trainer_state.json

{
  "best_metric": 0.9394894844530909,
  "best_model_checkpoint": "model_saves/electra-large_spell_5k_3_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": 0.9394894844530909,
      "eval_loss": 0.43710747361183167,
      "eval_runtime": 3.432,
      "eval_samples_per_second": 1276.817,
      "eval_steps_per_second": 10.198,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.938913673006183,
      "eval_loss": 0.4479462206363678,
      "eval_runtime": 3.4391,
      "eval_samples_per_second": 1274.157,
      "eval_steps_per_second": 10.177,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9375943380404521,
      "eval_loss": 0.4736363887786865,
      "eval_runtime": 3.4359,
      "eval_samples_per_second": 1275.346,
      "eval_steps_per_second": 10.186,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3205,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9370856114223102,
      "eval_loss": 0.49852246046066284,
      "eval_runtime": 3.8609,
      "eval_samples_per_second": 1134.962,
      "eval_steps_per_second": 9.065,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3909022142889984e+16,
      "train_loss": 0.31548746486208334,
      "train_runtime": 330.9383,
      "train_samples_per_second": 1554.852,
      "train_steps_per_second": 12.147
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3909022142889984e+16,
  "trial_name": null,
  "trial_params": null
}