{
  "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": 0.6061979532241821,
      "learning_rate": 0.00019605083868932084,
      "loss": 0.5436,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8741935483870967,
      "eval_loss": 0.08529900014400482,
      "eval_runtime": 2.7094,
      "eval_samples_per_second": 1144.165,
      "eval_steps_per_second": 23.991,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.7015511393547058,
      "learning_rate": 0.00027117893259218497,
      "loss": 0.0831,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9151612903225806,
      "eval_loss": 0.06519128382205963,
      "eval_runtime": 2.647,
      "eval_samples_per_second": 1171.156,
      "eval_steps_per_second": 24.556,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.3628714978694916,
      "learning_rate": 0.00025187859095955747,
      "loss": 0.0543,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.927741935483871,
      "eval_loss": 0.0552055686712265,
      "eval_runtime": 2.6505,
      "eval_samples_per_second": 1169.573,
      "eval_steps_per_second": 24.523,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.4606303572654724,
      "learning_rate": 0.00023257824932693003,
      "loss": 0.0392,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9229032258064516,
      "eval_loss": 0.054187335073947906,
      "eval_runtime": 2.6589,
      "eval_samples_per_second": 1165.89,
      "eval_steps_per_second": 24.446,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.1761247217655182,
      "learning_rate": 0.0002132779076943026,
      "loss": 0.0318,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9345161290322581,
      "eval_loss": 0.04057912528514862,
      "eval_runtime": 2.629,
      "eval_samples_per_second": 1179.148,
      "eval_steps_per_second": 24.724,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.2396286129951477,
      "learning_rate": 0.00019397756606167517,
      "loss": 0.0287,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9406451612903226,
      "eval_loss": 0.0397094301879406,
      "eval_runtime": 2.6735,
      "eval_samples_per_second": 1159.523,
      "eval_steps_per_second": 24.313,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.17782621085643768,
      "learning_rate": 0.00017467722442904773,
      "loss": 0.0251,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9383870967741935,
      "eval_loss": 0.03664569929242134,
      "eval_runtime": 2.6609,
      "eval_samples_per_second": 1165.001,
      "eval_steps_per_second": 24.427,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.09906672686338425,
      "learning_rate": 0.00015537688279642029,
      "loss": 0.0225,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9416129032258065,
      "eval_loss": 0.033652398735284805,
      "eval_runtime": 2.6704,
      "eval_samples_per_second": 1160.878,
      "eval_steps_per_second": 24.341,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.12453433126211166,
      "learning_rate": 0.00013607654116379284,
      "loss": 0.021,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9438709677419355,
      "eval_loss": 0.03149178624153137,
      "eval_runtime": 2.6699,
      "eval_samples_per_second": 1161.086,
      "eval_steps_per_second": 24.345,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.09386735409498215,
      "learning_rate": 0.0001167761995311654,
      "loss": 0.0198,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9412903225806452,
      "eval_loss": 0.029914697632193565,
      "eval_runtime": 2.6674,
      "eval_samples_per_second": 1162.17,
      "eval_steps_per_second": 24.368,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.0862426608800888,
      "learning_rate": 9.747585789853797e-05,
      "loss": 0.0185,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9435483870967742,
      "eval_loss": 0.02804023027420044,
      "eval_runtime": 2.6984,
      "eval_samples_per_second": 1148.825,
      "eval_steps_per_second": 24.088,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.08728446811437607,
      "learning_rate": 7.817551626591052e-05,
      "loss": 0.0178,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9464516129032258,
      "eval_loss": 0.02766558527946472,
      "eval_runtime": 2.6557,
      "eval_samples_per_second": 1167.283,
      "eval_steps_per_second": 24.475,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.0797114223241806,
      "learning_rate": 5.8875174633283076e-05,
      "loss": 0.0167,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9461290322580646,
      "eval_loss": 0.027048034593462944,
      "eval_runtime": 2.6519,
      "eval_samples_per_second": 1168.952,
      "eval_steps_per_second": 24.51,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.07728114724159241,
      "learning_rate": 3.9574833000655634e-05,
      "loss": 0.016,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9448387096774193,
      "eval_loss": 0.02572786808013916,
      "eval_runtime": 2.6537,
      "eval_samples_per_second": 1168.161,
      "eval_steps_per_second": 24.494,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.07841521501541138,
      "learning_rate": 2.0274491368028198e-05,
      "loss": 0.0153,
      "step": 4755
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9438709677419355,
      "eval_loss": 0.025068214163184166,
      "eval_runtime": 2.6663,
      "eval_samples_per_second": 1162.674,
      "eval_steps_per_second": 24.379,
      "step": 4770
    },
    {
      "epoch": 15.949685534591195,
      "grad_norm": 0.06940767914056778,
      "learning_rate": 9.741497354007542e-07,
      "loss": 0.0147,
      "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": 0.43552464457666973,
    "fp16": true,
    "learning_rate": 0.0002820163483985183,
    "lr_scheduler": "linear",
    "num_train_epochs": 16,
    "temperature": 2,
    "warmup_steps": 456,
    "weight_decay": 0.22663780810649276
  }
}