run-15/checkpoint-1066/trainer_state.json

{
  "best_metric": 0.6102677583694458,
  "best_model_checkpoint": "BERT-WMM/run-15/checkpoint-1066",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1066,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.94,
      "grad_norm": 8.692239761352539,
      "learning_rate": 1.3731493717417162e-05,
      "loss": 0.7076,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7530516431924883,
      "eval_loss": 0.6211019158363342,
      "eval_runtime": 1.9926,
      "eval_samples_per_second": 1068.944,
      "eval_steps_per_second": 67.248,
      "step": 533
    },
    {
      "epoch": 1.88,
      "grad_norm": 11.677574157714844,
      "learning_rate": 1.0560248055196108e-05,
      "loss": 0.4535,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7680751173708921,
      "eval_loss": 0.6102677583694458,
      "eval_runtime": 2.2362,
      "eval_samples_per_second": 952.498,
      "eval_steps_per_second": 59.922,
      "step": 1066
    }
  ],
  "logging_steps": 500,
  "max_steps": 2665,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 348929503378848.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.6902739379638215e-05,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 16,
    "weight_decay": 0.019154977870537472
  }
}