{
  "best_metric": 0.7932044522554189,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-22/checkpoint-2568",
  "epoch": 12.0,
  "eval_steps": 500,
  "global_step": 2568,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 6.217133522033691,
      "learning_rate": 0.00032487441871616396,
      "loss": 0.5489,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7164616285881664,
      "eval_f1": 0.39800995024875624,
      "eval_loss": 0.5112811923027039,
      "eval_mcc": 0.2945528643838721,
      "eval_precision": 0.6808510638297872,
      "eval_recall": 0.281195079086116,
      "eval_runtime": 3.1646,
      "eval_samples_per_second": 539.397,
      "eval_steps_per_second": 17.064,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.936843991279602,
      "learning_rate": 0.00030777576509952376,
      "loss": 0.4958,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7363796133567663,
      "eval_f1": 0.4966442953020134,
      "eval_loss": 0.4881761372089386,
      "eval_mcc": 0.3597855187518067,
      "eval_precision": 0.683076923076923,
      "eval_recall": 0.39015817223198596,
      "eval_runtime": 3.9406,
      "eval_samples_per_second": 433.184,
      "eval_steps_per_second": 13.704,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.401139497756958,
      "learning_rate": 0.0002906771114828835,
      "loss": 0.481,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7246631517281781,
      "eval_f1": 0.35792349726775957,
      "eval_loss": 0.4936811029911041,
      "eval_mcc": 0.32418684508290907,
      "eval_precision": 0.803680981595092,
      "eval_recall": 0.23022847100175747,
      "eval_runtime": 3.1821,
      "eval_samples_per_second": 536.439,
      "eval_steps_per_second": 16.97,
      "step": 642
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.7682132720947266,
      "learning_rate": 0.00027357845786624336,
      "loss": 0.47,
      "step": 856
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7510251903925015,
      "eval_f1": 0.47595561035758327,
      "eval_loss": 0.4672105610370636,
      "eval_mcc": 0.4002104345910371,
      "eval_precision": 0.7975206611570248,
      "eval_recall": 0.3391915641476274,
      "eval_runtime": 3.3008,
      "eval_samples_per_second": 517.142,
      "eval_steps_per_second": 16.359,
      "step": 856
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.235007286071777,
      "learning_rate": 0.0002564798042496031,
      "loss": 0.4564,
      "step": 1070
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7639132981839485,
      "eval_f1": 0.6385650224215247,
      "eval_loss": 0.46502017974853516,
      "eval_mcc": 0.46359962281417727,
      "eval_precision": 0.652014652014652,
      "eval_recall": 0.6256590509666081,
      "eval_runtime": 3.1354,
      "eval_samples_per_second": 544.434,
      "eval_steps_per_second": 17.223,
      "step": 1070
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.1712186336517334,
      "learning_rate": 0.0002393811506329629,
      "loss": 0.4494,
      "step": 1284
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7650849443468073,
      "eval_f1": 0.6125603864734299,
      "eval_loss": 0.45874524116516113,
      "eval_mcc": 0.4509702487241507,
      "eval_precision": 0.6802575107296137,
      "eval_recall": 0.5571177504393673,
      "eval_runtime": 3.3104,
      "eval_samples_per_second": 515.646,
      "eval_steps_per_second": 16.312,
      "step": 1284
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.33981466293335,
      "learning_rate": 0.0002222824970163227,
      "loss": 0.4385,
      "step": 1498
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7721148213239601,
      "eval_f1": 0.5682574916759155,
      "eval_loss": 0.4540397524833679,
      "eval_mcc": 0.4563006637386888,
      "eval_precision": 0.7710843373493976,
      "eval_recall": 0.44991212653778556,
      "eval_runtime": 3.165,
      "eval_samples_per_second": 539.344,
      "eval_steps_per_second": 17.062,
      "step": 1498
    },
    {
      "epoch": 8.0,
      "grad_norm": 2.767106294631958,
      "learning_rate": 0.00020518384339968248,
      "loss": 0.431,
      "step": 1712
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7727006444053895,
      "eval_f1": 0.5773420479302832,
      "eval_loss": 0.4424216151237488,
      "eval_mcc": 0.4580968046120123,
      "eval_precision": 0.7593123209169055,
      "eval_recall": 0.46572934973637964,
      "eval_runtime": 4.1398,
      "eval_samples_per_second": 412.342,
      "eval_steps_per_second": 13.044,
      "step": 1712
    },
    {
      "epoch": 9.0,
      "grad_norm": 7.568024635314941,
      "learning_rate": 0.0001880851897830423,
      "loss": 0.4269,
      "step": 1926
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7691857059168131,
      "eval_f1": 0.5553047404063206,
      "eval_loss": 0.4476867616176605,
      "eval_mcc": 0.4484665614833703,
      "eval_precision": 0.7760252365930599,
      "eval_recall": 0.43233743409490333,
      "eval_runtime": 3.1524,
      "eval_samples_per_second": 541.497,
      "eval_steps_per_second": 17.13,
      "step": 1926
    },
    {
      "epoch": 10.0,
      "grad_norm": 3.2477447986602783,
      "learning_rate": 0.00017098653616640208,
      "loss": 0.4186,
      "step": 2140
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7768014059753954,
      "eval_f1": 0.6193806193806194,
      "eval_loss": 0.45719364285469055,
      "eval_mcc": 0.4744794188942744,
      "eval_precision": 0.7175925925925926,
      "eval_recall": 0.5448154657293497,
      "eval_runtime": 3.2524,
      "eval_samples_per_second": 524.847,
      "eval_steps_per_second": 16.603,
      "step": 2140
    },
    {
      "epoch": 11.0,
      "grad_norm": 2.3632404804229736,
      "learning_rate": 0.00015388788254976188,
      "loss": 0.4109,
      "step": 2354
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.773286467486819,
      "eval_f1": 0.5587229190421892,
      "eval_loss": 0.44749346375465393,
      "eval_mcc": 0.45996910450651607,
      "eval_precision": 0.7954545454545454,
      "eval_recall": 0.4305799648506151,
      "eval_runtime": 3.1492,
      "eval_samples_per_second": 542.04,
      "eval_steps_per_second": 17.147,
      "step": 2354
    },
    {
      "epoch": 12.0,
      "grad_norm": 2.5821080207824707,
      "learning_rate": 0.00013678922893312168,
      "loss": 0.4098,
      "step": 2568
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7932044522554189,
      "eval_f1": 0.6672950047125353,
      "eval_loss": 0.43375280499458313,
      "eval_mcc": 0.5213017315650612,
      "eval_precision": 0.7195121951219512,
      "eval_recall": 0.6221441124780316,
      "eval_runtime": 3.3029,
      "eval_samples_per_second": 516.82,
      "eval_steps_per_second": 16.349,
      "step": 2568
    }
  ],
  "logging_steps": 500,
  "max_steps": 4280,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 6297307973280.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8988863482372167,
    "learning_rate": 0.00034197307233280416,
    "num_train_epochs": 20,
    "temperature": 20
  }
}