trainer_state.json

{
  "best_metric": 0.9397242813537718,
  "best_model_checkpoint": "model_saves/electra-large_spell_10k_2_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": 0.9397242813537718,
      "eval_loss": 0.44245752692222595,
      "eval_runtime": 3.4937,
      "eval_samples_per_second": 1254.269,
      "eval_steps_per_second": 10.018,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9393609051979561,
      "eval_loss": 0.4512821435928345,
      "eval_runtime": 3.4926,
      "eval_samples_per_second": 1254.636,
      "eval_steps_per_second": 10.021,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9392267355404242,
      "eval_loss": 0.4718455374240875,
      "eval_runtime": 3.5,
      "eval_samples_per_second": 1251.986,
      "eval_steps_per_second": 10.0,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3481,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.937722917295587,
      "eval_loss": 0.49437105655670166,
      "eval_runtime": 3.4985,
      "eval_samples_per_second": 1252.54,
      "eval_steps_per_second": 10.004,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.4139683831283712e+16,
      "train_loss": 0.34312588243342157,
      "train_runtime": 322.8928,
      "train_samples_per_second": 1593.594,
      "train_steps_per_second": 12.45
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.4139683831283712e+16,
  "trial_name": null,
  "trial_params": null
}