{
  "best_metric": 0.8371559633027523,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-15/checkpoint-527",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 3689,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 20.894563674926758,
      "learning_rate": 0.00025961167449966676,
      "loss": 1.6678,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8371559633027523,
      "eval_loss": 1.635141134262085,
      "eval_runtime": 2.8174,
      "eval_samples_per_second": 309.502,
      "eval_steps_per_second": 2.485,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.817446708679199,
      "learning_rate": 0.00021634306208305565,
      "loss": 0.738,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8188073394495413,
      "eval_loss": 1.8634682893753052,
      "eval_runtime": 2.8123,
      "eval_samples_per_second": 310.062,
      "eval_steps_per_second": 2.489,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": NaN,
      "learning_rate": 0.00017315655329531854,
      "loss": 0.5214,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8222477064220184,
      "eval_loss": 1.81102454662323,
      "eval_runtime": 2.8135,
      "eval_samples_per_second": 309.932,
      "eval_steps_per_second": 2.488,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.225761890411377,
      "learning_rate": 0.0001298879408787074,
      "loss": 0.4003,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8302752293577982,
      "eval_loss": 1.878706932067871,
      "eval_runtime": 2.8102,
      "eval_samples_per_second": 310.296,
      "eval_steps_per_second": 2.491,
      "step": 2108
    },
    {
      "epoch": 5.0,
      "grad_norm": 12.372098922729492,
      "learning_rate": 8.661932846209629e-05,
      "loss": 0.32,
      "step": 2635
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8256880733944955,
      "eval_loss": 1.8621728420257568,
      "eval_runtime": 2.8187,
      "eval_samples_per_second": 309.357,
      "eval_steps_per_second": 2.483,
      "step": 2635
    },
    {
      "epoch": 6.0,
      "grad_norm": 20.063739776611328,
      "learning_rate": 4.3350716045485156e-05,
      "loss": 0.2702,
      "step": 3162
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.823394495412844,
      "eval_loss": 1.8646079301834106,
      "eval_runtime": 2.813,
      "eval_samples_per_second": 309.994,
      "eval_steps_per_second": 2.488,
      "step": 3162
    },
    {
      "epoch": 7.0,
      "grad_norm": 8.405074119567871,
      "learning_rate": 8.21036288740249e-08,
      "loss": 0.2411,
      "step": 3689
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8268348623853211,
      "eval_loss": 1.880131483078003,
      "eval_runtime": 2.8069,
      "eval_samples_per_second": 310.665,
      "eval_steps_per_second": 2.494,
      "step": 3689
    }
  ],
  "logging_steps": 500,
  "max_steps": 3689,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 56628612372420.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.18003409042088625,
    "learning_rate": 0.0003028802869162779,
    "num_train_epochs": 7,
    "temperature": 27
  }
}