{
  "best_metric": 0.603524923324585,
  "best_model_checkpoint": "BERT-WMM/run-10/checkpoint-533",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1599,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.94,
      "grad_norm": 11.3766508102417,
      "learning_rate": 1.0605888092670128e-05,
      "loss": 0.7117,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7568075117370892,
      "eval_loss": 0.603524923324585,
      "eval_runtime": 2.4565,
      "eval_samples_per_second": 867.098,
      "eval_steps_per_second": 54.55,
      "step": 533
    },
    {
      "epoch": 1.88,
      "grad_norm": 19.02298927307129,
      "learning_rate": 5.780643282538131e-06,
      "loss": 0.4384,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7492957746478873,
      "eval_loss": 0.6247430443763733,
      "eval_runtime": 2.4597,
      "eval_samples_per_second": 865.958,
      "eval_steps_per_second": 54.478,
      "step": 1066
    },
    {
      "epoch": 2.81,
      "grad_norm": 25.672210693359375,
      "learning_rate": 9.553984724061353e-07,
      "loss": 0.2896,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7507042253521127,
      "eval_loss": 0.7414140701293945,
      "eval_runtime": 2.4706,
      "eval_samples_per_second": 862.126,
      "eval_steps_per_second": 54.237,
      "step": 1599
    }
  ],
  "logging_steps": 500,
  "max_steps": 1599,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 507646505902536.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "per_device_train_batch_size": 16
  }
}