{
  "best_metric": 0.8346379647749511,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-20/checkpoint-480",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 672,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 3.3483004570007324,
      "learning_rate": 0.00013855533928571864,
      "loss": 0.5685,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7671232876712328,
      "eval_f1": 0.8019966722129783,
      "eval_loss": 0.477300226688385,
      "eval_precision": 0.6975397973950795,
      "eval_recall": 0.9432485322896281,
      "eval_runtime": 25.843,
      "eval_samples_per_second": 39.546,
      "eval_steps_per_second": 1.238,
      "step": 96
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.558192729949951,
      "learning_rate": 0.00011546278273809888,
      "loss": 0.4736,
      "step": 192
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8072407045009785,
      "eval_f1": 0.8255093002657219,
      "eval_loss": 0.43477192521095276,
      "eval_precision": 0.7540453074433657,
      "eval_recall": 0.9119373776908023,
      "eval_runtime": 25.942,
      "eval_samples_per_second": 39.396,
      "eval_steps_per_second": 1.234,
      "step": 192
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.4191761016845703,
      "learning_rate": 9.237022619047909e-05,
      "loss": 0.4377,
      "step": 288
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8258317025440313,
      "eval_f1": 0.8381818181818183,
      "eval_loss": 0.4183538258075714,
      "eval_precision": 0.7826825127334465,
      "eval_recall": 0.9021526418786693,
      "eval_runtime": 25.665,
      "eval_samples_per_second": 39.821,
      "eval_steps_per_second": 1.247,
      "step": 288
    },
    {
      "epoch": 4.0,
      "grad_norm": 7.199775218963623,
      "learning_rate": 6.927766964285932e-05,
      "loss": 0.4231,
      "step": 384
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8199608610567515,
      "eval_f1": 0.8257575757575758,
      "eval_loss": 0.4184337258338928,
      "eval_precision": 0.8,
      "eval_recall": 0.8532289628180039,
      "eval_runtime": 25.0602,
      "eval_samples_per_second": 40.782,
      "eval_steps_per_second": 1.277,
      "step": 384
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.4152629375457764,
      "learning_rate": 4.6185113095239546e-05,
      "loss": 0.4131,
      "step": 480
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8346379647749511,
      "eval_f1": 0.8467815049864007,
      "eval_loss": 0.4067833125591278,
      "eval_precision": 0.7888513513513513,
      "eval_recall": 0.913894324853229,
      "eval_runtime": 24.9745,
      "eval_samples_per_second": 40.922,
      "eval_steps_per_second": 1.281,
      "step": 480
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.8709616661071777,
      "learning_rate": 2.3092556547619773e-05,
      "loss": 0.4025,
      "step": 576
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8307240704500979,
      "eval_f1": 0.8442844284428443,
      "eval_loss": 0.4088129699230194,
      "eval_precision": 0.7816666666666666,
      "eval_recall": 0.9178082191780822,
      "eval_runtime": 25.2849,
      "eval_samples_per_second": 40.419,
      "eval_steps_per_second": 1.266,
      "step": 576
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.1289896965026855,
      "learning_rate": 0.0,
      "loss": 0.398,
      "step": 672
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8287671232876712,
      "eval_f1": 0.839596700274977,
      "eval_loss": 0.40621256828308105,
      "eval_precision": 0.7896551724137931,
      "eval_recall": 0.8962818003913894,
      "eval_runtime": 25.662,
      "eval_samples_per_second": 39.825,
      "eval_steps_per_second": 1.247,
      "step": 672
    }
  ],
  "logging_steps": 500,
  "max_steps": 672,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 1649866380960.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8907547983808969,
    "learning_rate": 0.00016164789583333842,
    "num_train_epochs": 7,
    "temperature": 2
  }
}