{
  "best_metric": 0.5570505857467651,
  "best_model_checkpoint": "/data/jcanete/all_results/xnli/distillbeto/epochs_3_bs_32_lr_5e-5/checkpoint-24000",
  "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": 4.9323663624511086e-05,
      "loss": 0.9001,
      "step": 500
    },
    {
      "epoch": 0.08,
      "learning_rate": 4.864461103867884e-05,
      "loss": 0.8162,
      "step": 1000
    },
    {
      "epoch": 0.12,
      "learning_rate": 4.7965558452846594e-05,
      "loss": 0.7821,
      "step": 1500
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.728650586701434e-05,
      "loss": 0.742,
      "step": 2000
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.7164658634538152,
      "eval_loss": 0.689033567905426,
      "eval_runtime": 0.7191,
      "eval_samples_per_second": 3462.677,
      "eval_steps_per_second": 108.469,
      "step": 2000
    },
    {
      "epoch": 0.2,
      "learning_rate": 4.660881138635376e-05,
      "loss": 0.7456,
      "step": 2500
    },
    {
      "epoch": 0.24,
      "learning_rate": 4.5929758800521515e-05,
      "loss": 0.7257,
      "step": 3000
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.525070621468927e-05,
      "loss": 0.7183,
      "step": 3500
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.4571653628857017e-05,
      "loss": 0.6993,
      "step": 4000
    },
    {
      "epoch": 0.33,
      "eval_accuracy": 0.7176706827309237,
      "eval_loss": 0.6759888529777527,
      "eval_runtime": 0.7171,
      "eval_samples_per_second": 3472.257,
      "eval_steps_per_second": 108.77,
      "step": 4000
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.389260104302477e-05,
      "loss": 0.6877,
      "step": 4500
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.3214906562364196e-05,
      "loss": 0.6886,
      "step": 5000
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.2535853976531944e-05,
      "loss": 0.6822,
      "step": 5500
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.18568013906997e-05,
      "loss": 0.6738,
      "step": 6000
    },
    {
      "epoch": 0.49,
      "eval_accuracy": 0.7369477911646586,
      "eval_loss": 0.6459766030311584,
      "eval_runtime": 0.714,
      "eval_samples_per_second": 3487.194,
      "eval_steps_per_second": 109.237,
      "step": 6000
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.117774880486745e-05,
      "loss": 0.6721,
      "step": 6500
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.049869621903521e-05,
      "loss": 0.6681,
      "step": 7000
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.981964363320296e-05,
      "loss": 0.6638,
      "step": 7500
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.914059104737071e-05,
      "loss": 0.6511,
      "step": 8000
    },
    {
      "epoch": 0.65,
      "eval_accuracy": 0.7421686746987952,
      "eval_loss": 0.6142069101333618,
      "eval_runtime": 0.7168,
      "eval_samples_per_second": 3473.794,
      "eval_steps_per_second": 108.818,
      "step": 8000
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.846153846153846e-05,
      "loss": 0.6498,
      "step": 8500
    },
    {
      "epoch": 0.73,
      "learning_rate": 3.778384398087788e-05,
      "loss": 0.6465,
      "step": 9000
    },
    {
      "epoch": 0.77,
      "learning_rate": 3.7106149500217294e-05,
      "loss": 0.6422,
      "step": 9500
    },
    {
      "epoch": 0.81,
      "learning_rate": 3.642709691438505e-05,
      "loss": 0.6377,
      "step": 10000
    },
    {
      "epoch": 0.81,
      "eval_accuracy": 0.736144578313253,
      "eval_loss": 0.6204584240913391,
      "eval_runtime": 0.7121,
      "eval_samples_per_second": 3496.5,
      "eval_steps_per_second": 109.529,
      "step": 10000
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.57480443285528e-05,
      "loss": 0.6408,
      "step": 10500
    },
    {
      "epoch": 0.9,
      "learning_rate": 3.5068991742720556e-05,
      "loss": 0.6316,
      "step": 11000
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.438993915688831e-05,
      "loss": 0.6274,
      "step": 11500
    },
    {
      "epoch": 0.98,
      "learning_rate": 3.371224467622773e-05,
      "loss": 0.6332,
      "step": 12000
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.7602409638554217,
      "eval_loss": 0.583670437335968,
      "eval_runtime": 0.7748,
      "eval_samples_per_second": 3213.841,
      "eval_steps_per_second": 100.675,
      "step": 12000
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.3033192090395484e-05,
      "loss": 0.5784,
      "step": 12500
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.235413950456324e-05,
      "loss": 0.5255,
      "step": 13000
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.167508691873099e-05,
      "loss": 0.5146,
      "step": 13500
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.0997392438070404e-05,
      "loss": 0.5279,
      "step": 14000
    },
    {
      "epoch": 1.14,
      "eval_accuracy": 0.7594377510040161,
      "eval_loss": 0.5976436138153076,
      "eval_runtime": 0.7087,
      "eval_samples_per_second": 3513.652,
      "eval_steps_per_second": 110.066,
      "step": 14000
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.031833985223816e-05,
      "loss": 0.5312,
      "step": 14500
    },
    {
      "epoch": 1.22,
      "learning_rate": 2.9639287266405913e-05,
      "loss": 0.5353,
      "step": 15000
    },
    {
      "epoch": 1.26,
      "learning_rate": 2.8960234680573667e-05,
      "loss": 0.5151,
      "step": 15500
    },
    {
      "epoch": 1.3,
      "learning_rate": 2.828118209474142e-05,
      "loss": 0.516,
      "step": 16000
    },
    {
      "epoch": 1.3,
      "eval_accuracy": 0.7686746987951807,
      "eval_loss": 0.5902267098426819,
      "eval_runtime": 0.7185,
      "eval_samples_per_second": 3465.794,
      "eval_steps_per_second": 108.567,
      "step": 16000
    },
    {
      "epoch": 1.34,
      "learning_rate": 2.7603487614080837e-05,
      "loss": 0.5146,
      "step": 16500
    },
    {
      "epoch": 1.39,
      "learning_rate": 2.692443502824859e-05,
      "loss": 0.5205,
      "step": 17000
    },
    {
      "epoch": 1.43,
      "learning_rate": 2.6245382442416345e-05,
      "loss": 0.5209,
      "step": 17500
    },
    {
      "epoch": 1.47,
      "learning_rate": 2.5566329856584093e-05,
      "loss": 0.5072,
      "step": 18000
    },
    {
      "epoch": 1.47,
      "eval_accuracy": 0.7622489959839357,
      "eval_loss": 0.6036831736564636,
      "eval_runtime": 0.7061,
      "eval_samples_per_second": 3526.558,
      "eval_steps_per_second": 110.47,
      "step": 18000
    },
    {
      "epoch": 1.51,
      "learning_rate": 2.488727727075185e-05,
      "loss": 0.5105,
      "step": 18500
    },
    {
      "epoch": 1.55,
      "learning_rate": 2.4209582790091266e-05,
      "loss": 0.5132,
      "step": 19000
    },
    {
      "epoch": 1.59,
      "learning_rate": 2.353053020425902e-05,
      "loss": 0.5149,
      "step": 19500
    },
    {
      "epoch": 1.63,
      "learning_rate": 2.285147761842677e-05,
      "loss": 0.5144,
      "step": 20000
    },
    {
      "epoch": 1.63,
      "eval_accuracy": 0.7622489959839357,
      "eval_loss": 0.5975185632705688,
      "eval_runtime": 0.7116,
      "eval_samples_per_second": 3499.181,
      "eval_steps_per_second": 109.613,
      "step": 20000
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.2172425032594525e-05,
      "loss": 0.5161,
      "step": 20500
    },
    {
      "epoch": 1.71,
      "learning_rate": 2.1493372446762276e-05,
      "loss": 0.5122,
      "step": 21000
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.081431986093003e-05,
      "loss": 0.5052,
      "step": 21500
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.0135267275097785e-05,
      "loss": 0.502,
      "step": 22000
    },
    {
      "epoch": 1.79,
      "eval_accuracy": 0.7803212851405622,
      "eval_loss": 0.5712254047393799,
      "eval_runtime": 0.712,
      "eval_samples_per_second": 3497.159,
      "eval_steps_per_second": 109.55,
      "step": 22000
    },
    {
      "epoch": 1.83,
      "learning_rate": 1.94575727944372e-05,
      "loss": 0.5084,
      "step": 22500
    },
    {
      "epoch": 1.87,
      "learning_rate": 1.8778520208604955e-05,
      "loss": 0.5006,
      "step": 23000
    },
    {
      "epoch": 1.91,
      "learning_rate": 1.809946762277271e-05,
      "loss": 0.498,
      "step": 23500
    },
    {
      "epoch": 1.96,
      "learning_rate": 1.7420415036940463e-05,
      "loss": 0.4975,
      "step": 24000
    },
    {
      "epoch": 1.96,
      "eval_accuracy": 0.7730923694779116,
      "eval_loss": 0.5570505857467651,
      "eval_runtime": 0.7086,
      "eval_samples_per_second": 3514.146,
      "eval_steps_per_second": 110.082,
      "step": 24000
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.6741362451108214e-05,
      "loss": 0.506,
      "step": 24500
    },
    {
      "epoch": 2.04,
      "learning_rate": 1.6063667970447633e-05,
      "loss": 0.4089,
      "step": 25000
    },
    {
      "epoch": 2.08,
      "learning_rate": 1.5384615384615387e-05,
      "loss": 0.3805,
      "step": 25500
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.4705562798783138e-05,
      "loss": 0.3925,
      "step": 26000
    },
    {
      "epoch": 2.12,
      "eval_accuracy": 0.7698795180722892,
      "eval_loss": 0.6058809161186218,
      "eval_runtime": 0.7123,
      "eval_samples_per_second": 3495.574,
      "eval_steps_per_second": 109.5,
      "step": 26000
    },
    {
      "epoch": 2.16,
      "learning_rate": 1.402651021295089e-05,
      "loss": 0.3711,
      "step": 26500
    },
    {
      "epoch": 2.2,
      "learning_rate": 1.3347457627118645e-05,
      "loss": 0.3735,
      "step": 27000
    },
    {
      "epoch": 2.24,
      "learning_rate": 1.2669763146458064e-05,
      "loss": 0.3843,
      "step": 27500
    },
    {
      "epoch": 2.28,
      "learning_rate": 1.1990710560625816e-05,
      "loss": 0.3771,
      "step": 28000
    },
    {
      "epoch": 2.28,
      "eval_accuracy": 0.7746987951807229,
      "eval_loss": 0.6026178598403931,
      "eval_runtime": 0.7183,
      "eval_samples_per_second": 3466.55,
      "eval_steps_per_second": 108.591,
      "step": 28000
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.1311657974793569e-05,
      "loss": 0.3781,
      "step": 28500
    },
    {
      "epoch": 2.36,
      "learning_rate": 1.0632605388961322e-05,
      "loss": 0.3877,
      "step": 29000
    },
    {
      "epoch": 2.4,
      "learning_rate": 9.95491090830074e-06,
      "loss": 0.3759,
      "step": 29500
    },
    {
      "epoch": 2.44,
      "learning_rate": 9.275858322468491e-06,
      "loss": 0.3784,
      "step": 30000
    },
    {
      "epoch": 2.44,
      "eval_accuracy": 0.7855421686746988,
      "eval_loss": 0.6055970788002014,
      "eval_runtime": 0.7082,
      "eval_samples_per_second": 3516.156,
      "eval_steps_per_second": 110.145,
      "step": 30000
    },
    {
      "epoch": 2.49,
      "learning_rate": 8.596805736636245e-06,
      "loss": 0.3774,
      "step": 30500
    },
    {
      "epoch": 2.53,
      "learning_rate": 7.917753150803998e-06,
      "loss": 0.3783,
      "step": 31000
    },
    {
      "epoch": 2.57,
      "learning_rate": 7.240058670143417e-06,
      "loss": 0.3732,
      "step": 31500
    },
    {
      "epoch": 2.61,
      "learning_rate": 6.561006084311169e-06,
      "loss": 0.3863,
      "step": 32000
    },
    {
      "epoch": 2.61,
      "eval_accuracy": 0.7742971887550201,
      "eval_loss": 0.5977765321731567,
      "eval_runtime": 0.7194,
      "eval_samples_per_second": 3461.392,
      "eval_steps_per_second": 108.429,
      "step": 32000
    },
    {
      "epoch": 2.65,
      "learning_rate": 5.881953498478923e-06,
      "loss": 0.3817,
      "step": 32500
    },
    {
      "epoch": 2.69,
      "learning_rate": 5.2029009126466755e-06,
      "loss": 0.3748,
      "step": 33000
    },
    {
      "epoch": 2.73,
      "learning_rate": 4.525206431986093e-06,
      "loss": 0.3746,
      "step": 33500
    },
    {
      "epoch": 2.77,
      "learning_rate": 3.846153846153847e-06,
      "loss": 0.377,
      "step": 34000
    },
    {
      "epoch": 2.77,
      "eval_accuracy": 0.7755020080321285,
      "eval_loss": 0.6021597385406494,
      "eval_runtime": 0.71,
      "eval_samples_per_second": 3507.063,
      "eval_steps_per_second": 109.86,
      "step": 34000
    },
    {
      "epoch": 2.81,
      "learning_rate": 3.1671012603216e-06,
      "loss": 0.3745,
      "step": 34500
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.488048674489353e-06,
      "loss": 0.3798,
      "step": 35000
    },
    {
      "epoch": 2.89,
      "learning_rate": 1.8089960886571057e-06,
      "loss": 0.3751,
      "step": 35500
    },
    {
      "epoch": 2.93,
      "learning_rate": 1.1313016079965233e-06,
      "loss": 0.3602,
      "step": 36000
    },
    {
      "epoch": 2.93,
      "eval_accuracy": 0.7742971887550201,
      "eval_loss": 0.6019901633262634,
      "eval_runtime": 0.7094,
      "eval_samples_per_second": 3510.147,
      "eval_steps_per_second": 109.956,
      "step": 36000
    },
    {
      "epoch": 2.97,
      "learning_rate": 4.522490221642764e-07,
      "loss": 0.3712,
      "step": 36500
    },
    {
      "epoch": 3.0,
      "step": 36816,
      "total_flos": 2.996060719687747e+16,
      "train_loss": 0.5278426528443259,
      "train_runtime": 4215.3344,
      "train_samples_per_second": 279.481,
      "train_steps_per_second": 8.734
    }
  ],
  "max_steps": 36816,
  "num_train_epochs": 3,
  "total_flos": 2.996060719687747e+16,
  "trial_name": null,
  "trial_params": null
}