{
  "best_metric": 0.8356164383561644,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-0/checkpoint-576",
  "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": 2.648221731185913,
      "learning_rate": 0.0001778537404863438,
      "loss": 0.5789,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7700587084148728,
      "eval_f1": 0.8049792531120332,
      "eval_loss": 0.47850364446640015,
      "eval_precision": 0.6988472622478387,
      "eval_recall": 0.949119373776908,
      "eval_runtime": 31.8773,
      "eval_samples_per_second": 32.06,
      "eval_steps_per_second": 1.004,
      "step": 96
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.695119857788086,
      "learning_rate": 0.0001482114504052865,
      "loss": 0.4663,
      "step": 192
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8131115459882583,
      "eval_f1": 0.8323090430201932,
      "eval_loss": 0.42822203040122986,
      "eval_precision": 0.7547770700636943,
      "eval_recall": 0.9275929549902152,
      "eval_runtime": 31.0929,
      "eval_samples_per_second": 32.869,
      "eval_steps_per_second": 1.029,
      "step": 192
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.924405097961426,
      "learning_rate": 0.0001185691603242292,
      "loss": 0.4316,
      "step": 288
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8228962818003914,
      "eval_f1": 0.8331797235023042,
      "eval_loss": 0.4343426525592804,
      "eval_precision": 0.7874564459930313,
      "eval_recall": 0.8845401174168297,
      "eval_runtime": 31.8005,
      "eval_samples_per_second": 32.138,
      "eval_steps_per_second": 1.006,
      "step": 288
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.703495979309082,
      "learning_rate": 8.89268702431719e-05,
      "loss": 0.414,
      "step": 384
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8180039138943248,
      "eval_f1": 0.8245283018867925,
      "eval_loss": 0.41725680232048035,
      "eval_precision": 0.7959927140255009,
      "eval_recall": 0.8551859099804305,
      "eval_runtime": 32.0071,
      "eval_samples_per_second": 31.93,
      "eval_steps_per_second": 1.0,
      "step": 384
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.176375150680542,
      "learning_rate": 5.92845801621146e-05,
      "loss": 0.4042,
      "step": 480
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8336594911937377,
      "eval_f1": 0.8473967684021544,
      "eval_loss": 0.4033981263637543,
      "eval_precision": 0.7827529021558872,
      "eval_recall": 0.923679060665362,
      "eval_runtime": 31.1502,
      "eval_samples_per_second": 32.809,
      "eval_steps_per_second": 1.027,
      "step": 480
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.6770453453063965,
      "learning_rate": 2.96422900810573e-05,
      "loss": 0.3947,
      "step": 576
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8356164383561644,
      "eval_f1": 0.8497316636851521,
      "eval_loss": 0.4074985384941101,
      "eval_precision": 0.7825370675453048,
      "eval_recall": 0.9295499021526419,
      "eval_runtime": 31.7728,
      "eval_samples_per_second": 32.166,
      "eval_steps_per_second": 1.007,
      "step": 576
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.172855854034424,
      "learning_rate": 0.0,
      "loss": 0.3887,
      "step": 672
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8356164383561644,
      "eval_f1": 0.8467153284671534,
      "eval_loss": 0.4036194682121277,
      "eval_precision": 0.7931623931623931,
      "eval_recall": 0.9080234833659491,
      "eval_runtime": 32.108,
      "eval_samples_per_second": 31.83,
      "eval_steps_per_second": 0.997,
      "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": 0.9489576625421504,
    "learning_rate": 0.00020749603056740112,
    "num_train_epochs": 7,
    "temperature": 28
  }
}