run-4/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.9071100917431193,
  "best_model_checkpoint": "electra-distilled-sst\\run-4\\checkpoint-1054",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1054,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 26.105838775634766,
      "learning_rate": 2.1019516058113692e-05,
      "loss": 1.7548,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8922018348623854,
      "eval_loss": 0.9411500692367554,
      "eval_runtime": 1.4396,
      "eval_samples_per_second": 605.724,
      "eval_steps_per_second": 4.862,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 1.1942906851200962e-07,
      "loss": 0.93,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9071100917431193,
      "eval_loss": 0.9139881730079651,
      "eval_runtime": 1.4435,
      "eval_samples_per_second": 604.075,
      "eval_steps_per_second": 4.849,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 1054,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 373776249469380.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.4067254850025479,
    "learning_rate": 4.195941273721938e-05,
    "num_train_epochs": 2,
    "temperature": 4
  }
}