trainer_state.json

{
  "best_metric": 0.939796956584935,
  "best_model_checkpoint": "model_saves/electra-large_spell_5k_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": 0.939796956584935,
      "eval_loss": 0.4383164048194885,
      "eval_runtime": 3.4417,
      "eval_samples_per_second": 1273.208,
      "eval_steps_per_second": 10.169,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9389751674325518,
      "eval_loss": 0.45297756791114807,
      "eval_runtime": 3.4252,
      "eval_samples_per_second": 1279.351,
      "eval_steps_per_second": 10.218,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9389192634085801,
      "eval_loss": 0.47665515542030334,
      "eval_runtime": 3.4416,
      "eval_samples_per_second": 1273.235,
      "eval_steps_per_second": 10.17,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3217,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9377676405147642,
      "eval_loss": 0.5029319524765015,
      "eval_runtime": 3.4298,
      "eval_samples_per_second": 1277.636,
      "eval_steps_per_second": 10.205,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.390784196640768e+16,
      "train_loss": 0.31709314460185034,
      "train_runtime": 326.687,
      "train_samples_per_second": 1575.086,
      "train_steps_per_second": 12.305
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.390784196640768e+16,
  "trial_name": null,
  "trial_params": null
}