{
  "best_metric": 0.7850029291154071,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-26/checkpoint-1498",
  "epoch": 14.0,
  "eval_steps": 500,
  "global_step": 2996,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 7.895910739898682,
      "learning_rate": 0.00029747597659594095,
      "loss": 0.5937,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7152899824253075,
      "eval_f1": 0.425531914893617,
      "eval_loss": 0.5441433787345886,
      "eval_mcc": 0.2954832365575353,
      "eval_precision": 0.6498194945848376,
      "eval_recall": 0.3163444639718805,
      "eval_runtime": 3.1629,
      "eval_samples_per_second": 539.696,
      "eval_steps_per_second": 17.073,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.099026918411255,
      "learning_rate": 0.00028094953345172204,
      "loss": 0.5307,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7451669595782073,
      "eval_f1": 0.5435466946484785,
      "eval_loss": 0.5096856951713562,
      "eval_mcc": 0.3898809523809524,
      "eval_precision": 0.6744791666666666,
      "eval_recall": 0.45518453427065025,
      "eval_runtime": 3.1738,
      "eval_samples_per_second": 537.84,
      "eval_steps_per_second": 17.014,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.113579750061035,
      "learning_rate": 0.0002644230903075031,
      "loss": 0.505,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7393087287639133,
      "eval_f1": 0.4331210191082803,
      "eval_loss": 0.5280328392982483,
      "eval_mcc": 0.36632506494660955,
      "eval_precision": 0.7870370370370371,
      "eval_recall": 0.29876977152899825,
      "eval_runtime": 3.1389,
      "eval_samples_per_second": 543.826,
      "eval_steps_per_second": 17.204,
      "step": 642
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.820150852203369,
      "learning_rate": 0.00024789664716328417,
      "loss": 0.4888,
      "step": 856
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7510251903925015,
      "eval_f1": 0.4798041615667075,
      "eval_loss": 0.49989721179008484,
      "eval_mcc": 0.3996781742784298,
      "eval_precision": 0.7903225806451613,
      "eval_recall": 0.3444639718804921,
      "eval_runtime": 3.3284,
      "eval_samples_per_second": 512.86,
      "eval_steps_per_second": 16.224,
      "step": 856
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.9773123264312744,
      "learning_rate": 0.00023137020401906518,
      "loss": 0.475,
      "step": 1070
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7656707674282367,
      "eval_f1": 0.6190476190476191,
      "eval_loss": 0.48136404156684875,
      "eval_mcc": 0.4548772091771283,
      "eval_precision": 0.6756756756756757,
      "eval_recall": 0.5711775043936731,
      "eval_runtime": 3.1446,
      "eval_samples_per_second": 542.835,
      "eval_steps_per_second": 17.172,
      "step": 1070
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.4520063400268555,
      "learning_rate": 0.00021484376087484627,
      "loss": 0.4656,
      "step": 1284
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7650849443468073,
      "eval_f1": 0.6337899543378995,
      "eval_loss": 0.48336538672447205,
      "eval_mcc": 0.46203450521175543,
      "eval_precision": 0.6596958174904943,
      "eval_recall": 0.6098418277680141,
      "eval_runtime": 3.3889,
      "eval_samples_per_second": 503.697,
      "eval_steps_per_second": 15.934,
      "step": 1284
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.588683128356934,
      "learning_rate": 0.0001983173177306273,
      "loss": 0.4555,
      "step": 1498
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7850029291154071,
      "eval_f1": 0.608324439701174,
      "eval_loss": 0.48064079880714417,
      "eval_mcc": 0.4905686212612055,
      "eval_precision": 0.7744565217391305,
      "eval_recall": 0.5008787346221442,
      "eval_runtime": 3.1528,
      "eval_samples_per_second": 541.429,
      "eval_steps_per_second": 17.128,
      "step": 1498
    },
    {
      "epoch": 8.0,
      "grad_norm": 4.517602920532227,
      "learning_rate": 0.00018179087458640837,
      "loss": 0.4442,
      "step": 1712
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7797305213825425,
      "eval_f1": 0.6284584980237153,
      "eval_loss": 0.4647735357284546,
      "eval_mcc": 0.48287030915458007,
      "eval_precision": 0.7178329571106095,
      "eval_recall": 0.5588752196836555,
      "eval_runtime": 3.2447,
      "eval_samples_per_second": 526.083,
      "eval_steps_per_second": 16.642,
      "step": 1712
    },
    {
      "epoch": 9.0,
      "grad_norm": 5.497073650360107,
      "learning_rate": 0.0001652644314421894,
      "loss": 0.437,
      "step": 1926
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7691857059168131,
      "eval_f1": 0.606,
      "eval_loss": 0.4729827046394348,
      "eval_mcc": 0.4557733489579847,
      "eval_precision": 0.703016241299304,
      "eval_recall": 0.5325131810193322,
      "eval_runtime": 3.17,
      "eval_samples_per_second": 538.482,
      "eval_steps_per_second": 17.035,
      "step": 1926
    },
    {
      "epoch": 10.0,
      "grad_norm": 4.417614459991455,
      "learning_rate": 0.00014873798829797048,
      "loss": 0.4303,
      "step": 2140
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7691857059168131,
      "eval_f1": 0.64568345323741,
      "eval_loss": 0.4863553047180176,
      "eval_mcc": 0.4749521459303538,
      "eval_precision": 0.6611418047882136,
      "eval_recall": 0.6309314586994728,
      "eval_runtime": 3.1917,
      "eval_samples_per_second": 534.826,
      "eval_steps_per_second": 16.919,
      "step": 2140
    },
    {
      "epoch": 11.0,
      "grad_norm": 5.234562873840332,
      "learning_rate": 0.00013221154515375154,
      "loss": 0.4254,
      "step": 2354
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.7703573520796719,
      "eval_f1": 0.5908141962421711,
      "eval_loss": 0.4694356620311737,
      "eval_mcc": 0.4542660996178214,
      "eval_precision": 0.7275064267352185,
      "eval_recall": 0.4973637961335677,
      "eval_runtime": 3.1589,
      "eval_samples_per_second": 540.38,
      "eval_steps_per_second": 17.095,
      "step": 2354
    },
    {
      "epoch": 12.0,
      "grad_norm": 4.766694068908691,
      "learning_rate": 0.00011568510200953259,
      "loss": 0.4145,
      "step": 2568
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7709431751611013,
      "eval_f1": 0.6461538461538462,
      "eval_loss": 0.4714137315750122,
      "eval_mcc": 0.477505020147732,
      "eval_precision": 0.6660447761194029,
      "eval_recall": 0.6274165202108963,
      "eval_runtime": 3.2444,
      "eval_samples_per_second": 526.14,
      "eval_steps_per_second": 16.644,
      "step": 2568
    },
    {
      "epoch": 13.0,
      "grad_norm": 9.993538856506348,
      "learning_rate": 9.915865886531365e-05,
      "loss": 0.4113,
      "step": 2782
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.7685998828353837,
      "eval_f1": 0.6311858076563959,
      "eval_loss": 0.4818851053714752,
      "eval_mcc": 0.4654891699036773,
      "eval_precision": 0.6733067729083665,
      "eval_recall": 0.5940246045694201,
      "eval_runtime": 3.1513,
      "eval_samples_per_second": 541.687,
      "eval_steps_per_second": 17.136,
      "step": 2782
    },
    {
      "epoch": 14.0,
      "grad_norm": 14.592765808105469,
      "learning_rate": 8.26322157210947e-05,
      "loss": 0.4101,
      "step": 2996
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7738722905682484,
      "eval_f1": 0.6553571428571427,
      "eval_loss": 0.47235214710235596,
      "eval_mcc": 0.48729614444239394,
      "eval_precision": 0.6660617059891107,
      "eval_recall": 0.6449912126537786,
      "eval_runtime": 3.2584,
      "eval_samples_per_second": 523.878,
      "eval_steps_per_second": 16.573,
      "step": 2996
    }
  ],
  "logging_steps": 500,
  "max_steps": 4066,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 19,
  "save_steps": 500,
  "total_flos": 7346859302160.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9884393343643966,
    "learning_rate": 0.0003140024197401599,
    "num_train_epochs": 19,
    "temperature": 18
  }
}