run-4/checkpoint-633/trainer_state.json

{
  "best_metric": 0.930045871559633,
  "best_model_checkpoint": "bert-base-uncased-finetuned-sst2/run-4/checkpoint-633",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 633,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8807339449541285,
      "eval_loss": 0.2860465347766876,
      "eval_runtime": 2.3416,
      "eval_samples_per_second": 372.397,
      "eval_steps_per_second": 23.488,
      "step": 211
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9025229357798165,
      "eval_loss": 0.30859723687171936,
      "eval_runtime": 2.3653,
      "eval_samples_per_second": 368.662,
      "eval_steps_per_second": 23.253,
      "step": 422
    },
    {
      "epoch": 2.37,
      "grad_norm": 0.23388604819774628,
      "learning_rate": 1.099172968867278e-05,
      "loss": 0.2134,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.930045871559633,
      "eval_loss": 0.3024921715259552,
      "eval_runtime": 2.3782,
      "eval_samples_per_second": 366.664,
      "eval_steps_per_second": 23.127,
      "step": 633
    }
  ],
  "logging_steps": 500,
  "max_steps": 633,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 336939914716260.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 5.23140217513524e-05,
    "num_train_epochs": 3,
    "per_device_train_batch_size": 32,
    "seed": 3
  }
}