{
  "best_metric": 0.8268101761252447,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-5/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": 6.338156223297119,
      "learning_rate": 0.00018913271286694248,
      "loss": 0.5602,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7974559686888454,
      "eval_f1": 0.8172992056487202,
      "eval_loss": 0.450337290763855,
      "eval_mcc": 0.6094644742211153,
      "eval_precision": 0.7443729903536977,
      "eval_recall": 0.9060665362035225,
      "eval_runtime": 66.9138,
      "eval_samples_per_second": 15.273,
      "eval_steps_per_second": 0.478,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.966625690460205,
      "learning_rate": 0.00012608847524462832,
      "loss": 0.4665,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.821917808219178,
      "eval_f1": 0.8289473684210525,
      "eval_loss": 0.41706037521362305,
      "eval_mcc": 0.6460214094282304,
      "eval_precision": 0.7974683544303798,
      "eval_recall": 0.863013698630137,
      "eval_runtime": 66.4735,
      "eval_samples_per_second": 15.375,
      "eval_steps_per_second": 0.481,
      "step": 384
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.679861545562744,
      "learning_rate": 6.304423762231416e-05,
      "loss": 0.4212,
      "step": 576
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8268101761252447,
      "eval_f1": 0.8389444949954504,
      "eval_loss": 0.42093411087989807,
      "eval_mcc": 0.6611696946007627,
      "eval_precision": 0.7840136054421769,
      "eval_recall": 0.9021526418786693,
      "eval_runtime": 66.3714,
      "eval_samples_per_second": 15.398,
      "eval_steps_per_second": 0.482,
      "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": 0.9715047486461332,
    "learning_rate": 0.00025217695048925664,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 16,
    "temperature": 28
  }
}