trainer_state.json

{
  "best_metric": 0.9420163463366094,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-spell_5k_4_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": 0.9420163463366094,
      "eval_loss": 0.4153633117675781,
      "eval_runtime": 4.6508,
      "eval_samples_per_second": 942.213,
      "eval_steps_per_second": 7.526,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9409653506859423,
      "eval_loss": 0.44057878851890564,
      "eval_runtime": 4.6472,
      "eval_samples_per_second": 942.934,
      "eval_steps_per_second": 7.531,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9407137825780699,
      "eval_loss": 0.4833277463912964,
      "eval_runtime": 4.6824,
      "eval_samples_per_second": 935.835,
      "eval_steps_per_second": 7.475,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9396348349154172,
      "eval_loss": 0.5351554155349731,
      "eval_runtime": 4.6531,
      "eval_samples_per_second": 941.742,
      "eval_steps_per_second": 7.522,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.60866272018432e+16,
      "train_loss": 0.2470646803058795,
      "train_runtime": 467.3854,
      "train_samples_per_second": 1100.933,
      "train_steps_per_second": 8.601
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.60866272018432e+16,
  "trial_name": null,
  "trial_params": null
}