{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 6360,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9968553459119497,
      "grad_norm": 0.3776308596134186,
      "learning_rate": 0.00038685123471569015,
      "loss": 0.1977,
      "step": 317
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8580645161290322,
      "eval_loss": 0.06473647803068161,
      "eval_runtime": 2.6632,
      "eval_samples_per_second": 1164.002,
      "eval_steps_per_second": 24.406,
      "step": 318
    },
    {
      "epoch": 1.9937106918238994,
      "grad_norm": 0.25335493683815,
      "learning_rate": 0.0003665580291216353,
      "loss": 0.0555,
      "step": 634
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9019354838709678,
      "eval_loss": 0.05631404370069504,
      "eval_runtime": 2.6288,
      "eval_samples_per_second": 1179.25,
      "eval_steps_per_second": 24.726,
      "step": 636
    },
    {
      "epoch": 2.990566037735849,
      "grad_norm": 0.6674047112464905,
      "learning_rate": 0.0003462648235275804,
      "loss": 0.0401,
      "step": 951
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8874193548387097,
      "eval_loss": 0.05496141314506531,
      "eval_runtime": 2.6441,
      "eval_samples_per_second": 1172.424,
      "eval_steps_per_second": 24.583,
      "step": 954
    },
    {
      "epoch": 3.9874213836477987,
      "grad_norm": 0.38346582651138306,
      "learning_rate": 0.00032597161793352546,
      "loss": 0.0399,
      "step": 1268
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8932258064516129,
      "eval_loss": 0.0530259795486927,
      "eval_runtime": 2.6688,
      "eval_samples_per_second": 1161.578,
      "eval_steps_per_second": 24.356,
      "step": 1272
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.10067661851644516,
      "learning_rate": 0.00030567841233947055,
      "loss": 0.0336,
      "step": 1585
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9161290322580645,
      "eval_loss": 0.04296305030584335,
      "eval_runtime": 2.6372,
      "eval_samples_per_second": 1175.495,
      "eval_steps_per_second": 24.647,
      "step": 1590
    },
    {
      "epoch": 5.981132075471698,
      "grad_norm": 0.22013570368289948,
      "learning_rate": 0.0002853852067454157,
      "loss": 0.0287,
      "step": 1902
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9116129032258065,
      "eval_loss": 0.048792969435453415,
      "eval_runtime": 2.649,
      "eval_samples_per_second": 1170.263,
      "eval_steps_per_second": 24.538,
      "step": 1908
    },
    {
      "epoch": 6.977987421383648,
      "grad_norm": 0.08923449367284775,
      "learning_rate": 0.0002650920011513608,
      "loss": 0.0255,
      "step": 2219
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.912258064516129,
      "eval_loss": 0.044845063239336014,
      "eval_runtime": 2.6493,
      "eval_samples_per_second": 1170.106,
      "eval_steps_per_second": 24.534,
      "step": 2226
    },
    {
      "epoch": 7.9748427672955975,
      "grad_norm": 0.16984495520591736,
      "learning_rate": 0.00024479879555730586,
      "loss": 0.0245,
      "step": 2536
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.917741935483871,
      "eval_loss": 0.04449814185500145,
      "eval_runtime": 2.6634,
      "eval_samples_per_second": 1163.909,
      "eval_steps_per_second": 24.405,
      "step": 2544
    },
    {
      "epoch": 8.971698113207546,
      "grad_norm": 0.5007947087287903,
      "learning_rate": 0.00022450558996325092,
      "loss": 0.0244,
      "step": 2853
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9183870967741935,
      "eval_loss": 0.04302350431680679,
      "eval_runtime": 2.643,
      "eval_samples_per_second": 1172.907,
      "eval_steps_per_second": 24.593,
      "step": 2862
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.06756994128227234,
      "learning_rate": 0.00020421238436919606,
      "loss": 0.0219,
      "step": 3170
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9238709677419354,
      "eval_loss": 0.03625302389264107,
      "eval_runtime": 2.6402,
      "eval_samples_per_second": 1174.161,
      "eval_steps_per_second": 24.62,
      "step": 3180
    },
    {
      "epoch": 10.965408805031446,
      "grad_norm": 0.7381525635719299,
      "learning_rate": 0.00018391917877514115,
      "loss": 0.0199,
      "step": 3487
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9287096774193548,
      "eval_loss": 0.03676296025514603,
      "eval_runtime": 2.6489,
      "eval_samples_per_second": 1170.314,
      "eval_steps_per_second": 24.539,
      "step": 3498
    },
    {
      "epoch": 11.962264150943396,
      "grad_norm": 0.051013097167015076,
      "learning_rate": 0.00016362597318108623,
      "loss": 0.0184,
      "step": 3804
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9264516129032258,
      "eval_loss": 0.03529633954167366,
      "eval_runtime": 2.6771,
      "eval_samples_per_second": 1157.95,
      "eval_steps_per_second": 24.28,
      "step": 3816
    },
    {
      "epoch": 12.959119496855346,
      "grad_norm": 0.07709548622369766,
      "learning_rate": 0.00014333276758703132,
      "loss": 0.0171,
      "step": 4121
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9306451612903226,
      "eval_loss": 0.03273295238614082,
      "eval_runtime": 2.6553,
      "eval_samples_per_second": 1167.463,
      "eval_steps_per_second": 24.479,
      "step": 4134
    },
    {
      "epoch": 13.955974842767295,
      "grad_norm": 0.13161151111125946,
      "learning_rate": 0.0001230395619929764,
      "loss": 0.016,
      "step": 4438
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9309677419354838,
      "eval_loss": 0.03127666935324669,
      "eval_runtime": 2.6519,
      "eval_samples_per_second": 1168.993,
      "eval_steps_per_second": 24.511,
      "step": 4452
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.05098772048950195,
      "learning_rate": 0.00010274635639892152,
      "loss": 0.0151,
      "step": 4755
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9348387096774193,
      "eval_loss": 0.030557256191968918,
      "eval_runtime": 2.6688,
      "eval_samples_per_second": 1161.571,
      "eval_steps_per_second": 24.356,
      "step": 4770
    },
    {
      "epoch": 15.949685534591195,
      "grad_norm": 0.04677645117044449,
      "learning_rate": 8.245315080486662e-05,
      "loss": 0.0143,
      "step": 5072
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9338709677419355,
      "eval_loss": 0.03013915941119194,
      "eval_runtime": 2.6585,
      "eval_samples_per_second": 1166.065,
      "eval_steps_per_second": 24.45,
      "step": 5088
    },
    {
      "epoch": 16.946540880503143,
      "grad_norm": 0.05656523257493973,
      "learning_rate": 6.21599452108117e-05,
      "loss": 0.0138,
      "step": 5389
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9332258064516129,
      "eval_loss": 0.028866084292531013,
      "eval_runtime": 2.6651,
      "eval_samples_per_second": 1163.197,
      "eval_steps_per_second": 24.39,
      "step": 5406
    },
    {
      "epoch": 17.943396226415093,
      "grad_norm": 0.047965776175260544,
      "learning_rate": 4.1866739616756805e-05,
      "loss": 0.0132,
      "step": 5706
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9329032258064516,
      "eval_loss": 0.02866896614432335,
      "eval_runtime": 2.6577,
      "eval_samples_per_second": 1166.402,
      "eval_steps_per_second": 24.457,
      "step": 5724
    },
    {
      "epoch": 18.940251572327043,
      "grad_norm": 0.04392002522945404,
      "learning_rate": 2.15735340227019e-05,
      "loss": 0.0128,
      "step": 6023
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9338709677419355,
      "eval_loss": 0.02806813083589077,
      "eval_runtime": 2.6852,
      "eval_samples_per_second": 1154.456,
      "eval_steps_per_second": 24.206,
      "step": 6042
    },
    {
      "epoch": 19.937106918238992,
      "grad_norm": 0.043970681726932526,
      "learning_rate": 1.280328428646997e-06,
      "loss": 0.0124,
      "step": 6340
    }
  ],
  "logging_steps": 317,
  "max_steps": 6360,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "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": 1648461417340164.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.040278370777020456,
    "fp16": true,
    "learning_rate": 0.0004035595207095335,
    "lr_scheduler": "linear",
    "num_train_epochs": 20,
    "temperature": 19,
    "warmup_steps": 56,
    "weight_decay": 0.14418853843556578
  }
}