trainer_state.json

{
  "best_metric": 0.9412504612081978,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-spell_10k_1_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": 0.9412504612081978,
      "eval_loss": 0.4246947169303894,
      "eval_runtime": 4.6965,
      "eval_samples_per_second": 933.041,
      "eval_steps_per_second": 7.452,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.94108833953868,
      "eval_loss": 0.45120930671691895,
      "eval_runtime": 4.6963,
      "eval_samples_per_second": 933.071,
      "eval_steps_per_second": 7.453,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9405237088965663,
      "eval_loss": 0.4965062737464905,
      "eval_runtime": 4.6876,
      "eval_samples_per_second": 934.816,
      "eval_steps_per_second": 7.467,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9403839488366372,
      "eval_loss": 0.5336481928825378,
      "eval_runtime": 4.6888,
      "eval_samples_per_second": 934.575,
      "eval_steps_per_second": 7.465,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6315392326631424e+16,
      "train_loss": 0.24291392137755208,
      "train_runtime": 472.6427,
      "train_samples_per_second": 1088.687,
      "train_steps_per_second": 8.505
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6315392326631424e+16,
  "trial_name": null,
  "trial_params": null
}