{
  "best_metric": 0.8209393346379648,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-19/checkpoint-576",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 576,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.367055416107178,
      "learning_rate": 0.00018610347830365945,
      "loss": 0.5762,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7778864970645792,
      "eval_f1": 0.7781036168132942,
      "eval_loss": 0.4806343615055084,
      "eval_mcc": 0.5557740583380701,
      "eval_precision": 0.77734375,
      "eval_recall": 0.7788649706457925,
      "eval_runtime": 69.8858,
      "eval_samples_per_second": 14.624,
      "eval_steps_per_second": 0.458,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.6895294189453125,
      "learning_rate": 0.00012406898553577296,
      "loss": 0.4725,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8189823874755382,
      "eval_f1": 0.8304307974335472,
      "eval_loss": 0.42346441745758057,
      "eval_mcc": 0.6438615171758961,
      "eval_precision": 0.7810344827586206,
      "eval_recall": 0.8864970645792564,
      "eval_runtime": 67.0224,
      "eval_samples_per_second": 15.249,
      "eval_steps_per_second": 0.477,
      "step": 384
    },
    {
      "epoch": 3.0,
      "grad_norm": 10.282877922058105,
      "learning_rate": 6.203449276788648e-05,
      "loss": 0.4208,
      "step": 576
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8209393346379648,
      "eval_f1": 0.8325709057639525,
      "eval_loss": 0.4079805314540863,
      "eval_mcc": 0.6481656523622785,
      "eval_precision": 0.781786941580756,
      "eval_recall": 0.8904109589041096,
      "eval_runtime": 66.8138,
      "eval_samples_per_second": 15.296,
      "eval_steps_per_second": 0.479,
      "step": 576
    }
  ],
  "logging_steps": 500,
  "max_steps": 768,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 707085591840.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9975127064980498,
    "learning_rate": 0.00024813797107154593,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 16,
    "temperature": 45
  }
}