{
  "best_metric": 0.262811541557312,
  "best_model_checkpoint": "./output_samples_202302241350_lfbase/checkpoint-6000",
  "epoch": 4.245465071401004,
  "global_step": 22000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "learning_rate": 2.94245465071401e-05,
      "loss": 0.2869,
      "step": 500
    },
    {
      "epoch": 0.19,
      "learning_rate": 2.8845619451949058e-05,
      "loss": 0.1797,
      "step": 1000
    },
    {
      "epoch": 0.19,
      "eval_accuracy": 0.8138957619667053,
      "eval_loss": 0.49187058210372925,
      "eval_runtime": 490.7956,
      "eval_samples_per_second": 119.795,
      "eval_steps_per_second": 1.872,
      "step": 1000
    },
    {
      "epoch": 0.29,
      "learning_rate": 2.826669239675801e-05,
      "loss": 0.1589,
      "step": 1500
    },
    {
      "epoch": 0.39,
      "learning_rate": 2.7687765341566965e-05,
      "loss": 0.1369,
      "step": 2000
    },
    {
      "epoch": 0.39,
      "eval_accuracy": 0.8265668749809265,
      "eval_loss": 0.4808556139469147,
      "eval_runtime": 472.7272,
      "eval_samples_per_second": 124.374,
      "eval_steps_per_second": 1.944,
      "step": 2000
    },
    {
      "epoch": 0.48,
      "learning_rate": 2.71099961404863e-05,
      "loss": 0.128,
      "step": 2500
    },
    {
      "epoch": 0.58,
      "learning_rate": 2.6531069085295255e-05,
      "loss": 0.1274,
      "step": 3000
    },
    {
      "epoch": 0.58,
      "eval_accuracy": 0.8054426312446594,
      "eval_loss": 0.6082525253295898,
      "eval_runtime": 485.5844,
      "eval_samples_per_second": 121.081,
      "eval_steps_per_second": 1.893,
      "step": 3000
    },
    {
      "epoch": 0.68,
      "learning_rate": 2.5953299884214592e-05,
      "loss": 0.1161,
      "step": 3500
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.5374372829023545e-05,
      "loss": 0.1109,
      "step": 4000
    },
    {
      "epoch": 0.77,
      "eval_accuracy": 0.8235734105110168,
      "eval_loss": 0.510552704334259,
      "eval_runtime": 488.158,
      "eval_samples_per_second": 120.443,
      "eval_steps_per_second": 1.883,
      "step": 4000
    },
    {
      "epoch": 0.87,
      "learning_rate": 2.47954457738325e-05,
      "loss": 0.1078,
      "step": 4500
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.4216518718641452e-05,
      "loss": 0.1036,
      "step": 5000
    },
    {
      "epoch": 0.96,
      "eval_accuracy": 0.8834084272384644,
      "eval_loss": 0.312045156955719,
      "eval_runtime": 486.8674,
      "eval_samples_per_second": 120.762,
      "eval_steps_per_second": 1.888,
      "step": 5000
    },
    {
      "epoch": 1.06,
      "learning_rate": 2.3637591663450405e-05,
      "loss": 0.0823,
      "step": 5500
    },
    {
      "epoch": 1.16,
      "learning_rate": 2.305866460825936e-05,
      "loss": 0.0718,
      "step": 6000
    },
    {
      "epoch": 1.16,
      "eval_accuracy": 0.9076281785964966,
      "eval_loss": 0.262811541557312,
      "eval_runtime": 488.7776,
      "eval_samples_per_second": 120.29,
      "eval_steps_per_second": 1.88,
      "step": 6000
    },
    {
      "epoch": 1.25,
      "learning_rate": 2.2479737553068312e-05,
      "loss": 0.0714,
      "step": 6500
    },
    {
      "epoch": 1.35,
      "learning_rate": 2.190081049787727e-05,
      "loss": 0.0729,
      "step": 7000
    },
    {
      "epoch": 1.35,
      "eval_accuracy": 0.8230121731758118,
      "eval_loss": 0.6299644112586975,
      "eval_runtime": 480.4556,
      "eval_samples_per_second": 122.373,
      "eval_steps_per_second": 1.913,
      "step": 7000
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.1321883442686222e-05,
      "loss": 0.0687,
      "step": 7500
    },
    {
      "epoch": 1.54,
      "learning_rate": 2.0742956387495175e-05,
      "loss": 0.0664,
      "step": 8000
    },
    {
      "epoch": 1.54,
      "eval_accuracy": 0.8701080083847046,
      "eval_loss": 0.5767059922218323,
      "eval_runtime": 472.3986,
      "eval_samples_per_second": 124.461,
      "eval_steps_per_second": 1.945,
      "step": 8000
    },
    {
      "epoch": 1.64,
      "learning_rate": 2.0164029332304132e-05,
      "loss": 0.0665,
      "step": 8500
    },
    {
      "epoch": 1.74,
      "learning_rate": 1.9585102277113085e-05,
      "loss": 0.0622,
      "step": 9000
    },
    {
      "epoch": 1.74,
      "eval_accuracy": 0.8938174843788147,
      "eval_loss": 0.46610942482948303,
      "eval_runtime": 472.4306,
      "eval_samples_per_second": 124.452,
      "eval_steps_per_second": 1.945,
      "step": 9000
    },
    {
      "epoch": 1.83,
      "learning_rate": 1.9007333076032422e-05,
      "loss": 0.0639,
      "step": 9500
    },
    {
      "epoch": 1.93,
      "learning_rate": 1.8428406020841375e-05,
      "loss": 0.0638,
      "step": 10000
    },
    {
      "epoch": 1.93,
      "eval_accuracy": 0.913428008556366,
      "eval_loss": 0.27155911922454834,
      "eval_runtime": 484.5228,
      "eval_samples_per_second": 121.346,
      "eval_steps_per_second": 1.897,
      "step": 10000
    },
    {
      "epoch": 2.03,
      "learning_rate": 1.784947896565033e-05,
      "loss": 0.0564,
      "step": 10500
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.7270551910459282e-05,
      "loss": 0.0385,
      "step": 11000
    },
    {
      "epoch": 2.12,
      "eval_accuracy": 0.9116081595420837,
      "eval_loss": 0.38816747069358826,
      "eval_runtime": 488.2826,
      "eval_samples_per_second": 120.412,
      "eval_steps_per_second": 1.882,
      "step": 11000
    },
    {
      "epoch": 2.22,
      "learning_rate": 1.669162485526824e-05,
      "loss": 0.0412,
      "step": 11500
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.6113855654187572e-05,
      "loss": 0.0406,
      "step": 12000
    },
    {
      "epoch": 2.32,
      "eval_accuracy": 0.863440752029419,
      "eval_loss": 0.6258434653282166,
      "eval_runtime": 473.3305,
      "eval_samples_per_second": 124.216,
      "eval_steps_per_second": 1.942,
      "step": 12000
    },
    {
      "epoch": 2.41,
      "learning_rate": 1.553492859899653e-05,
      "loss": 0.0391,
      "step": 12500
    },
    {
      "epoch": 2.51,
      "learning_rate": 1.495600154380548e-05,
      "loss": 0.037,
      "step": 13000
    },
    {
      "epoch": 2.51,
      "eval_accuracy": 0.891011118888855,
      "eval_loss": 0.48729509115219116,
      "eval_runtime": 473.8644,
      "eval_samples_per_second": 124.076,
      "eval_steps_per_second": 1.939,
      "step": 13000
    },
    {
      "epoch": 2.61,
      "learning_rate": 1.4377074488614436e-05,
      "loss": 0.04,
      "step": 13500
    },
    {
      "epoch": 2.7,
      "learning_rate": 1.3798147433423389e-05,
      "loss": 0.0382,
      "step": 14000
    },
    {
      "epoch": 2.7,
      "eval_accuracy": 0.8833404183387756,
      "eval_loss": 0.5917666554450989,
      "eval_runtime": 473.8023,
      "eval_samples_per_second": 124.092,
      "eval_steps_per_second": 1.94,
      "step": 14000
    },
    {
      "epoch": 2.8,
      "learning_rate": 1.3220378232342726e-05,
      "loss": 0.0392,
      "step": 14500
    },
    {
      "epoch": 2.89,
      "learning_rate": 1.2642609031262061e-05,
      "loss": 0.0367,
      "step": 15000
    },
    {
      "epoch": 2.89,
      "eval_accuracy": 0.8793604969978333,
      "eval_loss": 0.569683313369751,
      "eval_runtime": 474.4494,
      "eval_samples_per_second": 123.923,
      "eval_steps_per_second": 1.937,
      "step": 15000
    },
    {
      "epoch": 2.99,
      "learning_rate": 1.2063681976071016e-05,
      "loss": 0.0358,
      "step": 15500
    },
    {
      "epoch": 3.09,
      "learning_rate": 1.148475492087997e-05,
      "loss": 0.0216,
      "step": 16000
    },
    {
      "epoch": 3.09,
      "eval_accuracy": 0.8624373078346252,
      "eval_loss": 0.8281469345092773,
      "eval_runtime": 473.8459,
      "eval_samples_per_second": 124.08,
      "eval_steps_per_second": 1.939,
      "step": 16000
    },
    {
      "epoch": 3.18,
      "learning_rate": 1.0905827865688923e-05,
      "loss": 0.0211,
      "step": 16500
    },
    {
      "epoch": 3.28,
      "learning_rate": 1.032805866460826e-05,
      "loss": 0.0227,
      "step": 17000
    },
    {
      "epoch": 3.28,
      "eval_accuracy": 0.8539331555366516,
      "eval_loss": 0.9032775163650513,
      "eval_runtime": 474.5611,
      "eval_samples_per_second": 123.893,
      "eval_steps_per_second": 1.937,
      "step": 17000
    },
    {
      "epoch": 3.38,
      "learning_rate": 9.749131609417213e-06,
      "loss": 0.0225,
      "step": 17500
    },
    {
      "epoch": 3.47,
      "learning_rate": 9.170204554226166e-06,
      "loss": 0.0216,
      "step": 18000
    },
    {
      "epoch": 3.47,
      "eval_accuracy": 0.888221800327301,
      "eval_loss": 0.6849754452705383,
      "eval_runtime": 487.296,
      "eval_samples_per_second": 120.656,
      "eval_steps_per_second": 1.886,
      "step": 18000
    },
    {
      "epoch": 3.57,
      "learning_rate": 8.591277499035123e-06,
      "loss": 0.0217,
      "step": 18500
    },
    {
      "epoch": 3.67,
      "learning_rate": 8.012350443844076e-06,
      "loss": 0.0199,
      "step": 19000
    },
    {
      "epoch": 3.67,
      "eval_accuracy": 0.8972702026367188,
      "eval_loss": 0.6052428483963013,
      "eval_runtime": 490.4243,
      "eval_samples_per_second": 119.886,
      "eval_steps_per_second": 1.874,
      "step": 19000
    },
    {
      "epoch": 3.76,
      "learning_rate": 7.43342338865303e-06,
      "loss": 0.0207,
      "step": 19500
    },
    {
      "epoch": 3.86,
      "learning_rate": 6.855654187572366e-06,
      "loss": 0.0206,
      "step": 20000
    },
    {
      "epoch": 3.86,
      "eval_accuracy": 0.8873884081840515,
      "eval_loss": 0.6865639090538025,
      "eval_runtime": 487.3484,
      "eval_samples_per_second": 120.643,
      "eval_steps_per_second": 1.886,
      "step": 20000
    },
    {
      "epoch": 3.96,
      "learning_rate": 6.276727132381321e-06,
      "loss": 0.0187,
      "step": 20500
    },
    {
      "epoch": 4.05,
      "learning_rate": 5.697800077190274e-06,
      "loss": 0.0148,
      "step": 21000
    },
    {
      "epoch": 4.05,
      "eval_accuracy": 0.8811633586883545,
      "eval_loss": 0.9089483618736267,
      "eval_runtime": 487.5135,
      "eval_samples_per_second": 120.602,
      "eval_steps_per_second": 1.885,
      "step": 21000
    },
    {
      "epoch": 4.15,
      "learning_rate": 5.118873021999228e-06,
      "loss": 0.0107,
      "step": 21500
    },
    {
      "epoch": 4.25,
      "learning_rate": 4.5399459668081825e-06,
      "loss": 0.0121,
      "step": 22000
    },
    {
      "epoch": 4.25,
      "eval_accuracy": 0.9049409031867981,
      "eval_loss": 0.641972541809082,
      "eval_runtime": 487.5431,
      "eval_samples_per_second": 120.594,
      "eval_steps_per_second": 1.885,
      "step": 22000
    }
  ],
  "max_steps": 25910,
  "num_train_epochs": 5,
  "total_flos": 1.8496468450847293e+18,
  "trial_name": null,
  "trial_params": null
}