{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 32919,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 4.92405601628239e-05,
      "loss": 3.2169,
      "step": 500
    },
    {
      "epoch": 0.09,
      "learning_rate": 4.848112032564781e-05,
      "loss": 2.872,
      "step": 1000
    },
    {
      "epoch": 0.14,
      "learning_rate": 4.77216804884717e-05,
      "loss": 2.6524,
      "step": 1500
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.6962240651295605e-05,
      "loss": 2.5352,
      "step": 2000
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.620280081411951e-05,
      "loss": 2.4293,
      "step": 2500
    },
    {
      "epoch": 0.27,
      "learning_rate": 4.544336097694341e-05,
      "loss": 2.3569,
      "step": 3000
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.468392113976731e-05,
      "loss": 2.2906,
      "step": 3500
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.392448130259121e-05,
      "loss": 2.2189,
      "step": 4000
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.316504146541511e-05,
      "loss": 2.1845,
      "step": 4500
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.240560162823901e-05,
      "loss": 2.1222,
      "step": 5000
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.1646161791062915e-05,
      "loss": 2.0704,
      "step": 5500
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.088672195388681e-05,
      "loss": 2.0429,
      "step": 6000
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.012728211671072e-05,
      "loss": 1.9951,
      "step": 6500
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.9367842279534616e-05,
      "loss": 1.9488,
      "step": 7000
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.860840244235852e-05,
      "loss": 1.9073,
      "step": 7500
    },
    {
      "epoch": 0.73,
      "learning_rate": 3.784896260518242e-05,
      "loss": 1.8649,
      "step": 8000
    },
    {
      "epoch": 0.77,
      "learning_rate": 3.7089522768006316e-05,
      "loss": 1.8362,
      "step": 8500
    },
    {
      "epoch": 0.82,
      "learning_rate": 3.6330082930830225e-05,
      "loss": 1.8117,
      "step": 9000
    },
    {
      "epoch": 0.87,
      "learning_rate": 3.557064309365412e-05,
      "loss": 1.7883,
      "step": 9500
    },
    {
      "epoch": 0.91,
      "learning_rate": 3.4811203256478024e-05,
      "loss": 1.7724,
      "step": 10000
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.4051763419301926e-05,
      "loss": 1.7215,
      "step": 10500
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.329232358212583e-05,
      "loss": 1.671,
      "step": 11000
    },
    {
      "epoch": 1.05,
      "learning_rate": 3.2532883744949724e-05,
      "loss": 1.542,
      "step": 11500
    },
    {
      "epoch": 1.09,
      "learning_rate": 3.1773443907773626e-05,
      "loss": 1.5211,
      "step": 12000
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.101400407059753e-05,
      "loss": 1.537,
      "step": 12500
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.0254564233421428e-05,
      "loss": 1.519,
      "step": 13000
    },
    {
      "epoch": 1.23,
      "learning_rate": 2.9495124396245334e-05,
      "loss": 1.4805,
      "step": 13500
    },
    {
      "epoch": 1.28,
      "learning_rate": 2.8735684559069233e-05,
      "loss": 1.467,
      "step": 14000
    },
    {
      "epoch": 1.32,
      "learning_rate": 2.7976244721893135e-05,
      "loss": 1.4591,
      "step": 14500
    },
    {
      "epoch": 1.37,
      "learning_rate": 2.7216804884717034e-05,
      "loss": 1.4131,
      "step": 15000
    },
    {
      "epoch": 1.41,
      "learning_rate": 2.6457365047540933e-05,
      "loss": 1.4217,
      "step": 15500
    },
    {
      "epoch": 1.46,
      "learning_rate": 2.5697925210364836e-05,
      "loss": 1.3952,
      "step": 16000
    },
    {
      "epoch": 1.5,
      "learning_rate": 2.4938485373188738e-05,
      "loss": 1.3792,
      "step": 16500
    },
    {
      "epoch": 1.55,
      "learning_rate": 2.4179045536012637e-05,
      "loss": 1.3583,
      "step": 17000
    },
    {
      "epoch": 1.59,
      "learning_rate": 2.341960569883654e-05,
      "loss": 1.3674,
      "step": 17500
    },
    {
      "epoch": 1.64,
      "learning_rate": 2.266016586166044e-05,
      "loss": 1.3277,
      "step": 18000
    },
    {
      "epoch": 1.69,
      "learning_rate": 2.190072602448434e-05,
      "loss": 1.2849,
      "step": 18500
    },
    {
      "epoch": 1.73,
      "learning_rate": 2.1141286187308243e-05,
      "loss": 1.2953,
      "step": 19000
    },
    {
      "epoch": 1.78,
      "learning_rate": 2.0381846350132146e-05,
      "loss": 1.2479,
      "step": 19500
    },
    {
      "epoch": 1.82,
      "learning_rate": 1.9622406512956045e-05,
      "loss": 1.2357,
      "step": 20000
    },
    {
      "epoch": 1.87,
      "learning_rate": 1.8862966675779944e-05,
      "loss": 1.2418,
      "step": 20500
    },
    {
      "epoch": 1.91,
      "learning_rate": 1.8103526838603846e-05,
      "loss": 1.2233,
      "step": 21000
    },
    {
      "epoch": 1.96,
      "learning_rate": 1.734408700142775e-05,
      "loss": 1.1791,
      "step": 21500
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.6584647164251648e-05,
      "loss": 1.1664,
      "step": 22000
    },
    {
      "epoch": 2.05,
      "learning_rate": 1.582520732707555e-05,
      "loss": 0.9787,
      "step": 22500
    },
    {
      "epoch": 2.1,
      "learning_rate": 1.5065767489899452e-05,
      "loss": 0.9935,
      "step": 23000
    },
    {
      "epoch": 2.14,
      "learning_rate": 1.4306327652723353e-05,
      "loss": 0.9884,
      "step": 23500
    },
    {
      "epoch": 2.19,
      "learning_rate": 1.3546887815547252e-05,
      "loss": 0.9676,
      "step": 24000
    },
    {
      "epoch": 2.23,
      "learning_rate": 1.2787447978371153e-05,
      "loss": 0.9653,
      "step": 24500
    },
    {
      "epoch": 2.28,
      "learning_rate": 1.2028008141195055e-05,
      "loss": 0.9492,
      "step": 25000
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.1268568304018956e-05,
      "loss": 0.9345,
      "step": 25500
    },
    {
      "epoch": 2.37,
      "learning_rate": 1.0509128466842857e-05,
      "loss": 0.9182,
      "step": 26000
    },
    {
      "epoch": 2.42,
      "learning_rate": 9.749688629666759e-06,
      "loss": 0.9127,
      "step": 26500
    },
    {
      "epoch": 2.46,
      "learning_rate": 8.990248792490658e-06,
      "loss": 0.9039,
      "step": 27000
    },
    {
      "epoch": 2.51,
      "learning_rate": 8.23080895531456e-06,
      "loss": 0.8962,
      "step": 27500
    },
    {
      "epoch": 2.55,
      "learning_rate": 7.471369118138462e-06,
      "loss": 0.8984,
      "step": 28000
    },
    {
      "epoch": 2.6,
      "learning_rate": 6.711929280962362e-06,
      "loss": 0.8718,
      "step": 28500
    },
    {
      "epoch": 2.64,
      "learning_rate": 5.9524894437862635e-06,
      "loss": 0.8662,
      "step": 29000
    },
    {
      "epoch": 2.69,
      "learning_rate": 5.193049606610165e-06,
      "loss": 0.864,
      "step": 29500
    },
    {
      "epoch": 2.73,
      "learning_rate": 4.433609769434066e-06,
      "loss": 0.8497,
      "step": 30000
    },
    {
      "epoch": 2.78,
      "learning_rate": 3.674169932257967e-06,
      "loss": 0.8242,
      "step": 30500
    },
    {
      "epoch": 2.83,
      "learning_rate": 2.9147300950818677e-06,
      "loss": 0.8302,
      "step": 31000
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.155290257905769e-06,
      "loss": 0.7988,
      "step": 31500
    },
    {
      "epoch": 2.92,
      "learning_rate": 1.3958504207296697e-06,
      "loss": 0.8169,
      "step": 32000
    },
    {
      "epoch": 2.96,
      "learning_rate": 6.364105835535709e-07,
      "loss": 0.8069,
      "step": 32500
    },
    {
      "epoch": 3.0,
      "step": 32919,
      "total_flos": 1.4064216261264e+17,
      "train_loss": 1.4714843301307519,
      "train_runtime": 14279.912,
      "train_samples_per_second": 73.765,
      "train_steps_per_second": 2.305
    }
  ],
  "max_steps": 32919,
  "num_train_epochs": 3,
  "total_flos": 1.4064216261264e+17,
  "trial_name": null,
  "trial_params": null
}