{
  "best_metric": 0.7094317516110135,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-642",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 642,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 3.8868563175201416,
      "learning_rate": 0.00018706603069859162,
      "loss": 0.3842,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6848271821909784,
      "eval_f1": 0.1487341772151899,
      "eval_loss": 0.3658275604248047,
      "eval_mcc": 0.17137932324690125,
      "eval_precision": 0.746031746031746,
      "eval_recall": 0.08260105448154657,
      "eval_runtime": 3.1455,
      "eval_samples_per_second": 542.68,
      "eval_steps_per_second": 17.167,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.3482353687286377,
      "learning_rate": 0.0001247106871323944,
      "loss": 0.3585,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6877562975981254,
      "eval_f1": 0.14446227929373995,
      "eval_loss": 0.35682281851768494,
      "eval_mcc": 0.19170644527903724,
      "eval_precision": 0.8333333333333334,
      "eval_recall": 0.07908611599297012,
      "eval_runtime": 3.1465,
      "eval_samples_per_second": 542.514,
      "eval_steps_per_second": 17.162,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.6936368942260742,
      "learning_rate": 6.23553435661972e-05,
      "loss": 0.3507,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7094317516110135,
      "eval_f1": 0.2811594202898551,
      "eval_loss": 0.3534710705280304,
      "eval_mcc": 0.27440350027007715,
      "eval_precision": 0.8016528925619835,
      "eval_recall": 0.1704745166959578,
      "eval_runtime": 3.1505,
      "eval_samples_per_second": 541.819,
      "eval_steps_per_second": 17.14,
      "step": 642
    }
  ],
  "logging_steps": 500,
  "max_steps": 856,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 1574326993320.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6014075049725551,
    "learning_rate": 0.0002494213742647888,
    "num_train_epochs": 4,
    "temperature": 26
  }
}