{
  "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": 0.3145744502544403,
      "learning_rate": 0.00013773896055969116,
      "loss": 0.2562,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8893548387096775,
      "eval_loss": 0.05541645735502243,
      "eval_runtime": 2.6126,
      "eval_samples_per_second": 1186.566,
      "eval_steps_per_second": 24.88,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.2050488293170929,
      "learning_rate": 0.00012858715789831575,
      "loss": 0.0488,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9303225806451613,
      "eval_loss": 0.030376020818948746,
      "eval_runtime": 2.6008,
      "eval_samples_per_second": 1191.932,
      "eval_steps_per_second": 24.992,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.20619478821754456,
      "learning_rate": 0.00011943535523694032,
      "loss": 0.0305,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9383870967741935,
      "eval_loss": 0.023558245971798897,
      "eval_runtime": 2.6497,
      "eval_samples_per_second": 1169.948,
      "eval_steps_per_second": 24.531,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.13936841487884521,
      "learning_rate": 0.00011028355257556492,
      "loss": 0.0246,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9390322580645162,
      "eval_loss": 0.02127992734313011,
      "eval_runtime": 2.6543,
      "eval_samples_per_second": 1167.921,
      "eval_steps_per_second": 24.489,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.09692877531051636,
      "learning_rate": 0.0001011317499141895,
      "loss": 0.0215,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9403225806451613,
      "eval_loss": 0.01949879713356495,
      "eval_runtime": 2.6309,
      "eval_samples_per_second": 1178.314,
      "eval_steps_per_second": 24.707,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.09292752295732498,
      "learning_rate": 9.197994725281409e-05,
      "loss": 0.0197,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9435483870967742,
      "eval_loss": 0.01897437684237957,
      "eval_runtime": 2.6528,
      "eval_samples_per_second": 1168.577,
      "eval_steps_per_second": 24.502,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.07617080211639404,
      "learning_rate": 8.282814459143868e-05,
      "loss": 0.0183,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9441935483870968,
      "eval_loss": 0.018397612497210503,
      "eval_runtime": 2.6546,
      "eval_samples_per_second": 1167.789,
      "eval_steps_per_second": 24.486,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.07391500473022461,
      "learning_rate": 7.367634193006326e-05,
      "loss": 0.0171,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9461290322580646,
      "eval_loss": 0.017443224787712097,
      "eval_runtime": 2.6528,
      "eval_samples_per_second": 1168.563,
      "eval_steps_per_second": 24.502,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.09221338480710983,
      "learning_rate": 6.452453926868784e-05,
      "loss": 0.0162,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9435483870967742,
      "eval_loss": 0.017002489417791367,
      "eval_runtime": 2.6461,
      "eval_samples_per_second": 1171.522,
      "eval_steps_per_second": 24.564,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.07394809275865555,
      "learning_rate": 5.537273660731243e-05,
      "loss": 0.0155,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9429032258064516,
      "eval_loss": 0.01677101105451584,
      "eval_runtime": 2.635,
      "eval_samples_per_second": 1176.456,
      "eval_steps_per_second": 24.668,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.0590691976249218,
      "learning_rate": 4.622093394593702e-05,
      "loss": 0.0149,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9464516129032258,
      "eval_loss": 0.016378937289118767,
      "eval_runtime": 2.6318,
      "eval_samples_per_second": 1177.923,
      "eval_steps_per_second": 24.698,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.05711551755666733,
      "learning_rate": 3.7069131284561605e-05,
      "loss": 0.0143,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9461290322580646,
      "eval_loss": 0.015853669494390488,
      "eval_runtime": 2.6492,
      "eval_samples_per_second": 1170.169,
      "eval_steps_per_second": 24.536,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.05840861052274704,
      "learning_rate": 2.7917328623186196e-05,
      "loss": 0.0137,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9461290322580646,
      "eval_loss": 0.015562719665467739,
      "eval_runtime": 2.6419,
      "eval_samples_per_second": 1173.405,
      "eval_steps_per_second": 24.604,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.0555766336619854,
      "learning_rate": 1.8765525961810784e-05,
      "loss": 0.0133,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9467741935483871,
      "eval_loss": 0.015005242079496384,
      "eval_runtime": 2.6572,
      "eval_samples_per_second": 1166.643,
      "eval_steps_per_second": 24.462,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.05531783401966095,
      "learning_rate": 9.61372330043537e-06,
      "loss": 0.0129,
      "step": 4755
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9461290322580646,
      "eval_loss": 0.01490978617221117,
      "eval_runtime": 2.6507,
      "eval_samples_per_second": 1169.499,
      "eval_steps_per_second": 24.522,
      "step": 4770
    },
    {
      "epoch": 15.949685534591195,
      "grad_norm": 0.05344949662685394,
      "learning_rate": 4.6192063905995774e-07,
      "loss": 0.0125,
      "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": 1319419533829824.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8496454567338295,
    "fp16": false,
    "learning_rate": 0.0001430799179488219,
    "lr_scheduler": "cosine",
    "num_train_epochs": 16,
    "temperature": 12,
    "warmup_steps": 132,
    "weight_decay": 0.20337985521529534
  }
}