
    },
    {
      "epoch": 5.0,
      "step": 1085,
      "total_flos": 6.483891179980063e+18,
      "train_loss": 0.246805760827482,
      "train_runtime": 4287.6356,
      "train_samples_per_second": 2.021,
      "train_steps_per_second": 0.253
    }
  ],
  "max_steps": 1085,
  "num_train_epochs": 5,
  "total_flos": 6.483891179980063e+18,
  "trial_name": null,
  "trial_params": null
}