trainer_state.json

{
  "best_metric": 0.939763414170552,
  "best_model_checkpoint": "model_saves/electra-large_spell_10k_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": 0.939763414170552,
      "eval_loss": 0.44610655307769775,
      "eval_runtime": 3.4844,
      "eval_samples_per_second": 1257.615,
      "eval_steps_per_second": 10.045,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9390981562852894,
      "eval_loss": 0.457300066947937,
      "eval_runtime": 3.4792,
      "eval_samples_per_second": 1259.491,
      "eval_steps_per_second": 10.06,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9380080278178423,
      "eval_loss": 0.4832221567630768,
      "eval_runtime": 3.4817,
      "eval_samples_per_second": 1258.598,
      "eval_steps_per_second": 10.053,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.3206,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9375328436140833,
      "eval_loss": 0.5064626932144165,
      "eval_runtime": 3.4792,
      "eval_samples_per_second": 1259.483,
      "eval_steps_per_second": 10.06,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.414088420753408e+16,
      "train_loss": 0.3156794149484207,
      "train_runtime": 320.9081,
      "train_samples_per_second": 1603.45,
      "train_steps_per_second": 12.527
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.414088420753408e+16,
  "trial_name": null,
  "trial_params": null
}