{
  "best_metric": 0.6534653465346535,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-6/checkpoint-976",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 1708,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.8588551878929138,
      "learning_rate": 0.0001547989976805511,
      "loss": 0.5028,
      "step": 244
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5089108910891089,
      "eval_f1": 0.1205673758865248,
      "eval_loss": 0.504381000995636,
      "eval_precision": 0.5666666666666667,
      "eval_recall": 0.06746031746031746,
      "eval_runtime": 14.9415,
      "eval_samples_per_second": 33.799,
      "eval_steps_per_second": 1.071,
      "step": 244
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.0281460285186768,
      "learning_rate": 0.00012899916473379258,
      "loss": 0.4912,
      "step": 488
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5782178217821782,
      "eval_f1": 0.4714640198511167,
      "eval_loss": 0.484104186296463,
      "eval_precision": 0.6291390728476821,
      "eval_recall": 0.376984126984127,
      "eval_runtime": 15.1104,
      "eval_samples_per_second": 33.421,
      "eval_steps_per_second": 1.059,
      "step": 488
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.708383321762085,
      "learning_rate": 0.00010319933178703406,
      "loss": 0.4797,
      "step": 732
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6415841584158416,
      "eval_f1": 0.6578449905482042,
      "eval_loss": 0.4840504229068756,
      "eval_precision": 0.628158844765343,
      "eval_recall": 0.6904761904761905,
      "eval_runtime": 14.8003,
      "eval_samples_per_second": 34.121,
      "eval_steps_per_second": 1.081,
      "step": 732
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.030291199684143,
      "learning_rate": 7.739949884027555e-05,
      "loss": 0.4669,
      "step": 976
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6534653465346535,
      "eval_f1": 0.6777163904235726,
      "eval_loss": 0.47764360904693604,
      "eval_precision": 0.6323024054982818,
      "eval_recall": 0.7301587301587301,
      "eval_runtime": 15.0188,
      "eval_samples_per_second": 33.624,
      "eval_steps_per_second": 1.065,
      "step": 976
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.2075254917144775,
      "learning_rate": 5.159966589351703e-05,
      "loss": 0.4603,
      "step": 1220
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6435643564356436,
      "eval_f1": 0.64,
      "eval_loss": 0.4682587683200836,
      "eval_precision": 0.6451612903225806,
      "eval_recall": 0.6349206349206349,
      "eval_runtime": 15.051,
      "eval_samples_per_second": 33.553,
      "eval_steps_per_second": 1.063,
      "step": 1220
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.7989599704742432,
      "learning_rate": 2.5799832946758514e-05,
      "loss": 0.4558,
      "step": 1464
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6455445544554456,
      "eval_f1": 0.6483300589390963,
      "eval_loss": 0.4689685106277466,
      "eval_precision": 0.642023346303502,
      "eval_recall": 0.6547619047619048,
      "eval_runtime": 15.4127,
      "eval_samples_per_second": 32.765,
      "eval_steps_per_second": 1.038,
      "step": 1464
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.2156615257263184,
      "learning_rate": 0.0,
      "loss": 0.4531,
      "step": 1708
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6495049504950495,
      "eval_f1": 0.6424242424242423,
      "eval_loss": 0.46652600169181824,
      "eval_precision": 0.654320987654321,
      "eval_recall": 0.6309523809523809,
      "eval_runtime": 14.8821,
      "eval_samples_per_second": 33.933,
      "eval_steps_per_second": 1.075,
      "step": 1708
    }
  ],
  "logging_steps": 500,
  "max_steps": 1708,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 5444902981980.0,
  "train_batch_size": 39,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6658046915260407,
    "learning_rate": 0.0001805988306273096,
    "num_train_epochs": 7,
    "per_device_train_batch_size": 39,
    "temperature": 15
  }
}