{
  "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": 0.40019768476486206,
      "learning_rate": 0.00016071954069252033,
      "loss": 0.2885,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8745161290322581,
      "eval_loss": 0.06336037814617157,
      "eval_runtime": 2.6312,
      "eval_samples_per_second": 1178.163,
      "eval_steps_per_second": 24.703,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.2962985634803772,
      "learning_rate": 0.00015070812501173257,
      "loss": 0.0531,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9241935483870968,
      "eval_loss": 0.034031569957733154,
      "eval_runtime": 2.6487,
      "eval_samples_per_second": 1170.371,
      "eval_steps_per_second": 24.54,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.19826966524124146,
      "learning_rate": 0.0001406967093309448,
      "loss": 0.0314,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9406451612903226,
      "eval_loss": 0.024853697046637535,
      "eval_runtime": 2.6706,
      "eval_samples_per_second": 1160.807,
      "eval_steps_per_second": 24.34,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.11921875178813934,
      "learning_rate": 0.00013068529365015705,
      "loss": 0.0247,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.944516129032258,
      "eval_loss": 0.02187591791152954,
      "eval_runtime": 2.6534,
      "eval_samples_per_second": 1168.295,
      "eval_steps_per_second": 24.497,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.09643226116895676,
      "learning_rate": 0.00012067387796936926,
      "loss": 0.0215,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9448387096774193,
      "eval_loss": 0.020509159192442894,
      "eval_runtime": 2.6308,
      "eval_samples_per_second": 1178.365,
      "eval_steps_per_second": 24.708,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.14147548377513885,
      "learning_rate": 0.0001106624622885815,
      "loss": 0.0197,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9441935483870968,
      "eval_loss": 0.02022239752113819,
      "eval_runtime": 2.6786,
      "eval_samples_per_second": 1157.315,
      "eval_steps_per_second": 24.266,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.07504777610301971,
      "learning_rate": 0.00010065104660779374,
      "loss": 0.0183,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9441935483870968,
      "eval_loss": 0.019469449296593666,
      "eval_runtime": 2.6497,
      "eval_samples_per_second": 1169.936,
      "eval_steps_per_second": 24.531,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.07543275505304337,
      "learning_rate": 9.063963092700595e-05,
      "loss": 0.0172,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9454838709677419,
      "eval_loss": 0.019471565261483192,
      "eval_runtime": 2.6321,
      "eval_samples_per_second": 1177.77,
      "eval_steps_per_second": 24.695,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.07022672146558762,
      "learning_rate": 8.062821524621819e-05,
      "loss": 0.0162,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9454838709677419,
      "eval_loss": 0.01835792325437069,
      "eval_runtime": 2.6569,
      "eval_samples_per_second": 1166.766,
      "eval_steps_per_second": 24.464,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.06286010891199112,
      "learning_rate": 7.061679956543043e-05,
      "loss": 0.0154,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9467741935483871,
      "eval_loss": 0.017796719446778297,
      "eval_runtime": 2.6735,
      "eval_samples_per_second": 1159.509,
      "eval_steps_per_second": 24.312,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.05888282507658005,
      "learning_rate": 6.060538388464266e-05,
      "loss": 0.0147,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9464516129032258,
      "eval_loss": 0.01736345887184143,
      "eval_runtime": 2.6469,
      "eval_samples_per_second": 1171.197,
      "eval_steps_per_second": 24.557,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.05614669620990753,
      "learning_rate": 5.0593968203854894e-05,
      "loss": 0.0142,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9480645161290323,
      "eval_loss": 0.01718168891966343,
      "eval_runtime": 2.6286,
      "eval_samples_per_second": 1179.354,
      "eval_steps_per_second": 24.728,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.05945774167776108,
      "learning_rate": 4.0582552523067126e-05,
      "loss": 0.0136,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9496774193548387,
      "eval_loss": 0.016656547784805298,
      "eval_runtime": 2.6659,
      "eval_samples_per_second": 1162.853,
      "eval_steps_per_second": 24.382,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.0536658801138401,
      "learning_rate": 3.0571136842279366e-05,
      "loss": 0.0132,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9493548387096774,
      "eval_loss": 0.015961430966854095,
      "eval_runtime": 2.6639,
      "eval_samples_per_second": 1163.709,
      "eval_steps_per_second": 24.4,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.052763283252716064,
      "learning_rate": 2.0559721161491598e-05,
      "loss": 0.0127,
      "step": 4755
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9480645161290323,
      "eval_loss": 0.015751173719763756,
      "eval_runtime": 2.6449,
      "eval_samples_per_second": 1172.047,
      "eval_steps_per_second": 24.575,
      "step": 4770
    },
    {
      "epoch": 15.949685534591195,
      "grad_norm": 0.05405045673251152,
      "learning_rate": 1.0548305480703831e-05,
      "loss": 0.0123,
      "step": 5072
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9490322580645161,
      "eval_loss": 0.015423502773046494,
      "eval_runtime": 2.6549,
      "eval_samples_per_second": 1167.639,
      "eval_steps_per_second": 24.483,
      "step": 5088
    },
    {
      "epoch": 16.946540880503143,
      "grad_norm": 0.054720040410757065,
      "learning_rate": 5.368897999160632e-07,
      "loss": 0.0121,
      "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": 0.23184342374680478,
    "fp16": false,
    "learning_rate": 0.00016308817216273826,
    "lr_scheduler": "cosine",
    "num_train_epochs": 17,
    "temperature": 12,
    "warmup_steps": 242,
    "weight_decay": 0.1818121677410312
  }
}