{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 14.0,
  "eval_steps": 500,
  "global_step": 4452,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9968553459119497,
      "grad_norm": 0.47506219148635864,
      "learning_rate": 9.46923688245487e-06,
      "loss": 0.4817,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.33064516129032256,
      "eval_loss": 0.3162553906440735,
      "eval_runtime": 2.713,
      "eval_samples_per_second": 1142.656,
      "eval_steps_per_second": 23.959,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.5554618239402771,
      "learning_rate": 8.743300221816854e-06,
      "loss": 0.2672,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6270967741935484,
      "eval_loss": 0.18112796545028687,
      "eval_runtime": 2.624,
      "eval_samples_per_second": 1181.409,
      "eval_steps_per_second": 24.771,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.6009519100189209,
      "learning_rate": 8.017363561178839e-06,
      "loss": 0.1758,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7761290322580645,
      "eval_loss": 0.12200149148702621,
      "eval_runtime": 2.6358,
      "eval_samples_per_second": 1176.11,
      "eval_steps_per_second": 24.66,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.5635930299758911,
      "learning_rate": 7.291426900540823e-06,
      "loss": 0.1322,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8258064516129032,
      "eval_loss": 0.09276342391967773,
      "eval_runtime": 2.6813,
      "eval_samples_per_second": 1156.15,
      "eval_steps_per_second": 24.242,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.3592129349708557,
      "learning_rate": 6.5654902399028074e-06,
      "loss": 0.1073,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8583870967741936,
      "eval_loss": 0.07522393763065338,
      "eval_runtime": 2.6613,
      "eval_samples_per_second": 1164.858,
      "eval_steps_per_second": 24.424,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.5057026147842407,
      "learning_rate": 5.839553579264793e-06,
      "loss": 0.0919,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8806451612903226,
      "eval_loss": 0.06308811157941818,
      "eval_runtime": 2.6307,
      "eval_samples_per_second": 1178.407,
      "eval_steps_per_second": 24.709,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.35764896869659424,
      "learning_rate": 5.1136169186267775e-06,
      "loss": 0.0813,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8877419354838709,
      "eval_loss": 0.05563005059957504,
      "eval_runtime": 2.6575,
      "eval_samples_per_second": 1166.531,
      "eval_steps_per_second": 24.46,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.3608967363834381,
      "learning_rate": 4.387680257988761e-06,
      "loss": 0.0733,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8977419354838709,
      "eval_loss": 0.05038844048976898,
      "eval_runtime": 2.6676,
      "eval_samples_per_second": 1162.09,
      "eval_steps_per_second": 24.366,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.43566691875457764,
      "learning_rate": 3.661743597350746e-06,
      "loss": 0.0679,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9019354838709678,
      "eval_loss": 0.045959629118442535,
      "eval_runtime": 2.69,
      "eval_samples_per_second": 1152.43,
      "eval_steps_per_second": 24.164,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.3841456472873688,
      "learning_rate": 2.935806936712731e-06,
      "loss": 0.0636,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9038709677419355,
      "eval_loss": 0.04300825670361519,
      "eval_runtime": 2.6725,
      "eval_samples_per_second": 1159.945,
      "eval_steps_per_second": 24.321,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.3714827597141266,
      "learning_rate": 2.2098702760747155e-06,
      "loss": 0.0605,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.905483870967742,
      "eval_loss": 0.04122327268123627,
      "eval_runtime": 2.6669,
      "eval_samples_per_second": 1162.413,
      "eval_steps_per_second": 24.373,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.28599628806114197,
      "learning_rate": 1.4839336154367003e-06,
      "loss": 0.0584,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9080645161290323,
      "eval_loss": 0.039800167083740234,
      "eval_runtime": 2.6582,
      "eval_samples_per_second": 1166.221,
      "eval_steps_per_second": 24.453,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.3374980390071869,
      "learning_rate": 7.579969547986848e-07,
      "loss": 0.0568,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9103225806451613,
      "eval_loss": 0.03894847258925438,
      "eval_runtime": 2.6631,
      "eval_samples_per_second": 1164.046,
      "eval_steps_per_second": 24.407,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.2531619071960449,
      "learning_rate": 3.206029416066945e-08,
      "loss": 0.0562,
      "step": 4438
    }
  ],
  "logging_steps": 317,
  "max_steps": 4452,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 14,
  "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": 1234918384253064.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6907680425307066,
    "fp16": true,
    "learning_rate": 1.015395316488631e-05,
    "lr_scheduler": "cosine",
    "num_train_epochs": 14.721187464037747,
    "temperature": 13,
    "warmup_steps": 18,
    "weight_decay": 0.2870223156122959
  }
}