{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 17.0,
  "eval_steps": 500,
  "global_step": 5406,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9968553459119497,
      "grad_norm": 0.37894660234451294,
      "learning_rate": 8.9089921371132e-05,
      "loss": 0.2936,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8738709677419355,
      "eval_loss": 0.06461235135793686,
      "eval_runtime": 2.6511,
      "eval_samples_per_second": 1169.335,
      "eval_steps_per_second": 24.518,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.2703763544559479,
      "learning_rate": 8.354040180448849e-05,
      "loss": 0.0584,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9264516129032258,
      "eval_loss": 0.03086796961724758,
      "eval_runtime": 2.6512,
      "eval_samples_per_second": 1169.284,
      "eval_steps_per_second": 24.517,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.20945589244365692,
      "learning_rate": 7.799088223784498e-05,
      "loss": 0.035,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9361290322580645,
      "eval_loss": 0.023069918155670166,
      "eval_runtime": 2.6554,
      "eval_samples_per_second": 1167.42,
      "eval_steps_per_second": 24.478,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.1196790486574173,
      "learning_rate": 7.244136267120147e-05,
      "loss": 0.0276,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9390322580645162,
      "eval_loss": 0.02050953172147274,
      "eval_runtime": 2.6346,
      "eval_samples_per_second": 1176.652,
      "eval_steps_per_second": 24.672,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.11527035385370255,
      "learning_rate": 6.689184310455794e-05,
      "loss": 0.0239,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9361290322580645,
      "eval_loss": 0.019045764580368996,
      "eval_runtime": 2.6584,
      "eval_samples_per_second": 1166.129,
      "eval_steps_per_second": 24.451,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.10317819565534592,
      "learning_rate": 6.134232353791443e-05,
      "loss": 0.0217,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9412903225806452,
      "eval_loss": 0.018115758895874023,
      "eval_runtime": 2.6563,
      "eval_samples_per_second": 1167.048,
      "eval_steps_per_second": 24.47,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.11120817810297012,
      "learning_rate": 5.5792803971270916e-05,
      "loss": 0.0202,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9435483870967742,
      "eval_loss": 0.017409062013030052,
      "eval_runtime": 2.664,
      "eval_samples_per_second": 1163.673,
      "eval_steps_per_second": 24.4,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.08610443770885468,
      "learning_rate": 5.0243284404627405e-05,
      "loss": 0.0189,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9441935483870968,
      "eval_loss": 0.01702069491147995,
      "eval_runtime": 2.6577,
      "eval_samples_per_second": 1166.417,
      "eval_steps_per_second": 24.457,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.09259846806526184,
      "learning_rate": 4.469376483798389e-05,
      "loss": 0.018,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9458064516129032,
      "eval_loss": 0.016187822446227074,
      "eval_runtime": 2.6545,
      "eval_samples_per_second": 1167.839,
      "eval_steps_per_second": 24.487,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.07796873152256012,
      "learning_rate": 3.914424527134037e-05,
      "loss": 0.0171,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9438709677419355,
      "eval_loss": 0.015915485098958015,
      "eval_runtime": 2.6415,
      "eval_samples_per_second": 1173.555,
      "eval_steps_per_second": 24.607,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.06984979659318924,
      "learning_rate": 3.359472570469686e-05,
      "loss": 0.0164,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9416129032258065,
      "eval_loss": 0.01606527529656887,
      "eval_runtime": 2.6279,
      "eval_samples_per_second": 1179.632,
      "eval_steps_per_second": 24.734,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.06347791105508804,
      "learning_rate": 2.8045206138053343e-05,
      "loss": 0.0159,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9461290322580646,
      "eval_loss": 0.015618913806974888,
      "eval_runtime": 2.6558,
      "eval_samples_per_second": 1167.25,
      "eval_steps_per_second": 24.475,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.08197327703237534,
      "learning_rate": 2.249568657140983e-05,
      "loss": 0.0153,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9438709677419355,
      "eval_loss": 0.015260354615747929,
      "eval_runtime": 2.6542,
      "eval_samples_per_second": 1167.962,
      "eval_steps_per_second": 24.49,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.06349155306816101,
      "learning_rate": 1.6946167004766316e-05,
      "loss": 0.0149,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9467741935483871,
      "eval_loss": 0.014791840687394142,
      "eval_runtime": 2.639,
      "eval_samples_per_second": 1174.705,
      "eval_steps_per_second": 24.631,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.0693543404340744,
      "learning_rate": 1.1396647438122802e-05,
      "loss": 0.0144,
      "step": 4755
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9464516129032258,
      "eval_loss": 0.014759526588022709,
      "eval_runtime": 2.629,
      "eval_samples_per_second": 1179.148,
      "eval_steps_per_second": 24.724,
      "step": 4770
    },
    {
      "epoch": 15.949685534591195,
      "grad_norm": 0.06401927769184113,
      "learning_rate": 5.847127871479286e-06,
      "loss": 0.0141,
      "step": 5072
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.947741935483871,
      "eval_loss": 0.014390287920832634,
      "eval_runtime": 2.6548,
      "eval_samples_per_second": 1167.693,
      "eval_steps_per_second": 24.484,
      "step": 5088
    },
    {
      "epoch": 16.946540880503143,
      "grad_norm": 0.06713937968015671,
      "learning_rate": 2.976083048357721e-07,
      "loss": 0.0139,
      "step": 5389
    }
  ],
  "logging_steps": 317,
  "max_steps": 5406,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 17,
  "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": 1401394002013824.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.28743091275198973,
    "fp16": true,
    "learning_rate": 9.215353627385319e-05,
    "lr_scheduler": "cosine",
    "num_train_epochs": 17,
    "temperature": 11,
    "warmup_steps": 142,
    "weight_decay": 0.1789135641130362
  }
}