run-17/checkpoint-534/trainer_state.json

{
  "best_metric": 0.603742778301239,
  "best_model_checkpoint": "BERT-WMM/run-17/checkpoint-534",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 534,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7530516431924883,
      "eval_loss": 0.6066157817840576,
      "eval_runtime": 2.6354,
      "eval_samples_per_second": 808.213,
      "eval_steps_per_second": 50.845,
      "step": 267
    },
    {
      "epoch": 1.87,
      "grad_norm": 9.448416709899902,
      "learning_rate": 9.825065893169892e-07,
      "loss": 0.6083,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7615023474178404,
      "eval_loss": 0.603742778301239,
      "eval_runtime": 2.6051,
      "eval_samples_per_second": 817.639,
      "eval_steps_per_second": 51.438,
      "step": 534
    }
  ],
  "logging_steps": 500,
  "max_steps": 534,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 349221394826640.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "weight_decay": 0.015498593504464519
  }
}