{
  "best_metric": 85.6444218033622,
  "best_model_checkpoint": "outputs/bert-large/weak_tiny_poe/mnli_21/checkpoint-36816",
  "epoch": 3.0,
  "global_step": 36816,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "learning_rate": 1.9728378965667103e-05,
      "loss": 1.039,
      "step": 500
    },
    {
      "epoch": 0.08,
      "learning_rate": 1.9456757931334205e-05,
      "loss": 0.9373,
      "step": 1000
    },
    {
      "epoch": 0.12,
      "learning_rate": 1.9185136897001307e-05,
      "loss": 0.8501,
      "step": 1500
    },
    {
      "epoch": 0.16,
      "learning_rate": 1.8913515862668405e-05,
      "loss": 0.8176,
      "step": 2000
    },
    {
      "epoch": 0.2,
      "learning_rate": 1.8641894828335507e-05,
      "loss": 0.7608,
      "step": 2500
    },
    {
      "epoch": 0.24,
      "learning_rate": 1.837027379400261e-05,
      "loss": 0.7572,
      "step": 3000
    },
    {
      "epoch": 0.29,
      "learning_rate": 1.809865275966971e-05,
      "loss": 0.7447,
      "step": 3500
    },
    {
      "epoch": 0.33,
      "learning_rate": 1.7827031725336812e-05,
      "loss": 0.743,
      "step": 4000
    },
    {
      "epoch": 0.37,
      "learning_rate": 1.7555410691003914e-05,
      "loss": 0.721,
      "step": 4500
    },
    {
      "epoch": 0.41,
      "learning_rate": 1.7283789656671015e-05,
      "loss": 0.727,
      "step": 5000
    },
    {
      "epoch": 0.45,
      "learning_rate": 1.7012168622338114e-05,
      "loss": 0.7244,
      "step": 5500
    },
    {
      "epoch": 0.49,
      "learning_rate": 1.6740547588005215e-05,
      "loss": 0.7052,
      "step": 6000
    },
    {
      "epoch": 0.53,
      "learning_rate": 1.6468926553672317e-05,
      "loss": 0.7041,
      "step": 6500
    },
    {
      "epoch": 0.57,
      "learning_rate": 1.619730551933942e-05,
      "loss": 0.6996,
      "step": 7000
    },
    {
      "epoch": 0.61,
      "learning_rate": 1.592568448500652e-05,
      "loss": 0.7108,
      "step": 7500
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.5654063450673622e-05,
      "loss": 0.6975,
      "step": 8000
    },
    {
      "epoch": 0.69,
      "learning_rate": 1.5382442416340724e-05,
      "loss": 0.6902,
      "step": 8500
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.5110821382007822e-05,
      "loss": 0.7043,
      "step": 9000
    },
    {
      "epoch": 0.77,
      "learning_rate": 1.4839200347674924e-05,
      "loss": 0.6857,
      "step": 9500
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.4567579313342026e-05,
      "loss": 0.6969,
      "step": 10000
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4295958279009128e-05,
      "loss": 0.6837,
      "step": 10500
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.402433724467623e-05,
      "loss": 0.6917,
      "step": 11000
    },
    {
      "epoch": 0.94,
      "learning_rate": 1.3752716210343331e-05,
      "loss": 0.6871,
      "step": 11500
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3481095176010431e-05,
      "loss": 0.6809,
      "step": 12000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 84.2384105960265,
      "eval_combined_score": 84.2384105960265,
      "eval_loss": 0.4313635230064392,
      "eval_runtime": 123.7058,
      "eval_samples_per_second": 79.341,
      "eval_steps_per_second": 0.315,
      "step": 12272
    },
    {
      "epoch": 1.02,
      "learning_rate": 1.3209474141677531e-05,
      "loss": 0.6063,
      "step": 12500
    },
    {
      "epoch": 1.06,
      "learning_rate": 1.2937853107344633e-05,
      "loss": 0.5482,
      "step": 13000
    },
    {
      "epoch": 1.1,
      "learning_rate": 1.2666232073011735e-05,
      "loss": 0.5298,
      "step": 13500
    },
    {
      "epoch": 1.14,
      "learning_rate": 1.2394611038678836e-05,
      "loss": 0.547,
      "step": 14000
    },
    {
      "epoch": 1.18,
      "learning_rate": 1.2122990004345938e-05,
      "loss": 0.534,
      "step": 14500
    },
    {
      "epoch": 1.22,
      "learning_rate": 1.185136897001304e-05,
      "loss": 0.552,
      "step": 15000
    },
    {
      "epoch": 1.26,
      "learning_rate": 1.157974793568014e-05,
      "loss": 0.534,
      "step": 15500
    },
    {
      "epoch": 1.3,
      "learning_rate": 1.130812690134724e-05,
      "loss": 0.5504,
      "step": 16000
    },
    {
      "epoch": 1.34,
      "learning_rate": 1.1036505867014341e-05,
      "loss": 0.5388,
      "step": 16500
    },
    {
      "epoch": 1.39,
      "learning_rate": 1.0764884832681443e-05,
      "loss": 0.5439,
      "step": 17000
    },
    {
      "epoch": 1.43,
      "learning_rate": 1.0493263798348545e-05,
      "loss": 0.535,
      "step": 17500
    },
    {
      "epoch": 1.47,
      "learning_rate": 1.0221642764015647e-05,
      "loss": 0.5426,
      "step": 18000
    },
    {
      "epoch": 1.51,
      "learning_rate": 9.950021729682747e-06,
      "loss": 0.5432,
      "step": 18500
    },
    {
      "epoch": 1.55,
      "learning_rate": 9.678400695349848e-06,
      "loss": 0.5337,
      "step": 19000
    },
    {
      "epoch": 1.59,
      "learning_rate": 9.40677966101695e-06,
      "loss": 0.5384,
      "step": 19500
    },
    {
      "epoch": 1.63,
      "learning_rate": 9.13515862668405e-06,
      "loss": 0.533,
      "step": 20000
    },
    {
      "epoch": 1.67,
      "learning_rate": 8.863537592351152e-06,
      "loss": 0.5301,
      "step": 20500
    },
    {
      "epoch": 1.71,
      "learning_rate": 8.591916558018254e-06,
      "loss": 0.5391,
      "step": 21000
    },
    {
      "epoch": 1.75,
      "learning_rate": 8.320295523685355e-06,
      "loss": 0.5351,
      "step": 21500
    },
    {
      "epoch": 1.79,
      "learning_rate": 8.048674489352455e-06,
      "loss": 0.5348,
      "step": 22000
    },
    {
      "epoch": 1.83,
      "learning_rate": 7.777053455019557e-06,
      "loss": 0.5541,
      "step": 22500
    },
    {
      "epoch": 1.87,
      "learning_rate": 7.505432420686659e-06,
      "loss": 0.5424,
      "step": 23000
    },
    {
      "epoch": 1.91,
      "learning_rate": 7.23381138635376e-06,
      "loss": 0.5276,
      "step": 23500
    },
    {
      "epoch": 1.96,
      "learning_rate": 6.962190352020861e-06,
      "loss": 0.535,
      "step": 24000
    },
    {
      "epoch": 2.0,
      "learning_rate": 6.690569317687963e-06,
      "loss": 0.5339,
      "step": 24500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 85.1146204788589,
      "eval_combined_score": 85.1146204788589,
      "eval_loss": 0.43441787362098694,
      "eval_runtime": 123.3351,
      "eval_samples_per_second": 79.58,
      "eval_steps_per_second": 0.316,
      "step": 24544
    },
    {
      "epoch": 2.04,
      "learning_rate": 6.418948283355063e-06,
      "loss": 0.4181,
      "step": 25000
    },
    {
      "epoch": 2.08,
      "learning_rate": 6.147327249022165e-06,
      "loss": 0.4102,
      "step": 25500
    },
    {
      "epoch": 2.12,
      "learning_rate": 5.8757062146892665e-06,
      "loss": 0.4134,
      "step": 26000
    },
    {
      "epoch": 2.16,
      "learning_rate": 5.6040851803563665e-06,
      "loss": 0.4213,
      "step": 26500
    },
    {
      "epoch": 2.2,
      "learning_rate": 5.332464146023468e-06,
      "loss": 0.4175,
      "step": 27000
    },
    {
      "epoch": 2.24,
      "learning_rate": 5.06084311169057e-06,
      "loss": 0.4067,
      "step": 27500
    },
    {
      "epoch": 2.28,
      "learning_rate": 4.789222077357671e-06,
      "loss": 0.4142,
      "step": 28000
    },
    {
      "epoch": 2.32,
      "learning_rate": 4.5176010430247726e-06,
      "loss": 0.4101,
      "step": 28500
    },
    {
      "epoch": 2.36,
      "learning_rate": 4.2459800086918734e-06,
      "loss": 0.4069,
      "step": 29000
    },
    {
      "epoch": 2.4,
      "learning_rate": 3.974358974358974e-06,
      "loss": 0.4076,
      "step": 29500
    },
    {
      "epoch": 2.44,
      "learning_rate": 3.702737940026076e-06,
      "loss": 0.3995,
      "step": 30000
    },
    {
      "epoch": 2.49,
      "learning_rate": 3.431116905693177e-06,
      "loss": 0.4072,
      "step": 30500
    },
    {
      "epoch": 2.53,
      "learning_rate": 3.1594958713602786e-06,
      "loss": 0.4032,
      "step": 31000
    },
    {
      "epoch": 2.57,
      "learning_rate": 2.8878748370273795e-06,
      "loss": 0.406,
      "step": 31500
    },
    {
      "epoch": 2.61,
      "learning_rate": 2.6162538026944812e-06,
      "loss": 0.407,
      "step": 32000
    },
    {
      "epoch": 2.65,
      "learning_rate": 2.344632768361582e-06,
      "loss": 0.4135,
      "step": 32500
    },
    {
      "epoch": 2.69,
      "learning_rate": 2.0730117340286834e-06,
      "loss": 0.4042,
      "step": 33000
    },
    {
      "epoch": 2.73,
      "learning_rate": 1.8013906996957845e-06,
      "loss": 0.4111,
      "step": 33500
    },
    {
      "epoch": 2.77,
      "learning_rate": 1.5297696653628858e-06,
      "loss": 0.3987,
      "step": 34000
    },
    {
      "epoch": 2.81,
      "learning_rate": 1.258148631029987e-06,
      "loss": 0.4094,
      "step": 34500
    },
    {
      "epoch": 2.85,
      "learning_rate": 9.865275966970884e-07,
      "loss": 0.3965,
      "step": 35000
    },
    {
      "epoch": 2.89,
      "learning_rate": 7.149065623641896e-07,
      "loss": 0.4046,
      "step": 35500
    },
    {
      "epoch": 2.93,
      "learning_rate": 4.432855280312908e-07,
      "loss": 0.4016,
      "step": 36000
    },
    {
      "epoch": 2.97,
      "learning_rate": 1.7166449369839202e-07,
      "loss": 0.3946,
      "step": 36500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 85.6444218033622,
      "eval_combined_score": 85.6444218033622,
      "eval_loss": 0.5070326328277588,
      "eval_runtime": 121.7901,
      "eval_samples_per_second": 80.589,
      "eval_steps_per_second": 0.32,
      "step": 36816
    },
    {
      "epoch": 3.0,
      "step": 36816,
      "total_flos": 0.0,
      "train_loss": 0.5627925765871846,
      "train_runtime": 42917.181,
      "train_samples_per_second": 27.451,
      "train_steps_per_second": 0.858
    }
  ],
  "max_steps": 36816,
  "num_train_epochs": 3,
  "total_flos": 0.0,
  "trial_name": null,
  "trial_params": null
}