{
  "best_metric": 0.7396078431372549,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-32/checkpoint-480",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 480,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.5651652812957764,
      "learning_rate": 0.0003078789246710064,
      "loss": 0.4451,
      "step": 160
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6886274509803921,
      "eval_f1": 0.22912621359223304,
      "eval_loss": 0.40980812907218933,
      "eval_mcc": 0.18837524271387296,
      "eval_precision": 0.6555555555555556,
      "eval_recall": 0.1388235294117647,
      "eval_runtime": 1.8755,
      "eval_samples_per_second": 679.803,
      "eval_steps_per_second": 21.327,
      "step": 160
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.201909065246582,
      "learning_rate": 0.00026939405908713064,
      "loss": 0.4166,
      "step": 320
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7231372549019608,
      "eval_f1": 0.400679117147708,
      "eval_loss": 0.40082836151123047,
      "eval_mcc": 0.3147455869861136,
      "eval_precision": 0.7195121951219512,
      "eval_recall": 0.2776470588235294,
      "eval_runtime": 1.915,
      "eval_samples_per_second": 665.803,
      "eval_steps_per_second": 20.888,
      "step": 320
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.4556416273117065,
      "learning_rate": 0.00023090919350325482,
      "loss": 0.4092,
      "step": 480
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7396078431372549,
      "eval_f1": 0.4860681114551084,
      "eval_loss": 0.39845094084739685,
      "eval_mcc": 0.36627608925493965,
      "eval_precision": 0.7104072398190046,
      "eval_recall": 0.36941176470588233,
      "eval_runtime": 1.8771,
      "eval_samples_per_second": 679.255,
      "eval_steps_per_second": 21.31,
      "step": 480
    }
  ],
  "logging_steps": 500,
  "max_steps": 1440,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "total_flos": 875266313760.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6927304568601392,
    "learning_rate": 0.00034636379025488223,
    "num_train_epochs": 9,
    "temperature": 14
  }
}