{
  "best_metric": 0.8986041101202016,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-42/checkpoint-1935",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1935,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.0985572338104248,
      "learning_rate": 0.00015144713450034053,
      "loss": 0.3039,
      "step": 645
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8945327646374563,
      "eval_f1": 0.19287833827893175,
      "eval_loss": 0.2821837365627289,
      "eval_mcc": 0.23408659817261018,
      "eval_precision": 0.6190476190476191,
      "eval_recall": 0.11423550087873462,
      "eval_runtime": 9.9917,
      "eval_samples_per_second": 516.228,
      "eval_steps_per_second": 16.213,
      "step": 645
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5096140503883362,
      "learning_rate": 0.0001135853508752554,
      "loss": 0.2797,
      "step": 1290
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8972469949592865,
      "eval_f1": 0.2182890855457227,
      "eval_loss": 0.27733176946640015,
      "eval_mcc": 0.26666876109371096,
      "eval_precision": 0.6788990825688074,
      "eval_recall": 0.13005272407732865,
      "eval_runtime": 9.9433,
      "eval_samples_per_second": 518.739,
      "eval_steps_per_second": 16.292,
      "step": 1290
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6613279581069946,
      "learning_rate": 7.572356725017026e-05,
      "loss": 0.2739,
      "step": 1935
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8986041101202016,
      "eval_f1": 0.2015267175572519,
      "eval_loss": 0.2698163688182831,
      "eval_mcc": 0.27313390239686397,
      "eval_precision": 0.7674418604651163,
      "eval_recall": 0.11599297012302284,
      "eval_runtime": 9.9136,
      "eval_samples_per_second": 520.296,
      "eval_steps_per_second": 16.341,
      "step": 1935
    }
  ],
  "logging_steps": 500,
  "max_steps": 3225,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 4760827597800.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8695064116191267,
    "learning_rate": 0.00018930891812542566,
    "num_train_epochs": 5,
    "temperature": 14
  }
}