{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "eval_steps": 500,
  "global_step": 4770,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9968553459119497,
      "grad_norm": 0.24409309029579163,
      "learning_rate": 0.00014590873969244836,
      "loss": 0.1983,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8987096774193548,
      "eval_loss": 0.04829428717494011,
      "eval_runtime": 5.4639,
      "eval_samples_per_second": 567.356,
      "eval_steps_per_second": 11.896,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.2151748687028885,
      "learning_rate": 0.0001411733693365559,
      "loss": 0.0453,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9312903225806451,
      "eval_loss": 0.03353011608123779,
      "eval_runtime": 5.4346,
      "eval_samples_per_second": 570.424,
      "eval_steps_per_second": 11.96,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.1853768676519394,
      "learning_rate": 0.00013350993615255516,
      "loss": 0.0302,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9380645161290323,
      "eval_loss": 0.024508632719516754,
      "eval_runtime": 5.4066,
      "eval_samples_per_second": 573.375,
      "eval_steps_per_second": 12.022,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.12661713361740112,
      "learning_rate": 0.00012325127343070825,
      "loss": 0.0241,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9390322580645162,
      "eval_loss": 0.022521397098898888,
      "eval_runtime": 5.4552,
      "eval_samples_per_second": 568.268,
      "eval_steps_per_second": 11.915,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.07092136144638062,
      "learning_rate": 0.00011084292879770157,
      "loss": 0.0211,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9432258064516129,
      "eval_loss": 0.020574456080794334,
      "eval_runtime": 5.4348,
      "eval_samples_per_second": 570.399,
      "eval_steps_per_second": 11.96,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.08544526994228363,
      "learning_rate": 9.682381347931996e-05,
      "loss": 0.0191,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9425806451612904,
      "eval_loss": 0.020394276827573776,
      "eval_runtime": 5.3752,
      "eval_samples_per_second": 576.72,
      "eval_steps_per_second": 12.093,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.07930737733840942,
      "learning_rate": 8.180279665546596e-05,
      "loss": 0.0178,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.942258064516129,
      "eval_loss": 0.019140785560011864,
      "eval_runtime": 5.4099,
      "eval_samples_per_second": 573.027,
      "eval_steps_per_second": 12.015,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.07150762528181076,
      "learning_rate": 6.643226144641048e-05,
      "loss": 0.0166,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9429032258064516,
      "eval_loss": 0.018344789743423462,
      "eval_runtime": 5.4691,
      "eval_samples_per_second": 566.817,
      "eval_steps_per_second": 11.885,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.07256153225898743,
      "learning_rate": 5.137977102963379e-05,
      "loss": 0.0156,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9435483870967742,
      "eval_loss": 0.018101442605257034,
      "eval_runtime": 5.4494,
      "eval_samples_per_second": 568.87,
      "eval_steps_per_second": 11.928,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.0548202246427536,
      "learning_rate": 3.729907546620694e-05,
      "loss": 0.0149,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9470967741935484,
      "eval_loss": 0.01727716624736786,
      "eval_runtime": 5.4078,
      "eval_samples_per_second": 573.243,
      "eval_steps_per_second": 12.02,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.06869664788246155,
      "learning_rate": 2.480171844954125e-05,
      "loss": 0.0141,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9451612903225807,
      "eval_loss": 0.016744110733270645,
      "eval_runtime": 5.3758,
      "eval_samples_per_second": 576.654,
      "eval_steps_per_second": 12.091,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.058046165853738785,
      "learning_rate": 1.4430477133891922e-05,
      "loss": 0.0137,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9464516129032258,
      "eval_loss": 0.016505638137459755,
      "eval_runtime": 5.4235,
      "eval_samples_per_second": 571.587,
      "eval_steps_per_second": 11.985,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.07068239897489548,
      "learning_rate": 6.6357885868641785e-06,
      "loss": 0.0133,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9464516129032258,
      "eval_loss": 0.016218330711126328,
      "eval_runtime": 5.4782,
      "eval_samples_per_second": 565.88,
      "eval_steps_per_second": 11.865,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.055857669562101364,
      "learning_rate": 1.7561866970409828e-06,
      "loss": 0.0131,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9467741935483871,
      "eval_loss": 0.01611991599202156,
      "eval_runtime": 5.4357,
      "eval_samples_per_second": 570.308,
      "eval_steps_per_second": 11.958,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.04982820898294449,
      "learning_rate": 3.599188350283289e-09,
      "loss": 0.0129,
      "step": 4755
    }
  ],
  "logging_steps": 317,
  "max_steps": 4770,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "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": 1236796014635076.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8539116398349643,
    "learning_rate": 0.00014751038366731588,
    "lr_scheduler_type": "cosine",
    "num_train_epochs": 15,
    "temperature": 10.068159940358916,
    "weight_decay": 0.299840395098859
  }
}