{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 1,
  "global_step": 160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0625,
      "grad_norm": 268.0,
      "learning_rate": 2.5e-05,
      "loss": 2.5702,
      "step": 1
    },
    {
      "epoch": 0.0625,
      "eval_accuracy": 0.508,
      "eval_loss": 2.8505053520202637,
      "eval_runtime": 4.8412,
      "eval_samples_per_second": 51.64,
      "eval_steps_per_second": 1.652,
      "step": 1
    },
    {
      "epoch": 0.125,
      "grad_norm": 227.0,
      "learning_rate": 5e-05,
      "loss": 2.6389,
      "step": 2
    },
    {
      "epoch": 0.125,
      "eval_accuracy": 0.528,
      "eval_loss": 2.2588462829589844,
      "eval_runtime": 4.8967,
      "eval_samples_per_second": 51.054,
      "eval_steps_per_second": 1.634,
      "step": 2
    },
    {
      "epoch": 0.1875,
      "grad_norm": 236.0,
      "learning_rate": 4.968354430379747e-05,
      "loss": 2.3551,
      "step": 3
    },
    {
      "epoch": 0.1875,
      "eval_accuracy": 0.536,
      "eval_loss": 1.4990876913070679,
      "eval_runtime": 4.8929,
      "eval_samples_per_second": 51.095,
      "eval_steps_per_second": 1.635,
      "step": 3
    },
    {
      "epoch": 0.25,
      "grad_norm": 89.0,
      "learning_rate": 4.936708860759494e-05,
      "loss": 1.5552,
      "step": 4
    },
    {
      "epoch": 0.25,
      "eval_accuracy": 0.544,
      "eval_loss": 1.8495231866836548,
      "eval_runtime": 4.8705,
      "eval_samples_per_second": 51.33,
      "eval_steps_per_second": 1.643,
      "step": 4
    },
    {
      "epoch": 0.3125,
      "grad_norm": 130.0,
      "learning_rate": 4.905063291139241e-05,
      "loss": 1.6022,
      "step": 5
    },
    {
      "epoch": 0.3125,
      "eval_accuracy": 0.552,
      "eval_loss": 1.5980929136276245,
      "eval_runtime": 4.8961,
      "eval_samples_per_second": 51.061,
      "eval_steps_per_second": 1.634,
      "step": 5
    },
    {
      "epoch": 0.375,
      "grad_norm": 190.0,
      "learning_rate": 4.8734177215189874e-05,
      "loss": 1.8478,
      "step": 6
    },
    {
      "epoch": 0.375,
      "eval_accuracy": 0.62,
      "eval_loss": 1.179288625717163,
      "eval_runtime": 4.8667,
      "eval_samples_per_second": 51.37,
      "eval_steps_per_second": 1.644,
      "step": 6
    },
    {
      "epoch": 0.4375,
      "grad_norm": 92.5,
      "learning_rate": 4.8417721518987346e-05,
      "loss": 1.1437,
      "step": 7
    },
    {
      "epoch": 0.4375,
      "eval_accuracy": 0.624,
      "eval_loss": 1.0324233770370483,
      "eval_runtime": 4.8984,
      "eval_samples_per_second": 51.037,
      "eval_steps_per_second": 1.633,
      "step": 7
    },
    {
      "epoch": 0.5,
      "grad_norm": 52.25,
      "learning_rate": 4.810126582278481e-05,
      "loss": 0.7386,
      "step": 8
    },
    {
      "epoch": 0.5,
      "eval_accuracy": 0.628,
      "eval_loss": 1.0248664617538452,
      "eval_runtime": 4.8699,
      "eval_samples_per_second": 51.335,
      "eval_steps_per_second": 1.643,
      "step": 8
    },
    {
      "epoch": 0.5625,
      "grad_norm": 69.0,
      "learning_rate": 4.778481012658228e-05,
      "loss": 1.0636,
      "step": 9
    },
    {
      "epoch": 0.5625,
      "eval_accuracy": 0.656,
      "eval_loss": 0.9664216041564941,
      "eval_runtime": 4.8671,
      "eval_samples_per_second": 51.365,
      "eval_steps_per_second": 1.644,
      "step": 9
    },
    {
      "epoch": 0.625,
      "grad_norm": 104.0,
      "learning_rate": 4.7468354430379746e-05,
      "loss": 0.9956,
      "step": 10
    },
    {
      "epoch": 0.625,
      "eval_accuracy": 0.696,
      "eval_loss": 0.8491688966751099,
      "eval_runtime": 4.8707,
      "eval_samples_per_second": 51.328,
      "eval_steps_per_second": 1.642,
      "step": 10
    },
    {
      "epoch": 0.6875,
      "grad_norm": 51.75,
      "learning_rate": 4.715189873417722e-05,
      "loss": 0.7195,
      "step": 11
    },
    {
      "epoch": 0.6875,
      "eval_accuracy": 0.724,
      "eval_loss": 0.7724543809890747,
      "eval_runtime": 4.8634,
      "eval_samples_per_second": 51.405,
      "eval_steps_per_second": 1.645,
      "step": 11
    },
    {
      "epoch": 0.75,
      "grad_norm": 24.625,
      "learning_rate": 4.683544303797468e-05,
      "loss": 0.7704,
      "step": 12
    },
    {
      "epoch": 0.75,
      "eval_accuracy": 0.728,
      "eval_loss": 0.7519556283950806,
      "eval_runtime": 4.8881,
      "eval_samples_per_second": 51.144,
      "eval_steps_per_second": 1.637,
      "step": 12
    },
    {
      "epoch": 0.8125,
      "grad_norm": 18.625,
      "learning_rate": 4.6518987341772154e-05,
      "loss": 0.4636,
      "step": 13
    },
    {
      "epoch": 0.8125,
      "eval_accuracy": 0.744,
      "eval_loss": 0.7026970386505127,
      "eval_runtime": 4.8695,
      "eval_samples_per_second": 51.34,
      "eval_steps_per_second": 1.643,
      "step": 13
    },
    {
      "epoch": 0.875,
      "grad_norm": 22.25,
      "learning_rate": 4.6202531645569625e-05,
      "loss": 0.771,
      "step": 14
    },
    {
      "epoch": 0.875,
      "eval_accuracy": 0.788,
      "eval_loss": 0.6045461893081665,
      "eval_runtime": 4.8902,
      "eval_samples_per_second": 51.123,
      "eval_steps_per_second": 1.636,
      "step": 14
    },
    {
      "epoch": 0.9375,
      "grad_norm": 15.5,
      "learning_rate": 4.588607594936709e-05,
      "loss": 0.7138,
      "step": 15
    },
    {
      "epoch": 0.9375,
      "eval_accuracy": 0.82,
      "eval_loss": 0.5602908730506897,
      "eval_runtime": 4.8902,
      "eval_samples_per_second": 51.123,
      "eval_steps_per_second": 1.636,
      "step": 15
    },
    {
      "epoch": 1.0,
      "grad_norm": 30.875,
      "learning_rate": 4.556962025316456e-05,
      "loss": 0.7263,
      "step": 16
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.804,
      "eval_loss": 0.5666066408157349,
      "eval_runtime": 4.8404,
      "eval_samples_per_second": 51.649,
      "eval_steps_per_second": 1.653,
      "step": 16
    },
    {
      "epoch": 1.0625,
      "grad_norm": 77.5,
      "learning_rate": 4.525316455696203e-05,
      "loss": 0.3919,
      "step": 17
    },
    {
      "epoch": 1.0625,
      "eval_accuracy": 0.808,
      "eval_loss": 0.5534331202507019,
      "eval_runtime": 4.816,
      "eval_samples_per_second": 51.911,
      "eval_steps_per_second": 1.661,
      "step": 17
    },
    {
      "epoch": 1.125,
      "grad_norm": 44.5,
      "learning_rate": 4.49367088607595e-05,
      "loss": 0.2448,
      "step": 18
    },
    {
      "epoch": 1.125,
      "eval_accuracy": 0.836,
      "eval_loss": 0.5073856711387634,
      "eval_runtime": 4.8889,
      "eval_samples_per_second": 51.137,
      "eval_steps_per_second": 1.636,
      "step": 18
    },
    {
      "epoch": 1.1875,
      "grad_norm": 31.75,
      "learning_rate": 4.462025316455696e-05,
      "loss": 0.2112,
      "step": 19
    },
    {
      "epoch": 1.1875,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4510812759399414,
      "eval_runtime": 4.8632,
      "eval_samples_per_second": 51.407,
      "eval_steps_per_second": 1.645,
      "step": 19
    },
    {
      "epoch": 1.25,
      "grad_norm": 33.25,
      "learning_rate": 4.430379746835443e-05,
      "loss": 0.335,
      "step": 20
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.828,
      "eval_loss": 0.4451099932193756,
      "eval_runtime": 4.8972,
      "eval_samples_per_second": 51.05,
      "eval_steps_per_second": 1.634,
      "step": 20
    },
    {
      "epoch": 1.3125,
      "grad_norm": 17.375,
      "learning_rate": 4.3987341772151904e-05,
      "loss": 0.196,
      "step": 21
    },
    {
      "epoch": 1.3125,
      "eval_accuracy": 0.82,
      "eval_loss": 0.47517192363739014,
      "eval_runtime": 4.8852,
      "eval_samples_per_second": 51.175,
      "eval_steps_per_second": 1.638,
      "step": 21
    },
    {
      "epoch": 1.375,
      "grad_norm": 45.75,
      "learning_rate": 4.367088607594937e-05,
      "loss": 0.1765,
      "step": 22
    },
    {
      "epoch": 1.375,
      "eval_accuracy": 0.82,
      "eval_loss": 0.4687165319919586,
      "eval_runtime": 4.8161,
      "eval_samples_per_second": 51.909,
      "eval_steps_per_second": 1.661,
      "step": 22
    },
    {
      "epoch": 1.4375,
      "grad_norm": 18.625,
      "learning_rate": 4.3354430379746834e-05,
      "loss": 0.2245,
      "step": 23
    },
    {
      "epoch": 1.4375,
      "eval_accuracy": 0.824,
      "eval_loss": 0.4411630928516388,
      "eval_runtime": 4.8861,
      "eval_samples_per_second": 51.165,
      "eval_steps_per_second": 1.637,
      "step": 23
    },
    {
      "epoch": 1.5,
      "grad_norm": 10.0,
      "learning_rate": 4.3037974683544305e-05,
      "loss": 0.137,
      "step": 24
    },
    {
      "epoch": 1.5,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4266666769981384,
      "eval_runtime": 4.8898,
      "eval_samples_per_second": 51.127,
      "eval_steps_per_second": 1.636,
      "step": 24
    },
    {
      "epoch": 1.5625,
      "grad_norm": 13.125,
      "learning_rate": 4.2721518987341776e-05,
      "loss": 0.1957,
      "step": 25
    },
    {
      "epoch": 1.5625,
      "eval_accuracy": 0.86,
      "eval_loss": 0.42926496267318726,
      "eval_runtime": 4.8436,
      "eval_samples_per_second": 51.614,
      "eval_steps_per_second": 1.652,
      "step": 25
    },
    {
      "epoch": 1.625,
      "grad_norm": 62.0,
      "learning_rate": 4.240506329113924e-05,
      "loss": 0.3881,
      "step": 26
    },
    {
      "epoch": 1.625,
      "eval_accuracy": 0.864,
      "eval_loss": 0.43251192569732666,
      "eval_runtime": 4.8871,
      "eval_samples_per_second": 51.155,
      "eval_steps_per_second": 1.637,
      "step": 26
    },
    {
      "epoch": 1.6875,
      "grad_norm": 7.9375,
      "learning_rate": 4.208860759493671e-05,
      "loss": 0.1298,
      "step": 27
    },
    {
      "epoch": 1.6875,
      "eval_accuracy": 0.856,
      "eval_loss": 0.46340441703796387,
      "eval_runtime": 4.8777,
      "eval_samples_per_second": 51.254,
      "eval_steps_per_second": 1.64,
      "step": 27
    },
    {
      "epoch": 1.75,
      "grad_norm": 13.625,
      "learning_rate": 4.177215189873418e-05,
      "loss": 0.1124,
      "step": 28
    },
    {
      "epoch": 1.75,
      "eval_accuracy": 0.86,
      "eval_loss": 0.46482232213020325,
      "eval_runtime": 4.8682,
      "eval_samples_per_second": 51.353,
      "eval_steps_per_second": 1.643,
      "step": 28
    },
    {
      "epoch": 1.8125,
      "grad_norm": 59.25,
      "learning_rate": 4.145569620253165e-05,
      "loss": 0.2744,
      "step": 29
    },
    {
      "epoch": 1.8125,
      "eval_accuracy": 0.848,
      "eval_loss": 0.43966910243034363,
      "eval_runtime": 4.8941,
      "eval_samples_per_second": 51.082,
      "eval_steps_per_second": 1.635,
      "step": 29
    },
    {
      "epoch": 1.875,
      "grad_norm": 5.28125,
      "learning_rate": 4.113924050632912e-05,
      "loss": 0.0496,
      "step": 30
    },
    {
      "epoch": 1.875,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4173106849193573,
      "eval_runtime": 4.8998,
      "eval_samples_per_second": 51.023,
      "eval_steps_per_second": 1.633,
      "step": 30
    },
    {
      "epoch": 1.9375,
      "grad_norm": 29.625,
      "learning_rate": 4.0822784810126584e-05,
      "loss": 0.1658,
      "step": 31
    },
    {
      "epoch": 1.9375,
      "eval_accuracy": 0.832,
      "eval_loss": 0.41856226325035095,
      "eval_runtime": 4.8693,
      "eval_samples_per_second": 51.343,
      "eval_steps_per_second": 1.643,
      "step": 31
    },
    {
      "epoch": 2.0,
      "grad_norm": 12.625,
      "learning_rate": 4.050632911392405e-05,
      "loss": 0.1718,
      "step": 32
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.836,
      "eval_loss": 0.4255797863006592,
      "eval_runtime": 4.8685,
      "eval_samples_per_second": 51.35,
      "eval_steps_per_second": 1.643,
      "step": 32
    },
    {
      "epoch": 2.0625,
      "grad_norm": 14.125,
      "learning_rate": 4.018987341772152e-05,
      "loss": 0.0979,
      "step": 33
    },
    {
      "epoch": 2.0625,
      "eval_accuracy": 0.836,
      "eval_loss": 0.4206787645816803,
      "eval_runtime": 4.8232,
      "eval_samples_per_second": 51.833,
      "eval_steps_per_second": 1.659,
      "step": 33
    },
    {
      "epoch": 2.125,
      "grad_norm": 4.65625,
      "learning_rate": 3.987341772151899e-05,
      "loss": 0.0284,
      "step": 34
    },
    {
      "epoch": 2.125,
      "eval_accuracy": 0.844,
      "eval_loss": 0.41741058230400085,
      "eval_runtime": 4.8675,
      "eval_samples_per_second": 51.361,
      "eval_steps_per_second": 1.644,
      "step": 34
    },
    {
      "epoch": 2.1875,
      "grad_norm": 8.875,
      "learning_rate": 3.9556962025316456e-05,
      "loss": 0.0558,
      "step": 35
    },
    {
      "epoch": 2.1875,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4207935929298401,
      "eval_runtime": 4.7688,
      "eval_samples_per_second": 52.424,
      "eval_steps_per_second": 1.678,
      "step": 35
    },
    {
      "epoch": 2.25,
      "grad_norm": 4.75,
      "learning_rate": 3.924050632911392e-05,
      "loss": 0.1725,
      "step": 36
    },
    {
      "epoch": 2.25,
      "eval_accuracy": 0.84,
      "eval_loss": 0.43104878067970276,
      "eval_runtime": 4.8634,
      "eval_samples_per_second": 51.404,
      "eval_steps_per_second": 1.645,
      "step": 36
    },
    {
      "epoch": 2.3125,
      "grad_norm": 11.25,
      "learning_rate": 3.89240506329114e-05,
      "loss": 0.0363,
      "step": 37
    },
    {
      "epoch": 2.3125,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4319431185722351,
      "eval_runtime": 4.8918,
      "eval_samples_per_second": 51.106,
      "eval_steps_per_second": 1.635,
      "step": 37
    },
    {
      "epoch": 2.375,
      "grad_norm": 10.125,
      "learning_rate": 3.8607594936708864e-05,
      "loss": 0.0429,
      "step": 38
    },
    {
      "epoch": 2.375,
      "eval_accuracy": 0.856,
      "eval_loss": 0.4509444534778595,
      "eval_runtime": 4.8964,
      "eval_samples_per_second": 51.058,
      "eval_steps_per_second": 1.634,
      "step": 38
    },
    {
      "epoch": 2.4375,
      "grad_norm": 8.3125,
      "learning_rate": 3.829113924050633e-05,
      "loss": 0.0468,
      "step": 39
    },
    {
      "epoch": 2.4375,
      "eval_accuracy": 0.856,
      "eval_loss": 0.4592805802822113,
      "eval_runtime": 4.8908,
      "eval_samples_per_second": 51.117,
      "eval_steps_per_second": 1.636,
      "step": 39
    },
    {
      "epoch": 2.5,
      "grad_norm": 16.5,
      "learning_rate": 3.79746835443038e-05,
      "loss": 0.0525,
      "step": 40
    },
    {
      "epoch": 2.5,
      "eval_accuracy": 0.864,
      "eval_loss": 0.4509994387626648,
      "eval_runtime": 4.8936,
      "eval_samples_per_second": 51.088,
      "eval_steps_per_second": 1.635,
      "step": 40
    },
    {
      "epoch": 2.5625,
      "grad_norm": 3.609375,
      "learning_rate": 3.765822784810127e-05,
      "loss": 0.0232,
      "step": 41
    },
    {
      "epoch": 2.5625,
      "eval_accuracy": 0.852,
      "eval_loss": 0.45242950320243835,
      "eval_runtime": 4.8925,
      "eval_samples_per_second": 51.099,
      "eval_steps_per_second": 1.635,
      "step": 41
    },
    {
      "epoch": 2.625,
      "grad_norm": 1.734375,
      "learning_rate": 3.7341772151898736e-05,
      "loss": 0.0137,
      "step": 42
    },
    {
      "epoch": 2.625,
      "eval_accuracy": 0.852,
      "eval_loss": 0.46332496404647827,
      "eval_runtime": 4.8687,
      "eval_samples_per_second": 51.349,
      "eval_steps_per_second": 1.643,
      "step": 42
    },
    {
      "epoch": 2.6875,
      "grad_norm": 4.75,
      "learning_rate": 3.70253164556962e-05,
      "loss": 0.0202,
      "step": 43
    },
    {
      "epoch": 2.6875,
      "eval_accuracy": 0.86,
      "eval_loss": 0.5035133957862854,
      "eval_runtime": 4.8899,
      "eval_samples_per_second": 51.126,
      "eval_steps_per_second": 1.636,
      "step": 43
    },
    {
      "epoch": 2.75,
      "grad_norm": 15.6875,
      "learning_rate": 3.670886075949367e-05,
      "loss": 0.0383,
      "step": 44
    },
    {
      "epoch": 2.75,
      "eval_accuracy": 0.852,
      "eval_loss": 0.5293290615081787,
      "eval_runtime": 4.8882,
      "eval_samples_per_second": 51.144,
      "eval_steps_per_second": 1.637,
      "step": 44
    },
    {
      "epoch": 2.8125,
      "grad_norm": 24.75,
      "learning_rate": 3.639240506329114e-05,
      "loss": 0.073,
      "step": 45
    },
    {
      "epoch": 2.8125,
      "eval_accuracy": 0.856,
      "eval_loss": 0.5254361033439636,
      "eval_runtime": 4.865,
      "eval_samples_per_second": 51.387,
      "eval_steps_per_second": 1.644,
      "step": 45
    },
    {
      "epoch": 2.875,
      "grad_norm": 28.25,
      "learning_rate": 3.607594936708861e-05,
      "loss": 0.0725,
      "step": 46
    },
    {
      "epoch": 2.875,
      "eval_accuracy": 0.864,
      "eval_loss": 0.5004922151565552,
      "eval_runtime": 4.895,
      "eval_samples_per_second": 51.072,
      "eval_steps_per_second": 1.634,
      "step": 46
    },
    {
      "epoch": 2.9375,
      "grad_norm": 16.375,
      "learning_rate": 3.575949367088608e-05,
      "loss": 0.0516,
      "step": 47
    },
    {
      "epoch": 2.9375,
      "eval_accuracy": 0.864,
      "eval_loss": 0.463413268327713,
      "eval_runtime": 4.8708,
      "eval_samples_per_second": 51.326,
      "eval_steps_per_second": 1.642,
      "step": 47
    },
    {
      "epoch": 3.0,
      "grad_norm": 10.875,
      "learning_rate": 3.5443037974683544e-05,
      "loss": 0.0255,
      "step": 48
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.868,
      "eval_loss": 0.4455827474594116,
      "eval_runtime": 4.8928,
      "eval_samples_per_second": 51.096,
      "eval_steps_per_second": 1.635,
      "step": 48
    },
    {
      "epoch": 3.0625,
      "grad_norm": 7.9375,
      "learning_rate": 3.5126582278481015e-05,
      "loss": 0.0191,
      "step": 49
    },
    {
      "epoch": 3.0625,
      "eval_accuracy": 0.868,
      "eval_loss": 0.45068883895874023,
      "eval_runtime": 4.8384,
      "eval_samples_per_second": 51.67,
      "eval_steps_per_second": 1.653,
      "step": 49
    },
    {
      "epoch": 3.125,
      "grad_norm": 0.23828125,
      "learning_rate": 3.4810126582278487e-05,
      "loss": 0.0021,
      "step": 50
    },
    {
      "epoch": 3.125,
      "eval_accuracy": 0.868,
      "eval_loss": 0.4584948718547821,
      "eval_runtime": 4.838,
      "eval_samples_per_second": 51.675,
      "eval_steps_per_second": 1.654,
      "step": 50
    },
    {
      "epoch": 3.1875,
      "grad_norm": 1.4453125,
      "learning_rate": 3.449367088607595e-05,
      "loss": 0.0034,
      "step": 51
    },
    {
      "epoch": 3.1875,
      "eval_accuracy": 0.864,
      "eval_loss": 0.4626566767692566,
      "eval_runtime": 4.8918,
      "eval_samples_per_second": 51.105,
      "eval_steps_per_second": 1.635,
      "step": 51
    },
    {
      "epoch": 3.25,
      "grad_norm": 1.2890625,
      "learning_rate": 3.4177215189873416e-05,
      "loss": 0.0057,
      "step": 52
    },
    {
      "epoch": 3.25,
      "eval_accuracy": 0.864,
      "eval_loss": 0.46148645877838135,
      "eval_runtime": 4.8885,
      "eval_samples_per_second": 51.141,
      "eval_steps_per_second": 1.637,
      "step": 52
    },
    {
      "epoch": 3.3125,
      "grad_norm": 2.984375,
      "learning_rate": 3.386075949367089e-05,
      "loss": 0.0123,
      "step": 53
    },
    {
      "epoch": 3.3125,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4723862409591675,
      "eval_runtime": 4.8923,
      "eval_samples_per_second": 51.101,
      "eval_steps_per_second": 1.635,
      "step": 53
    },
    {
      "epoch": 3.375,
      "grad_norm": 5.75,
      "learning_rate": 3.354430379746836e-05,
      "loss": 0.0111,
      "step": 54
    },
    {
      "epoch": 3.375,
      "eval_accuracy": 0.856,
      "eval_loss": 0.463609904050827,
      "eval_runtime": 4.8347,
      "eval_samples_per_second": 51.71,
      "eval_steps_per_second": 1.655,
      "step": 54
    },
    {
      "epoch": 3.4375,
      "grad_norm": 4.21875,
      "learning_rate": 3.322784810126582e-05,
      "loss": 0.0072,
      "step": 55
    },
    {
      "epoch": 3.4375,
      "eval_accuracy": 0.88,
      "eval_loss": 0.44373881816864014,
      "eval_runtime": 4.8346,
      "eval_samples_per_second": 51.71,
      "eval_steps_per_second": 1.655,
      "step": 55
    },
    {
      "epoch": 3.5,
      "grad_norm": 2.6875,
      "learning_rate": 3.291139240506329e-05,
      "loss": 0.0044,
      "step": 56
    },
    {
      "epoch": 3.5,
      "eval_accuracy": 0.896,
      "eval_loss": 0.4286611080169678,
      "eval_runtime": 4.8885,
      "eval_samples_per_second": 51.14,
      "eval_steps_per_second": 1.636,
      "step": 56
    },
    {
      "epoch": 3.5625,
      "grad_norm": 0.34375,
      "learning_rate": 3.2594936708860766e-05,
      "loss": 0.0013,
      "step": 57
    },
    {
      "epoch": 3.5625,
      "eval_accuracy": 0.892,
      "eval_loss": 0.4324968755245209,
      "eval_runtime": 4.882,
      "eval_samples_per_second": 51.209,
      "eval_steps_per_second": 1.639,
      "step": 57
    },
    {
      "epoch": 3.625,
      "grad_norm": 0.111328125,
      "learning_rate": 3.227848101265823e-05,
      "loss": 0.0005,
      "step": 58
    },
    {
      "epoch": 3.625,
      "eval_accuracy": 0.896,
      "eval_loss": 0.45419201254844666,
      "eval_runtime": 4.8691,
      "eval_samples_per_second": 51.344,
      "eval_steps_per_second": 1.643,
      "step": 58
    },
    {
      "epoch": 3.6875,
      "grad_norm": 1.3984375,
      "learning_rate": 3.1962025316455695e-05,
      "loss": 0.0045,
      "step": 59
    },
    {
      "epoch": 3.6875,
      "eval_accuracy": 0.892,
      "eval_loss": 0.4625495672225952,
      "eval_runtime": 4.8885,
      "eval_samples_per_second": 51.14,
      "eval_steps_per_second": 1.636,
      "step": 59
    },
    {
      "epoch": 3.75,
      "grad_norm": 0.287109375,
      "learning_rate": 3.1645569620253167e-05,
      "loss": 0.0018,
      "step": 60
    },
    {
      "epoch": 3.75,
      "eval_accuracy": 0.888,
      "eval_loss": 0.4779915511608124,
      "eval_runtime": 4.8965,
      "eval_samples_per_second": 51.057,
      "eval_steps_per_second": 1.634,
      "step": 60
    },
    {
      "epoch": 3.8125,
      "grad_norm": 2.109375,
      "learning_rate": 3.132911392405064e-05,
      "loss": 0.0039,
      "step": 61
    },
    {
      "epoch": 3.8125,
      "eval_accuracy": 0.892,
      "eval_loss": 0.47442150115966797,
      "eval_runtime": 4.8755,
      "eval_samples_per_second": 51.277,
      "eval_steps_per_second": 1.641,
      "step": 61
    },
    {
      "epoch": 3.875,
      "grad_norm": 0.201171875,
      "learning_rate": 3.10126582278481e-05,
      "loss": 0.0007,
      "step": 62
    },
    {
      "epoch": 3.875,
      "eval_accuracy": 0.896,
      "eval_loss": 0.47328320145606995,
      "eval_runtime": 4.8693,
      "eval_samples_per_second": 51.342,
      "eval_steps_per_second": 1.643,
      "step": 62
    },
    {
      "epoch": 3.9375,
      "grad_norm": 0.56640625,
      "learning_rate": 3.0696202531645574e-05,
      "loss": 0.0009,
      "step": 63
    },
    {
      "epoch": 3.9375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.4680858850479126,
      "eval_runtime": 4.8481,
      "eval_samples_per_second": 51.567,
      "eval_steps_per_second": 1.65,
      "step": 63
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.259765625,
      "learning_rate": 3.0379746835443042e-05,
      "loss": 0.0011,
      "step": 64
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9,
      "eval_loss": 0.46620070934295654,
      "eval_runtime": 4.8197,
      "eval_samples_per_second": 51.871,
      "eval_steps_per_second": 1.66,
      "step": 64
    },
    {
      "epoch": 4.0625,
      "grad_norm": 0.01483154296875,
      "learning_rate": 3.0063291139240506e-05,
      "loss": 0.0001,
      "step": 65
    },
    {
      "epoch": 4.0625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.47483450174331665,
      "eval_runtime": 4.8902,
      "eval_samples_per_second": 51.123,
      "eval_steps_per_second": 1.636,
      "step": 65
    },
    {
      "epoch": 4.125,
      "grad_norm": 0.197265625,
      "learning_rate": 2.9746835443037974e-05,
      "loss": 0.0004,
      "step": 66
    },
    {
      "epoch": 4.125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.4847821593284607,
      "eval_runtime": 4.8909,
      "eval_samples_per_second": 51.116,
      "eval_steps_per_second": 1.636,
      "step": 66
    },
    {
      "epoch": 4.1875,
      "grad_norm": 0.283203125,
      "learning_rate": 2.9430379746835446e-05,
      "loss": 0.0008,
      "step": 67
    },
    {
      "epoch": 4.1875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.49961230158805847,
      "eval_runtime": 4.8722,
      "eval_samples_per_second": 51.312,
      "eval_steps_per_second": 1.642,
      "step": 67
    },
    {
      "epoch": 4.25,
      "grad_norm": 0.06103515625,
      "learning_rate": 2.9113924050632914e-05,
      "loss": 0.0001,
      "step": 68
    },
    {
      "epoch": 4.25,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5119830965995789,
      "eval_runtime": 4.872,
      "eval_samples_per_second": 51.314,
      "eval_steps_per_second": 1.642,
      "step": 68
    },
    {
      "epoch": 4.3125,
      "grad_norm": 0.10302734375,
      "learning_rate": 2.879746835443038e-05,
      "loss": 0.0002,
      "step": 69
    },
    {
      "epoch": 4.3125,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5265066623687744,
      "eval_runtime": 4.8698,
      "eval_samples_per_second": 51.337,
      "eval_steps_per_second": 1.643,
      "step": 69
    },
    {
      "epoch": 4.375,
      "grad_norm": 0.1904296875,
      "learning_rate": 2.848101265822785e-05,
      "loss": 0.0003,
      "step": 70
    },
    {
      "epoch": 4.375,
      "eval_accuracy": 0.904,
      "eval_loss": 0.53618323802948,
      "eval_runtime": 4.904,
      "eval_samples_per_second": 50.978,
      "eval_steps_per_second": 1.631,
      "step": 70
    },
    {
      "epoch": 4.4375,
      "grad_norm": 0.31640625,
      "learning_rate": 2.8164556962025318e-05,
      "loss": 0.0005,
      "step": 71
    },
    {
      "epoch": 4.4375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5419101119041443,
      "eval_runtime": 4.8958,
      "eval_samples_per_second": 51.064,
      "eval_steps_per_second": 1.634,
      "step": 71
    },
    {
      "epoch": 4.5,
      "grad_norm": 0.55859375,
      "learning_rate": 2.7848101265822786e-05,
      "loss": 0.001,
      "step": 72
    },
    {
      "epoch": 4.5,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5409899950027466,
      "eval_runtime": 4.8722,
      "eval_samples_per_second": 51.311,
      "eval_steps_per_second": 1.642,
      "step": 72
    },
    {
      "epoch": 4.5625,
      "grad_norm": 0.1376953125,
      "learning_rate": 2.7531645569620257e-05,
      "loss": 0.0002,
      "step": 73
    },
    {
      "epoch": 4.5625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5402071475982666,
      "eval_runtime": 4.8891,
      "eval_samples_per_second": 51.135,
      "eval_steps_per_second": 1.636,
      "step": 73
    },
    {
      "epoch": 4.625,
      "grad_norm": 0.107421875,
      "learning_rate": 2.7215189873417722e-05,
      "loss": 0.0002,
      "step": 74
    },
    {
      "epoch": 4.625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5370295643806458,
      "eval_runtime": 4.8731,
      "eval_samples_per_second": 51.303,
      "eval_steps_per_second": 1.642,
      "step": 74
    },
    {
      "epoch": 4.6875,
      "grad_norm": 0.34765625,
      "learning_rate": 2.689873417721519e-05,
      "loss": 0.0005,
      "step": 75
    },
    {
      "epoch": 4.6875,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5307853817939758,
      "eval_runtime": 4.868,
      "eval_samples_per_second": 51.355,
      "eval_steps_per_second": 1.643,
      "step": 75
    },
    {
      "epoch": 4.75,
      "grad_norm": 0.0732421875,
      "learning_rate": 2.6582278481012658e-05,
      "loss": 0.0001,
      "step": 76
    },
    {
      "epoch": 4.75,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5289453268051147,
      "eval_runtime": 4.8663,
      "eval_samples_per_second": 51.374,
      "eval_steps_per_second": 1.644,
      "step": 76
    },
    {
      "epoch": 4.8125,
      "grad_norm": 0.038818359375,
      "learning_rate": 2.626582278481013e-05,
      "loss": 0.0002,
      "step": 77
    },
    {
      "epoch": 4.8125,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5269708037376404,
      "eval_runtime": 4.9075,
      "eval_samples_per_second": 50.942,
      "eval_steps_per_second": 1.63,
      "step": 77
    },
    {
      "epoch": 4.875,
      "grad_norm": 0.015380859375,
      "learning_rate": 2.5949367088607597e-05,
      "loss": 0.0,
      "step": 78
    },
    {
      "epoch": 4.875,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5278732776641846,
      "eval_runtime": 4.899,
      "eval_samples_per_second": 51.031,
      "eval_steps_per_second": 1.633,
      "step": 78
    },
    {
      "epoch": 4.9375,
      "grad_norm": 0.056884765625,
      "learning_rate": 2.5632911392405062e-05,
      "loss": 0.0001,
      "step": 79
    },
    {
      "epoch": 4.9375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5320713520050049,
      "eval_runtime": 4.824,
      "eval_samples_per_second": 51.825,
      "eval_steps_per_second": 1.658,
      "step": 79
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.01025390625,
      "learning_rate": 2.5316455696202533e-05,
      "loss": 0.0,
      "step": 80
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5385279059410095,
      "eval_runtime": 4.8713,
      "eval_samples_per_second": 51.321,
      "eval_steps_per_second": 1.642,
      "step": 80
    },
    {
      "epoch": 5.0625,
      "grad_norm": 0.03076171875,
      "learning_rate": 2.5e-05,
      "loss": 0.0001,
      "step": 81
    },
    {
      "epoch": 5.0625,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5427132248878479,
      "eval_runtime": 4.8863,
      "eval_samples_per_second": 51.163,
      "eval_steps_per_second": 1.637,
      "step": 81
    },
    {
      "epoch": 5.125,
      "grad_norm": 0.010986328125,
      "learning_rate": 2.468354430379747e-05,
      "loss": 0.0,
      "step": 82
    },
    {
      "epoch": 5.125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5424291491508484,
      "eval_runtime": 4.869,
      "eval_samples_per_second": 51.345,
      "eval_steps_per_second": 1.643,
      "step": 82
    },
    {
      "epoch": 5.1875,
      "grad_norm": 0.00860595703125,
      "learning_rate": 2.4367088607594937e-05,
      "loss": 0.0,
      "step": 83
    },
    {
      "epoch": 5.1875,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5434779524803162,
      "eval_runtime": 4.8938,
      "eval_samples_per_second": 51.085,
      "eval_steps_per_second": 1.635,
      "step": 83
    },
    {
      "epoch": 5.25,
      "grad_norm": 0.046630859375,
      "learning_rate": 2.4050632911392405e-05,
      "loss": 0.0001,
      "step": 84
    },
    {
      "epoch": 5.25,
      "eval_accuracy": 0.9,
      "eval_loss": 0.552955687046051,
      "eval_runtime": 4.8992,
      "eval_samples_per_second": 51.028,
      "eval_steps_per_second": 1.633,
      "step": 84
    },
    {
      "epoch": 5.3125,
      "grad_norm": 0.0242919921875,
      "learning_rate": 2.3734177215189873e-05,
      "loss": 0.0001,
      "step": 85
    },
    {
      "epoch": 5.3125,
      "eval_accuracy": 0.908,
      "eval_loss": 0.5524942874908447,
      "eval_runtime": 4.8893,
      "eval_samples_per_second": 51.132,
      "eval_steps_per_second": 1.636,
      "step": 85
    },
    {
      "epoch": 5.375,
      "grad_norm": 0.10302734375,
      "learning_rate": 2.341772151898734e-05,
      "loss": 0.0002,
      "step": 86
    },
    {
      "epoch": 5.375,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5556187629699707,
      "eval_runtime": 4.8844,
      "eval_samples_per_second": 51.184,
      "eval_steps_per_second": 1.638,
      "step": 86
    },
    {
      "epoch": 5.4375,
      "grad_norm": 0.058837890625,
      "learning_rate": 2.3101265822784813e-05,
      "loss": 0.0001,
      "step": 87
    },
    {
      "epoch": 5.4375,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5552673935890198,
      "eval_runtime": 4.8868,
      "eval_samples_per_second": 51.158,
      "eval_steps_per_second": 1.637,
      "step": 87
    },
    {
      "epoch": 5.5,
      "grad_norm": 0.049560546875,
      "learning_rate": 2.278481012658228e-05,
      "loss": 0.0001,
      "step": 88
    },
    {
      "epoch": 5.5,
      "eval_accuracy": 0.908,
      "eval_loss": 0.5524196624755859,
      "eval_runtime": 4.8881,
      "eval_samples_per_second": 51.145,
      "eval_steps_per_second": 1.637,
      "step": 88
    },
    {
      "epoch": 5.5625,
      "grad_norm": 0.1201171875,
      "learning_rate": 2.246835443037975e-05,
      "loss": 0.0002,
      "step": 89
    },
    {
      "epoch": 5.5625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5546653866767883,
      "eval_runtime": 4.888,
      "eval_samples_per_second": 51.145,
      "eval_steps_per_second": 1.637,
      "step": 89
    },
    {
      "epoch": 5.625,
      "grad_norm": 0.011474609375,
      "learning_rate": 2.2151898734177217e-05,
      "loss": 0.0,
      "step": 90
    },
    {
      "epoch": 5.625,
      "eval_accuracy": 0.908,
      "eval_loss": 0.5581926107406616,
      "eval_runtime": 4.8965,
      "eval_samples_per_second": 51.056,
      "eval_steps_per_second": 1.634,
      "step": 90
    },
    {
      "epoch": 5.6875,
      "grad_norm": 0.00494384765625,
      "learning_rate": 2.1835443037974685e-05,
      "loss": 0.0,
      "step": 91
    },
    {
      "epoch": 5.6875,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5577874779701233,
      "eval_runtime": 4.8978,
      "eval_samples_per_second": 51.043,
      "eval_steps_per_second": 1.633,
      "step": 91
    },
    {
      "epoch": 5.75,
      "grad_norm": 0.001556396484375,
      "learning_rate": 2.1518987341772153e-05,
      "loss": 0.0,
      "step": 92
    },
    {
      "epoch": 5.75,
      "eval_accuracy": 0.908,
      "eval_loss": 0.5564088821411133,
      "eval_runtime": 4.9024,
      "eval_samples_per_second": 50.995,
      "eval_steps_per_second": 1.632,
      "step": 92
    },
    {
      "epoch": 5.8125,
      "grad_norm": 0.04248046875,
      "learning_rate": 2.120253164556962e-05,
      "loss": 0.0001,
      "step": 93
    },
    {
      "epoch": 5.8125,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5549535155296326,
      "eval_runtime": 4.868,
      "eval_samples_per_second": 51.355,
      "eval_steps_per_second": 1.643,
      "step": 93
    },
    {
      "epoch": 5.875,
      "grad_norm": 0.0031890869140625,
      "learning_rate": 2.088607594936709e-05,
      "loss": 0.0,
      "step": 94
    },
    {
      "epoch": 5.875,
      "eval_accuracy": 0.908,
      "eval_loss": 0.5585831999778748,
      "eval_runtime": 4.8902,
      "eval_samples_per_second": 51.123,
      "eval_steps_per_second": 1.636,
      "step": 94
    },
    {
      "epoch": 5.9375,
      "grad_norm": 0.0191650390625,
      "learning_rate": 2.056962025316456e-05,
      "loss": 0.0,
      "step": 95
    },
    {
      "epoch": 5.9375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5561901330947876,
      "eval_runtime": 4.899,
      "eval_samples_per_second": 51.031,
      "eval_steps_per_second": 1.633,
      "step": 95
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.0301513671875,
      "learning_rate": 2.0253164556962025e-05,
      "loss": 0.0001,
      "step": 96
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.908,
      "eval_loss": 0.557356595993042,
      "eval_runtime": 4.8952,
      "eval_samples_per_second": 51.071,
      "eval_steps_per_second": 1.634,
      "step": 96
    },
    {
      "epoch": 6.0625,
      "grad_norm": 0.0029754638671875,
      "learning_rate": 1.9936708860759496e-05,
      "loss": 0.0,
      "step": 97
    },
    {
      "epoch": 6.0625,
      "eval_accuracy": 0.9,
      "eval_loss": 0.557080090045929,
      "eval_runtime": 4.8905,
      "eval_samples_per_second": 51.119,
      "eval_steps_per_second": 1.636,
      "step": 97
    },
    {
      "epoch": 6.125,
      "grad_norm": 0.013427734375,
      "learning_rate": 1.962025316455696e-05,
      "loss": 0.0,
      "step": 98
    },
    {
      "epoch": 6.125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5583979487419128,
      "eval_runtime": 4.8684,
      "eval_samples_per_second": 51.352,
      "eval_steps_per_second": 1.643,
      "step": 98
    },
    {
      "epoch": 6.1875,
      "grad_norm": 0.0205078125,
      "learning_rate": 1.9303797468354432e-05,
      "loss": 0.0,
      "step": 99
    },
    {
      "epoch": 6.1875,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5552465319633484,
      "eval_runtime": 4.8752,
      "eval_samples_per_second": 51.28,
      "eval_steps_per_second": 1.641,
      "step": 99
    },
    {
      "epoch": 6.25,
      "grad_norm": 0.03857421875,
      "learning_rate": 1.89873417721519e-05,
      "loss": 0.0001,
      "step": 100
    },
    {
      "epoch": 6.25,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5585227608680725,
      "eval_runtime": 4.8896,
      "eval_samples_per_second": 51.129,
      "eval_steps_per_second": 1.636,
      "step": 100
    },
    {
      "epoch": 6.3125,
      "grad_norm": 0.047607421875,
      "learning_rate": 1.8670886075949368e-05,
      "loss": 0.0001,
      "step": 101
    },
    {
      "epoch": 6.3125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5532352328300476,
      "eval_runtime": 4.8702,
      "eval_samples_per_second": 51.332,
      "eval_steps_per_second": 1.643,
      "step": 101
    },
    {
      "epoch": 6.375,
      "grad_norm": 0.00836181640625,
      "learning_rate": 1.8354430379746836e-05,
      "loss": 0.0,
      "step": 102
    },
    {
      "epoch": 6.375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.555249035358429,
      "eval_runtime": 4.8964,
      "eval_samples_per_second": 51.058,
      "eval_steps_per_second": 1.634,
      "step": 102
    },
    {
      "epoch": 6.4375,
      "grad_norm": 0.01055908203125,
      "learning_rate": 1.8037974683544304e-05,
      "loss": 0.0,
      "step": 103
    },
    {
      "epoch": 6.4375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.556434690952301,
      "eval_runtime": 4.9004,
      "eval_samples_per_second": 51.016,
      "eval_steps_per_second": 1.633,
      "step": 103
    },
    {
      "epoch": 6.5,
      "grad_norm": 0.022216796875,
      "learning_rate": 1.7721518987341772e-05,
      "loss": 0.0,
      "step": 104
    },
    {
      "epoch": 6.5,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5526236891746521,
      "eval_runtime": 4.87,
      "eval_samples_per_second": 51.335,
      "eval_steps_per_second": 1.643,
      "step": 104
    },
    {
      "epoch": 6.5625,
      "grad_norm": 0.0027618408203125,
      "learning_rate": 1.7405063291139243e-05,
      "loss": 0.0,
      "step": 105
    },
    {
      "epoch": 6.5625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5563039779663086,
      "eval_runtime": 4.892,
      "eval_samples_per_second": 51.104,
      "eval_steps_per_second": 1.635,
      "step": 105
    },
    {
      "epoch": 6.625,
      "grad_norm": 0.007110595703125,
      "learning_rate": 1.7088607594936708e-05,
      "loss": 0.0,
      "step": 106
    },
    {
      "epoch": 6.625,
      "eval_accuracy": 0.892,
      "eval_loss": 0.5540825724601746,
      "eval_runtime": 4.8978,
      "eval_samples_per_second": 51.043,
      "eval_steps_per_second": 1.633,
      "step": 106
    },
    {
      "epoch": 6.6875,
      "grad_norm": 0.0031890869140625,
      "learning_rate": 1.677215189873418e-05,
      "loss": 0.0,
      "step": 107
    },
    {
      "epoch": 6.6875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5545496940612793,
      "eval_runtime": 4.8688,
      "eval_samples_per_second": 51.347,
      "eval_steps_per_second": 1.643,
      "step": 107
    },
    {
      "epoch": 6.75,
      "grad_norm": 0.0223388671875,
      "learning_rate": 1.6455696202531644e-05,
      "loss": 0.0,
      "step": 108
    },
    {
      "epoch": 6.75,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5558124780654907,
      "eval_runtime": 4.8649,
      "eval_samples_per_second": 51.389,
      "eval_steps_per_second": 1.644,
      "step": 108
    },
    {
      "epoch": 6.8125,
      "grad_norm": 0.007354736328125,
      "learning_rate": 1.6139240506329115e-05,
      "loss": 0.0,
      "step": 109
    },
    {
      "epoch": 6.8125,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5552881956100464,
      "eval_runtime": 4.8707,
      "eval_samples_per_second": 51.327,
      "eval_steps_per_second": 1.642,
      "step": 109
    },
    {
      "epoch": 6.875,
      "grad_norm": 0.006011962890625,
      "learning_rate": 1.5822784810126583e-05,
      "loss": 0.0,
      "step": 110
    },
    {
      "epoch": 6.875,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5591185092926025,
      "eval_runtime": 4.8653,
      "eval_samples_per_second": 51.384,
      "eval_steps_per_second": 1.644,
      "step": 110
    },
    {
      "epoch": 6.9375,
      "grad_norm": 0.03759765625,
      "learning_rate": 1.550632911392405e-05,
      "loss": 0.0001,
      "step": 111
    },
    {
      "epoch": 6.9375,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5544008612632751,
      "eval_runtime": 4.8902,
      "eval_samples_per_second": 51.122,
      "eval_steps_per_second": 1.636,
      "step": 111
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.00213623046875,
      "learning_rate": 1.5189873417721521e-05,
      "loss": 0.0,
      "step": 112
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5563123822212219,
      "eval_runtime": 4.8882,
      "eval_samples_per_second": 51.143,
      "eval_steps_per_second": 1.637,
      "step": 112
    },
    {
      "epoch": 7.0625,
      "grad_norm": 0.018798828125,
      "learning_rate": 1.4873417721518987e-05,
      "loss": 0.0,
      "step": 113
    },
    {
      "epoch": 7.0625,
      "eval_accuracy": 0.896,
      "eval_loss": 0.55401611328125,
      "eval_runtime": 4.8635,
      "eval_samples_per_second": 51.403,
      "eval_steps_per_second": 1.645,
      "step": 113
    },
    {
      "epoch": 7.125,
      "grad_norm": 0.0458984375,
      "learning_rate": 1.4556962025316457e-05,
      "loss": 0.0001,
      "step": 114
    },
    {
      "epoch": 7.125,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5561829805374146,
      "eval_runtime": 4.8843,
      "eval_samples_per_second": 51.184,
      "eval_steps_per_second": 1.638,
      "step": 114
    },
    {
      "epoch": 7.1875,
      "grad_norm": 0.01336669921875,
      "learning_rate": 1.4240506329113925e-05,
      "loss": 0.0,
      "step": 115
    },
    {
      "epoch": 7.1875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5537428855895996,
      "eval_runtime": 4.8887,
      "eval_samples_per_second": 51.138,
      "eval_steps_per_second": 1.636,
      "step": 115
    },
    {
      "epoch": 7.25,
      "grad_norm": 0.0113525390625,
      "learning_rate": 1.3924050632911393e-05,
      "loss": 0.0,
      "step": 116
    },
    {
      "epoch": 7.25,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5549358129501343,
      "eval_runtime": 4.8909,
      "eval_samples_per_second": 51.115,
      "eval_steps_per_second": 1.636,
      "step": 116
    },
    {
      "epoch": 7.3125,
      "grad_norm": 0.00762939453125,
      "learning_rate": 1.3607594936708861e-05,
      "loss": 0.0,
      "step": 117
    },
    {
      "epoch": 7.3125,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5518209338188171,
      "eval_runtime": 4.8812,
      "eval_samples_per_second": 51.217,
      "eval_steps_per_second": 1.639,
      "step": 117
    },
    {
      "epoch": 7.375,
      "grad_norm": 0.0038604736328125,
      "learning_rate": 1.3291139240506329e-05,
      "loss": 0.0,
      "step": 118
    },
    {
      "epoch": 7.375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5559548139572144,
      "eval_runtime": 4.889,
      "eval_samples_per_second": 51.135,
      "eval_steps_per_second": 1.636,
      "step": 118
    },
    {
      "epoch": 7.4375,
      "grad_norm": 0.0196533203125,
      "learning_rate": 1.2974683544303799e-05,
      "loss": 0.0,
      "step": 119
    },
    {
      "epoch": 7.4375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5553652048110962,
      "eval_runtime": 4.8764,
      "eval_samples_per_second": 51.268,
      "eval_steps_per_second": 1.641,
      "step": 119
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.00958251953125,
      "learning_rate": 1.2658227848101267e-05,
      "loss": 0.0,
      "step": 120
    },
    {
      "epoch": 7.5,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5554091930389404,
      "eval_runtime": 4.8866,
      "eval_samples_per_second": 51.161,
      "eval_steps_per_second": 1.637,
      "step": 120
    },
    {
      "epoch": 7.5625,
      "grad_norm": 0.0029754638671875,
      "learning_rate": 1.2341772151898735e-05,
      "loss": 0.0,
      "step": 121
    },
    {
      "epoch": 7.5625,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5547875761985779,
      "eval_runtime": 4.8661,
      "eval_samples_per_second": 51.376,
      "eval_steps_per_second": 1.644,
      "step": 121
    },
    {
      "epoch": 7.625,
      "grad_norm": 0.001556396484375,
      "learning_rate": 1.2025316455696203e-05,
      "loss": 0.0,
      "step": 122
    },
    {
      "epoch": 7.625,
      "eval_accuracy": 0.9,
      "eval_loss": 0.55680251121521,
      "eval_runtime": 4.8915,
      "eval_samples_per_second": 51.109,
      "eval_steps_per_second": 1.635,
      "step": 122
    },
    {
      "epoch": 7.6875,
      "grad_norm": 0.0174560546875,
      "learning_rate": 1.170886075949367e-05,
      "loss": 0.0001,
      "step": 123
    },
    {
      "epoch": 7.6875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5555019378662109,
      "eval_runtime": 4.8666,
      "eval_samples_per_second": 51.371,
      "eval_steps_per_second": 1.644,
      "step": 123
    },
    {
      "epoch": 7.75,
      "grad_norm": 0.02734375,
      "learning_rate": 1.139240506329114e-05,
      "loss": 0.0001,
      "step": 124
    },
    {
      "epoch": 7.75,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5568622946739197,
      "eval_runtime": 4.865,
      "eval_samples_per_second": 51.387,
      "eval_steps_per_second": 1.644,
      "step": 124
    },
    {
      "epoch": 7.8125,
      "grad_norm": 0.0242919921875,
      "learning_rate": 1.1075949367088608e-05,
      "loss": 0.0,
      "step": 125
    },
    {
      "epoch": 7.8125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5577536821365356,
      "eval_runtime": 4.8877,
      "eval_samples_per_second": 51.149,
      "eval_steps_per_second": 1.637,
      "step": 125
    },
    {
      "epoch": 7.875,
      "grad_norm": 0.0020904541015625,
      "learning_rate": 1.0759493670886076e-05,
      "loss": 0.0,
      "step": 126
    },
    {
      "epoch": 7.875,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5592789649963379,
      "eval_runtime": 4.8614,
      "eval_samples_per_second": 51.425,
      "eval_steps_per_second": 1.646,
      "step": 126
    },
    {
      "epoch": 7.9375,
      "grad_norm": 0.00732421875,
      "learning_rate": 1.0443037974683544e-05,
      "loss": 0.0,
      "step": 127
    },
    {
      "epoch": 7.9375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.556476891040802,
      "eval_runtime": 4.892,
      "eval_samples_per_second": 51.104,
      "eval_steps_per_second": 1.635,
      "step": 127
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.001861572265625,
      "learning_rate": 1.0126582278481012e-05,
      "loss": 0.0,
      "step": 128
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9,
      "eval_loss": 0.555403470993042,
      "eval_runtime": 4.8901,
      "eval_samples_per_second": 51.123,
      "eval_steps_per_second": 1.636,
      "step": 128
    },
    {
      "epoch": 8.0625,
      "grad_norm": 0.0111083984375,
      "learning_rate": 9.81012658227848e-06,
      "loss": 0.0,
      "step": 129
    },
    {
      "epoch": 8.0625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5534031987190247,
      "eval_runtime": 4.8858,
      "eval_samples_per_second": 51.169,
      "eval_steps_per_second": 1.637,
      "step": 129
    },
    {
      "epoch": 8.125,
      "grad_norm": 0.00604248046875,
      "learning_rate": 9.49367088607595e-06,
      "loss": 0.0,
      "step": 130
    },
    {
      "epoch": 8.125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5557398796081543,
      "eval_runtime": 4.8638,
      "eval_samples_per_second": 51.4,
      "eval_steps_per_second": 1.645,
      "step": 130
    },
    {
      "epoch": 8.1875,
      "grad_norm": 0.008544921875,
      "learning_rate": 9.177215189873418e-06,
      "loss": 0.0,
      "step": 131
    },
    {
      "epoch": 8.1875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5564188361167908,
      "eval_runtime": 4.8832,
      "eval_samples_per_second": 51.196,
      "eval_steps_per_second": 1.638,
      "step": 131
    },
    {
      "epoch": 8.25,
      "grad_norm": 0.003875732421875,
      "learning_rate": 8.860759493670886e-06,
      "loss": 0.0,
      "step": 132
    },
    {
      "epoch": 8.25,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5545617938041687,
      "eval_runtime": 4.8916,
      "eval_samples_per_second": 51.108,
      "eval_steps_per_second": 1.635,
      "step": 132
    },
    {
      "epoch": 8.3125,
      "grad_norm": 0.0322265625,
      "learning_rate": 8.544303797468354e-06,
      "loss": 0.0001,
      "step": 133
    },
    {
      "epoch": 8.3125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5550633668899536,
      "eval_runtime": 4.8868,
      "eval_samples_per_second": 51.158,
      "eval_steps_per_second": 1.637,
      "step": 133
    },
    {
      "epoch": 8.375,
      "grad_norm": 0.0296630859375,
      "learning_rate": 8.227848101265822e-06,
      "loss": 0.0001,
      "step": 134
    },
    {
      "epoch": 8.375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.553980827331543,
      "eval_runtime": 4.8628,
      "eval_samples_per_second": 51.411,
      "eval_steps_per_second": 1.645,
      "step": 134
    },
    {
      "epoch": 8.4375,
      "grad_norm": 0.00396728515625,
      "learning_rate": 7.911392405063292e-06,
      "loss": 0.0,
      "step": 135
    },
    {
      "epoch": 8.4375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5577096939086914,
      "eval_runtime": 4.8664,
      "eval_samples_per_second": 51.372,
      "eval_steps_per_second": 1.644,
      "step": 135
    },
    {
      "epoch": 8.5,
      "grad_norm": 0.00089263916015625,
      "learning_rate": 7.5949367088607605e-06,
      "loss": 0.0,
      "step": 136
    },
    {
      "epoch": 8.5,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5528184771537781,
      "eval_runtime": 4.8633,
      "eval_samples_per_second": 51.405,
      "eval_steps_per_second": 1.645,
      "step": 136
    },
    {
      "epoch": 8.5625,
      "grad_norm": 0.01080322265625,
      "learning_rate": 7.2784810126582285e-06,
      "loss": 0.0,
      "step": 137
    },
    {
      "epoch": 8.5625,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5544995069503784,
      "eval_runtime": 4.8638,
      "eval_samples_per_second": 51.4,
      "eval_steps_per_second": 1.645,
      "step": 137
    },
    {
      "epoch": 8.625,
      "grad_norm": 0.006561279296875,
      "learning_rate": 6.9620253164556965e-06,
      "loss": 0.0,
      "step": 138
    },
    {
      "epoch": 8.625,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5562748908996582,
      "eval_runtime": 4.8677,
      "eval_samples_per_second": 51.359,
      "eval_steps_per_second": 1.643,
      "step": 138
    },
    {
      "epoch": 8.6875,
      "grad_norm": 0.01007080078125,
      "learning_rate": 6.6455696202531645e-06,
      "loss": 0.0,
      "step": 139
    },
    {
      "epoch": 8.6875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5547692775726318,
      "eval_runtime": 4.8897,
      "eval_samples_per_second": 51.128,
      "eval_steps_per_second": 1.636,
      "step": 139
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.01171875,
      "learning_rate": 6.329113924050633e-06,
      "loss": 0.0,
      "step": 140
    },
    {
      "epoch": 8.75,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5561813712120056,
      "eval_runtime": 4.8875,
      "eval_samples_per_second": 51.151,
      "eval_steps_per_second": 1.637,
      "step": 140
    },
    {
      "epoch": 8.8125,
      "grad_norm": 0.0166015625,
      "learning_rate": 6.012658227848101e-06,
      "loss": 0.0,
      "step": 141
    },
    {
      "epoch": 8.8125,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5564705729484558,
      "eval_runtime": 4.8825,
      "eval_samples_per_second": 51.203,
      "eval_steps_per_second": 1.638,
      "step": 141
    },
    {
      "epoch": 8.875,
      "grad_norm": 0.002777099609375,
      "learning_rate": 5.69620253164557e-06,
      "loss": 0.0,
      "step": 142
    },
    {
      "epoch": 8.875,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5576943159103394,
      "eval_runtime": 4.8931,
      "eval_samples_per_second": 51.092,
      "eval_steps_per_second": 1.635,
      "step": 142
    },
    {
      "epoch": 8.9375,
      "grad_norm": 0.048583984375,
      "learning_rate": 5.379746835443038e-06,
      "loss": 0.0001,
      "step": 143
    },
    {
      "epoch": 8.9375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5525559186935425,
      "eval_runtime": 4.8612,
      "eval_samples_per_second": 51.427,
      "eval_steps_per_second": 1.646,
      "step": 143
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.0019683837890625,
      "learning_rate": 5.063291139240506e-06,
      "loss": 0.0,
      "step": 144
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5553069710731506,
      "eval_runtime": 4.8923,
      "eval_samples_per_second": 51.101,
      "eval_steps_per_second": 1.635,
      "step": 144
    },
    {
      "epoch": 9.0625,
      "grad_norm": 0.0279541015625,
      "learning_rate": 4.746835443037975e-06,
      "loss": 0.0,
      "step": 145
    },
    {
      "epoch": 9.0625,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5583046078681946,
      "eval_runtime": 4.8879,
      "eval_samples_per_second": 51.147,
      "eval_steps_per_second": 1.637,
      "step": 145
    },
    {
      "epoch": 9.125,
      "grad_norm": 0.0130615234375,
      "learning_rate": 4.430379746835443e-06,
      "loss": 0.0,
      "step": 146
    },
    {
      "epoch": 9.125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5578604936599731,
      "eval_runtime": 4.8946,
      "eval_samples_per_second": 51.077,
      "eval_steps_per_second": 1.634,
      "step": 146
    },
    {
      "epoch": 9.1875,
      "grad_norm": 0.00732421875,
      "learning_rate": 4.113924050632911e-06,
      "loss": 0.0,
      "step": 147
    },
    {
      "epoch": 9.1875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5572697520256042,
      "eval_runtime": 4.8664,
      "eval_samples_per_second": 51.372,
      "eval_steps_per_second": 1.644,
      "step": 147
    },
    {
      "epoch": 9.25,
      "grad_norm": 0.004364013671875,
      "learning_rate": 3.7974683544303802e-06,
      "loss": 0.0,
      "step": 148
    },
    {
      "epoch": 9.25,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5586851835250854,
      "eval_runtime": 4.8653,
      "eval_samples_per_second": 51.384,
      "eval_steps_per_second": 1.644,
      "step": 148
    },
    {
      "epoch": 9.3125,
      "grad_norm": 0.010986328125,
      "learning_rate": 3.4810126582278482e-06,
      "loss": 0.0,
      "step": 149
    },
    {
      "epoch": 9.3125,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5570804476737976,
      "eval_runtime": 4.8876,
      "eval_samples_per_second": 51.15,
      "eval_steps_per_second": 1.637,
      "step": 149
    },
    {
      "epoch": 9.375,
      "grad_norm": 0.0025787353515625,
      "learning_rate": 3.1645569620253167e-06,
      "loss": 0.0,
      "step": 150
    },
    {
      "epoch": 9.375,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5560451745986938,
      "eval_runtime": 4.8867,
      "eval_samples_per_second": 51.159,
      "eval_steps_per_second": 1.637,
      "step": 150
    },
    {
      "epoch": 9.4375,
      "grad_norm": 0.00616455078125,
      "learning_rate": 2.848101265822785e-06,
      "loss": 0.0,
      "step": 151
    },
    {
      "epoch": 9.4375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5584925413131714,
      "eval_runtime": 4.8614,
      "eval_samples_per_second": 51.425,
      "eval_steps_per_second": 1.646,
      "step": 151
    },
    {
      "epoch": 9.5,
      "grad_norm": 0.01312255859375,
      "learning_rate": 2.531645569620253e-06,
      "loss": 0.0,
      "step": 152
    },
    {
      "epoch": 9.5,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5539237260818481,
      "eval_runtime": 4.8841,
      "eval_samples_per_second": 51.186,
      "eval_steps_per_second": 1.638,
      "step": 152
    },
    {
      "epoch": 9.5625,
      "grad_norm": 0.00604248046875,
      "learning_rate": 2.2151898734177215e-06,
      "loss": 0.0,
      "step": 153
    },
    {
      "epoch": 9.5625,
      "eval_accuracy": 0.904,
      "eval_loss": 0.5538834929466248,
      "eval_runtime": 4.8602,
      "eval_samples_per_second": 51.438,
      "eval_steps_per_second": 1.646,
      "step": 153
    },
    {
      "epoch": 9.625,
      "grad_norm": 0.032958984375,
      "learning_rate": 1.8987341772151901e-06,
      "loss": 0.0001,
      "step": 154
    },
    {
      "epoch": 9.625,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5534281134605408,
      "eval_runtime": 4.8627,
      "eval_samples_per_second": 51.412,
      "eval_steps_per_second": 1.645,
      "step": 154
    },
    {
      "epoch": 9.6875,
      "grad_norm": 0.007354736328125,
      "learning_rate": 1.5822784810126583e-06,
      "loss": 0.0,
      "step": 155
    },
    {
      "epoch": 9.6875,
      "eval_accuracy": 0.9,
      "eval_loss": 0.5582576990127563,
      "eval_runtime": 4.8653,
      "eval_samples_per_second": 51.384,
      "eval_steps_per_second": 1.644,
      "step": 155
    },
    {
      "epoch": 9.75,
      "grad_norm": 0.00421142578125,
      "learning_rate": 1.2658227848101265e-06,
      "loss": 0.0,
      "step": 156
    },
    {
      "epoch": 9.75,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5575215816497803,
      "eval_runtime": 4.878,
      "eval_samples_per_second": 51.251,
      "eval_steps_per_second": 1.64,
      "step": 156
    },
    {
      "epoch": 9.8125,
      "grad_norm": 0.00482177734375,
      "learning_rate": 9.493670886075951e-07,
      "loss": 0.0,
      "step": 157
    },
    {
      "epoch": 9.8125,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5540234446525574,
      "eval_runtime": 4.8905,
      "eval_samples_per_second": 51.119,
      "eval_steps_per_second": 1.636,
      "step": 157
    },
    {
      "epoch": 9.875,
      "grad_norm": 0.0147705078125,
      "learning_rate": 6.329113924050633e-07,
      "loss": 0.0,
      "step": 158
    },
    {
      "epoch": 9.875,
      "eval_accuracy": 0.896,
      "eval_loss": 0.55516517162323,
      "eval_runtime": 4.8871,
      "eval_samples_per_second": 51.156,
      "eval_steps_per_second": 1.637,
      "step": 158
    },
    {
      "epoch": 9.9375,
      "grad_norm": 0.0033416748046875,
      "learning_rate": 3.1645569620253163e-07,
      "loss": 0.0,
      "step": 159
    },
    {
      "epoch": 9.9375,
      "eval_accuracy": 0.896,
      "eval_loss": 0.5555429458618164,
      "eval_runtime": 4.8568,
      "eval_samples_per_second": 51.474,
      "eval_steps_per_second": 1.647,
      "step": 159
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.0185546875,
      "learning_rate": 0.0,
      "loss": 0.0,
      "step": 160
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.896,
      "eval_loss": 0.557336688041687,
      "eval_runtime": 4.8633,
      "eval_samples_per_second": 51.405,
      "eval_steps_per_second": 1.645,
      "step": 160
    },
    {
      "epoch": 10.0,
      "step": 160,
      "total_flos": 7.024653098391962e+16,
      "train_loss": 0.1563696864293888,
      "train_runtime": 1332.2795,
      "train_samples_per_second": 7.506,
      "train_steps_per_second": 0.12
    }
  ],
  "logging_steps": 1,
  "max_steps": 160,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7.024653098391962e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}