trainer_state.json

{
  "best_metric": 0.9399870302664386,
  "best_model_checkpoint": "model_saves/electra-large_spell_5k_5_p3/checkpoint-268",
  "epoch": 4.0,
  "global_step": 1072,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9399870302664386,
      "eval_loss": 0.43308553099632263,
      "eval_runtime": 4.7345,
      "eval_samples_per_second": 925.552,
      "eval_steps_per_second": 7.393,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9393329531859703,
      "eval_loss": 0.44236716628074646,
      "eval_runtime": 4.5996,
      "eval_samples_per_second": 952.69,
      "eval_steps_per_second": 7.609,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9391987835284383,
      "eval_loss": 0.4650343954563141,
      "eval_runtime": 4.1,
      "eval_samples_per_second": 1068.789,
      "eval_steps_per_second": 8.537,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3503,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9382931383400978,
      "eval_loss": 0.4915383756160736,
      "eval_runtime": 3.5395,
      "eval_samples_per_second": 1238.012,
      "eval_steps_per_second": 9.888,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3900286313627648e+16,
      "train_loss": 0.34482939029807474,
      "train_runtime": 412.9972,
      "train_samples_per_second": 1245.916,
      "train_steps_per_second": 9.734
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3900286313627648e+16,
  "trial_name": null,
  "trial_params": null
}