{
  "best_metric": 0.926605504587156,
  "best_model_checkpoint": "electra-distilled-sst\\run-13\\checkpoint-3162",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 4743,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 22.835023880004883,
      "learning_rate": 4.3163581221778226e-05,
      "loss": 0.8573,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9162844036697247,
      "eval_loss": 0.5491362810134888,
      "eval_runtime": 1.4473,
      "eval_samples_per_second": 602.517,
      "eval_steps_per_second": 4.837,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 38.72591018676758,
      "learning_rate": 3.7769413020716535e-05,
      "loss": 0.4496,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9105504587155964,
      "eval_loss": 0.5399537682533264,
      "eval_runtime": 1.4499,
      "eval_samples_per_second": 601.429,
      "eval_steps_per_second": 4.828,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.218717575073242,
      "learning_rate": 3.237524481965485e-05,
      "loss": 0.3331,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 0.5268839001655579,
      "eval_runtime": 1.4424,
      "eval_samples_per_second": 604.54,
      "eval_steps_per_second": 4.853,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.019476890563965,
      "learning_rate": 2.6981076618593168e-05,
      "loss": 0.2593,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9174311926605505,
      "eval_loss": 0.5509608387947083,
      "eval_runtime": 1.453,
      "eval_samples_per_second": 600.123,
      "eval_steps_per_second": 4.818,
      "step": 2108
    },
    {
      "epoch": 5.0,
      "grad_norm": 19.7722110748291,
      "learning_rate": 2.1586908417531485e-05,
      "loss": 0.2191,
      "step": 2635
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9162844036697247,
      "eval_loss": 0.5370484590530396,
      "eval_runtime": 1.4546,
      "eval_samples_per_second": 599.496,
      "eval_steps_per_second": 4.812,
      "step": 2635
    },
    {
      "epoch": 6.0,
      "grad_norm": 12.621796607971191,
      "learning_rate": 1.6202975829754548e-05,
      "loss": 0.1874,
      "step": 3162
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.926605504587156,
      "eval_loss": 0.572212815284729,
      "eval_runtime": 1.4456,
      "eval_samples_per_second": 603.228,
      "eval_steps_per_second": 4.842,
      "step": 3162
    },
    {
      "epoch": 7.0,
      "grad_norm": 12.125116348266602,
      "learning_rate": 1.0808807628692863e-05,
      "loss": 0.1614,
      "step": 3689
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9197247706422018,
      "eval_loss": 0.5944486260414124,
      "eval_runtime": 1.4613,
      "eval_samples_per_second": 596.712,
      "eval_steps_per_second": 4.79,
      "step": 3689
    },
    {
      "epoch": 8.0,
      "grad_norm": 7.69290828704834,
      "learning_rate": 5.414639427631178e-06,
      "loss": 0.1463,
      "step": 4216
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 0.652703046798706,
      "eval_runtime": 1.4499,
      "eval_samples_per_second": 601.423,
      "eval_steps_per_second": 4.828,
      "step": 4216
    },
    {
      "epoch": 9.0,
      "grad_norm": 29.435096740722656,
      "learning_rate": 2.0471226569494057e-08,
      "loss": 0.1339,
      "step": 4743
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9220183486238532,
      "eval_loss": 0.6081395149230957,
      "eval_runtime": 1.4498,
      "eval_samples_per_second": 601.454,
      "eval_steps_per_second": 4.828,
      "step": 4743
    }
  ],
  "logging_steps": 500,
  "max_steps": 4743,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "total_flos": 1684816747339368.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8585651118406669,
    "learning_rate": 4.854751380955516e-05,
    "num_train_epochs": 9,
    "temperature": 23
  }
}