{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 16.0,
  "eval_steps": 500,
  "global_step": 5088,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9968553459119497,
      "grad_norm": 0.29921460151672363,
      "learning_rate": 0.00012819630238394863,
      "loss": 0.2627,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8874193548387097,
      "eval_loss": 0.05519754812121391,
      "eval_runtime": 2.7241,
      "eval_samples_per_second": 1138.01,
      "eval_steps_per_second": 23.862,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.20764172077178955,
      "learning_rate": 0.00011967854345380574,
      "loss": 0.0497,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9296774193548387,
      "eval_loss": 0.03099919483065605,
      "eval_runtime": 2.6324,
      "eval_samples_per_second": 1177.638,
      "eval_steps_per_second": 24.692,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.23950758576393127,
      "learning_rate": 0.00011116078452366286,
      "loss": 0.0309,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9425806451612904,
      "eval_loss": 0.022007087245583534,
      "eval_runtime": 2.609,
      "eval_samples_per_second": 1188.208,
      "eval_steps_per_second": 24.914,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.11315398663282394,
      "learning_rate": 0.00010264302559351995,
      "loss": 0.0249,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9387096774193548,
      "eval_loss": 0.020410971716046333,
      "eval_runtime": 2.663,
      "eval_samples_per_second": 1164.097,
      "eval_steps_per_second": 24.408,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.12218069285154343,
      "learning_rate": 9.412526666337707e-05,
      "loss": 0.0221,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9387096774193548,
      "eval_loss": 0.01974082551896572,
      "eval_runtime": 2.6648,
      "eval_samples_per_second": 1163.306,
      "eval_steps_per_second": 24.392,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.11072761565446854,
      "learning_rate": 8.56075077332342e-05,
      "loss": 0.02,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9480645161290323,
      "eval_loss": 0.018622873350977898,
      "eval_runtime": 2.6471,
      "eval_samples_per_second": 1171.078,
      "eval_steps_per_second": 24.555,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.1001732349395752,
      "learning_rate": 7.708974880309129e-05,
      "loss": 0.0185,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9451612903225807,
      "eval_loss": 0.017867427319288254,
      "eval_runtime": 2.6485,
      "eval_samples_per_second": 1170.477,
      "eval_steps_per_second": 24.542,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.07717470079660416,
      "learning_rate": 6.857198987294841e-05,
      "loss": 0.0173,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9470967741935484,
      "eval_loss": 0.01696794480085373,
      "eval_runtime": 2.6421,
      "eval_samples_per_second": 1173.287,
      "eval_steps_per_second": 24.601,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.0753188505768776,
      "learning_rate": 6.0054230942805527e-05,
      "loss": 0.0164,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9454838709677419,
      "eval_loss": 0.01644430123269558,
      "eval_runtime": 2.6339,
      "eval_samples_per_second": 1176.963,
      "eval_steps_per_second": 24.678,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.0695965364575386,
      "learning_rate": 5.153647201266264e-05,
      "loss": 0.0157,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9451612903225807,
      "eval_loss": 0.01595238223671913,
      "eval_runtime": 2.6568,
      "eval_samples_per_second": 1166.806,
      "eval_steps_per_second": 24.465,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.06090254336595535,
      "learning_rate": 4.301871308251975e-05,
      "loss": 0.015,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9474193548387096,
      "eval_loss": 0.015776390209794044,
      "eval_runtime": 2.6651,
      "eval_samples_per_second": 1163.194,
      "eval_steps_per_second": 24.39,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.05655613914132118,
      "learning_rate": 3.4500954152376865e-05,
      "loss": 0.0145,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9490322580645161,
      "eval_loss": 0.01548299752175808,
      "eval_runtime": 2.6549,
      "eval_samples_per_second": 1167.669,
      "eval_steps_per_second": 24.483,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.06078702211380005,
      "learning_rate": 2.598319522223398e-05,
      "loss": 0.0139,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9474193548387096,
      "eval_loss": 0.015226291492581367,
      "eval_runtime": 2.6558,
      "eval_samples_per_second": 1167.243,
      "eval_steps_per_second": 24.474,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.05542264133691788,
      "learning_rate": 1.7465436292091095e-05,
      "loss": 0.0135,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9483870967741935,
      "eval_loss": 0.014664444141089916,
      "eval_runtime": 2.6705,
      "eval_samples_per_second": 1160.811,
      "eval_steps_per_second": 24.34,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.05966860428452492,
      "learning_rate": 8.947677361948206e-06,
      "loss": 0.0131,
      "step": 4755
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9458064516129032,
      "eval_loss": 0.014484290964901447,
      "eval_runtime": 2.6603,
      "eval_samples_per_second": 1165.301,
      "eval_steps_per_second": 24.434,
      "step": 4770
    },
    {
      "epoch": 15.949685534591195,
      "grad_norm": 0.05474509298801422,
      "learning_rate": 4.299184318053192e-07,
      "loss": 0.0127,
      "step": 5072
    }
  ],
  "logging_steps": 317,
  "max_steps": 5088,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 16,
  "save_steps": 1000000000.0,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1319269593188652.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9923256140916163,
    "fp16": false,
    "learning_rate": 0.00013300601483977063,
    "lr_scheduler": "cosine",
    "num_train_epochs": 16,
    "temperature": 12,
    "warmup_steps": 138,
    "weight_decay": 0.2272381410096871
  }
}