{
  "best_metric": 0.6574257425742575,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-32/checkpoint-891",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 891,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.0489678382873535,
      "learning_rate": 0.00030782109688346106,
      "loss": 0.6591,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5366336633663367,
      "eval_f1": 0.20945945945945946,
      "eval_loss": 0.6587525606155396,
      "eval_mcc": 0.12699706368366073,
      "eval_precision": 0.7045454545454546,
      "eval_recall": 0.12301587301587301,
      "eval_runtime": 0.9347,
      "eval_samples_per_second": 540.253,
      "eval_steps_per_second": 17.117,
      "step": 297
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.570565104484558,
      "learning_rate": 0.00024625687750676884,
      "loss": 0.6201,
      "step": 594
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6514851485148515,
      "eval_f1": 0.7095709570957095,
      "eval_loss": 0.6164775490760803,
      "eval_mcc": 0.33175059993848016,
      "eval_precision": 0.6073446327683616,
      "eval_recall": 0.8531746031746031,
      "eval_runtime": 0.94,
      "eval_samples_per_second": 537.23,
      "eval_steps_per_second": 17.021,
      "step": 594
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.9003602266311646,
      "learning_rate": 0.00018469265813007664,
      "loss": 0.6,
      "step": 891
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6574257425742575,
      "eval_f1": 0.7131011608623549,
      "eval_loss": 0.6146110892295837,
      "eval_mcc": 0.3427822124548935,
      "eval_precision": 0.6125356125356125,
      "eval_recall": 0.8531746031746031,
      "eval_runtime": 0.9324,
      "eval_samples_per_second": 541.638,
      "eval_steps_per_second": 17.161,
      "step": 891
    }
  ],
  "logging_steps": 500,
  "max_steps": 1782,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 500,
  "total_flos": 2192103797940.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9361124407975862,
    "learning_rate": 0.0003693853162601533,
    "num_train_epochs": 6,
    "temperature": 21
  }
}