{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 1,
  "global_step": 90,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1111111111111111,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 0.8046,
      "step": 1
    },
    {
      "epoch": 0.1111111111111111,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9468807578086853,
      "eval_runtime": 1.4964,
      "eval_samples_per_second": 48.117,
      "eval_steps_per_second": 3.341,
      "step": 1
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 0.834,
      "step": 2
    },
    {
      "epoch": 0.2222222222222222,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9468807578086853,
      "eval_runtime": 1.4798,
      "eval_samples_per_second": 48.657,
      "eval_steps_per_second": 3.379,
      "step": 2
    },
    {
      "epoch": 0.3333333333333333,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 0.864,
      "step": 3
    },
    {
      "epoch": 0.3333333333333333,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9468807578086853,
      "eval_runtime": 1.4711,
      "eval_samples_per_second": 48.945,
      "eval_steps_per_second": 3.399,
      "step": 3
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 0.6596,
      "step": 4
    },
    {
      "epoch": 0.4444444444444444,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9468807578086853,
      "eval_runtime": 1.4781,
      "eval_samples_per_second": 48.711,
      "eval_steps_per_second": 3.383,
      "step": 4
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 15.292896270751953,
      "learning_rate": 2.5e-05,
      "loss": 0.9107,
      "step": 5
    },
    {
      "epoch": 0.5555555555555556,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9468807578086853,
      "eval_runtime": 1.4681,
      "eval_samples_per_second": 49.042,
      "eval_steps_per_second": 3.406,
      "step": 5
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": NaN,
      "learning_rate": 2.5e-05,
      "loss": 0.8285,
      "step": 6
    },
    {
      "epoch": 0.6666666666666666,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9468807578086853,
      "eval_runtime": 1.4656,
      "eval_samples_per_second": 49.125,
      "eval_steps_per_second": 3.411,
      "step": 6
    },
    {
      "epoch": 0.7777777777777778,
      "grad_norm": 38.5906867980957,
      "learning_rate": 5e-05,
      "loss": 0.8436,
      "step": 7
    },
    {
      "epoch": 0.7777777777777778,
      "eval_accuracy": 0.5416666666666666,
      "eval_loss": 0.8725653886795044,
      "eval_runtime": 1.466,
      "eval_samples_per_second": 49.113,
      "eval_steps_per_second": 3.411,
      "step": 7
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 11.768951416015625,
      "learning_rate": 4.943181818181818e-05,
      "loss": 0.8397,
      "step": 8
    },
    {
      "epoch": 0.8888888888888888,
      "eval_accuracy": 0.5138888888888888,
      "eval_loss": 0.89170241355896,
      "eval_runtime": 1.4681,
      "eval_samples_per_second": 49.044,
      "eval_steps_per_second": 3.406,
      "step": 8
    },
    {
      "epoch": 1.0,
      "grad_norm": 38.63240432739258,
      "learning_rate": 4.886363636363637e-05,
      "loss": 0.7961,
      "step": 9
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5833333333333334,
      "eval_loss": 0.9097018837928772,
      "eval_runtime": 1.4715,
      "eval_samples_per_second": 48.93,
      "eval_steps_per_second": 3.398,
      "step": 9
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 25.327926635742188,
      "learning_rate": 4.829545454545455e-05,
      "loss": 0.6728,
      "step": 10
    },
    {
      "epoch": 1.1111111111111112,
      "eval_accuracy": 0.5416666666666666,
      "eval_loss": 0.8724534511566162,
      "eval_runtime": 1.4733,
      "eval_samples_per_second": 48.871,
      "eval_steps_per_second": 3.394,
      "step": 10
    },
    {
      "epoch": 1.2222222222222223,
      "grad_norm": 41.665626525878906,
      "learning_rate": 4.772727272727273e-05,
      "loss": 0.8504,
      "step": 11
    },
    {
      "epoch": 1.2222222222222223,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.8109580278396606,
      "eval_runtime": 1.4661,
      "eval_samples_per_second": 49.109,
      "eval_steps_per_second": 3.41,
      "step": 11
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 28.891475677490234,
      "learning_rate": 4.715909090909091e-05,
      "loss": 0.6348,
      "step": 12
    },
    {
      "epoch": 1.3333333333333333,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.7847561240196228,
      "eval_runtime": 1.4756,
      "eval_samples_per_second": 48.794,
      "eval_steps_per_second": 3.388,
      "step": 12
    },
    {
      "epoch": 1.4444444444444444,
      "grad_norm": 15.667322158813477,
      "learning_rate": 4.659090909090909e-05,
      "loss": 0.7075,
      "step": 13
    },
    {
      "epoch": 1.4444444444444444,
      "eval_accuracy": 0.4861111111111111,
      "eval_loss": 0.8036965131759644,
      "eval_runtime": 1.4698,
      "eval_samples_per_second": 48.987,
      "eval_steps_per_second": 3.402,
      "step": 13
    },
    {
      "epoch": 1.5555555555555556,
      "grad_norm": 14.308302879333496,
      "learning_rate": 4.602272727272727e-05,
      "loss": 0.625,
      "step": 14
    },
    {
      "epoch": 1.5555555555555556,
      "eval_accuracy": 0.4444444444444444,
      "eval_loss": 0.844111979007721,
      "eval_runtime": 1.4649,
      "eval_samples_per_second": 49.152,
      "eval_steps_per_second": 3.413,
      "step": 14
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 33.23239517211914,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.7687,
      "step": 15
    },
    {
      "epoch": 1.6666666666666665,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.8427666425704956,
      "eval_runtime": 1.4735,
      "eval_samples_per_second": 48.862,
      "eval_steps_per_second": 3.393,
      "step": 15
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 23.251686096191406,
      "learning_rate": 4.488636363636364e-05,
      "loss": 0.5379,
      "step": 16
    },
    {
      "epoch": 1.7777777777777777,
      "eval_accuracy": 0.5555555555555556,
      "eval_loss": 0.7749015092849731,
      "eval_runtime": 1.4807,
      "eval_samples_per_second": 48.625,
      "eval_steps_per_second": 3.377,
      "step": 16
    },
    {
      "epoch": 1.8888888888888888,
      "grad_norm": 11.318711280822754,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.5477,
      "step": 17
    },
    {
      "epoch": 1.8888888888888888,
      "eval_accuracy": 0.6805555555555556,
      "eval_loss": 0.68642258644104,
      "eval_runtime": 1.4816,
      "eval_samples_per_second": 48.595,
      "eval_steps_per_second": 3.375,
      "step": 17
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.618575096130371,
      "learning_rate": 4.375e-05,
      "loss": 0.54,
      "step": 18
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7083333333333334,
      "eval_loss": 0.6404860019683838,
      "eval_runtime": 1.4695,
      "eval_samples_per_second": 48.996,
      "eval_steps_per_second": 3.402,
      "step": 18
    },
    {
      "epoch": 2.111111111111111,
      "grad_norm": 4.5261077880859375,
      "learning_rate": 4.318181818181819e-05,
      "loss": 0.4569,
      "step": 19
    },
    {
      "epoch": 2.111111111111111,
      "eval_accuracy": 0.6527777777777778,
      "eval_loss": 0.641287088394165,
      "eval_runtime": 1.4917,
      "eval_samples_per_second": 48.268,
      "eval_steps_per_second": 3.352,
      "step": 19
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 16.080116271972656,
      "learning_rate": 4.261363636363637e-05,
      "loss": 0.4211,
      "step": 20
    },
    {
      "epoch": 2.2222222222222223,
      "eval_accuracy": 0.6805555555555556,
      "eval_loss": 0.5904948115348816,
      "eval_runtime": 1.6079,
      "eval_samples_per_second": 44.778,
      "eval_steps_per_second": 3.11,
      "step": 20
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 10.328635215759277,
      "learning_rate": 4.204545454545455e-05,
      "loss": 0.4088,
      "step": 21
    },
    {
      "epoch": 2.3333333333333335,
      "eval_accuracy": 0.6666666666666666,
      "eval_loss": 0.5809131264686584,
      "eval_runtime": 1.4918,
      "eval_samples_per_second": 48.264,
      "eval_steps_per_second": 3.352,
      "step": 21
    },
    {
      "epoch": 2.4444444444444446,
      "grad_norm": 3.862455368041992,
      "learning_rate": 4.1477272727272734e-05,
      "loss": 0.3588,
      "step": 22
    },
    {
      "epoch": 2.4444444444444446,
      "eval_accuracy": 0.6388888888888888,
      "eval_loss": 0.5841391086578369,
      "eval_runtime": 1.4927,
      "eval_samples_per_second": 48.235,
      "eval_steps_per_second": 3.35,
      "step": 22
    },
    {
      "epoch": 2.5555555555555554,
      "grad_norm": 21.388599395751953,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.3224,
      "step": 23
    },
    {
      "epoch": 2.5555555555555554,
      "eval_accuracy": 0.6805555555555556,
      "eval_loss": 0.5694243311882019,
      "eval_runtime": 1.4741,
      "eval_samples_per_second": 48.845,
      "eval_steps_per_second": 3.392,
      "step": 23
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 18.525726318359375,
      "learning_rate": 4.034090909090909e-05,
      "loss": 0.3886,
      "step": 24
    },
    {
      "epoch": 2.6666666666666665,
      "eval_accuracy": 0.6944444444444444,
      "eval_loss": 0.5775553584098816,
      "eval_runtime": 1.4808,
      "eval_samples_per_second": 48.621,
      "eval_steps_per_second": 3.376,
      "step": 24
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 8.507179260253906,
      "learning_rate": 3.9772727272727275e-05,
      "loss": 0.4483,
      "step": 25
    },
    {
      "epoch": 2.7777777777777777,
      "eval_accuracy": 0.7083333333333334,
      "eval_loss": 0.5739086866378784,
      "eval_runtime": 1.5173,
      "eval_samples_per_second": 47.452,
      "eval_steps_per_second": 3.295,
      "step": 25
    },
    {
      "epoch": 2.888888888888889,
      "grad_norm": 10.042552947998047,
      "learning_rate": 3.9204545454545456e-05,
      "loss": 0.2571,
      "step": 26
    },
    {
      "epoch": 2.888888888888889,
      "eval_accuracy": 0.6944444444444444,
      "eval_loss": 0.5965893864631653,
      "eval_runtime": 1.513,
      "eval_samples_per_second": 47.589,
      "eval_steps_per_second": 3.305,
      "step": 26
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.674591541290283,
      "learning_rate": 3.8636363636363636e-05,
      "loss": 0.2023,
      "step": 27
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7222222222222222,
      "eval_loss": 0.6090915203094482,
      "eval_runtime": 1.4987,
      "eval_samples_per_second": 48.041,
      "eval_steps_per_second": 3.336,
      "step": 27
    },
    {
      "epoch": 3.111111111111111,
      "grad_norm": 4.555273532867432,
      "learning_rate": 3.8068181818181816e-05,
      "loss": 0.1835,
      "step": 28
    },
    {
      "epoch": 3.111111111111111,
      "eval_accuracy": 0.7222222222222222,
      "eval_loss": 0.6348134279251099,
      "eval_runtime": 1.5215,
      "eval_samples_per_second": 47.323,
      "eval_steps_per_second": 3.286,
      "step": 28
    },
    {
      "epoch": 3.2222222222222223,
      "grad_norm": 21.892303466796875,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.2978,
      "step": 29
    },
    {
      "epoch": 3.2222222222222223,
      "eval_accuracy": 0.75,
      "eval_loss": 0.6449768543243408,
      "eval_runtime": 1.4942,
      "eval_samples_per_second": 48.187,
      "eval_steps_per_second": 3.346,
      "step": 29
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 3.3101179599761963,
      "learning_rate": 3.6931818181818184e-05,
      "loss": 0.1375,
      "step": 30
    },
    {
      "epoch": 3.3333333333333335,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.6778775453567505,
      "eval_runtime": 1.581,
      "eval_samples_per_second": 45.542,
      "eval_steps_per_second": 3.163,
      "step": 30
    },
    {
      "epoch": 3.4444444444444446,
      "grad_norm": 5.253066062927246,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 0.0773,
      "step": 31
    },
    {
      "epoch": 3.4444444444444446,
      "eval_accuracy": 0.75,
      "eval_loss": 0.6996697783470154,
      "eval_runtime": 1.4905,
      "eval_samples_per_second": 48.307,
      "eval_steps_per_second": 3.355,
      "step": 31
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 6.1245317459106445,
      "learning_rate": 3.579545454545455e-05,
      "loss": 0.1398,
      "step": 32
    },
    {
      "epoch": 3.5555555555555554,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.7371202111244202,
      "eval_runtime": 1.5564,
      "eval_samples_per_second": 46.261,
      "eval_steps_per_second": 3.213,
      "step": 32
    },
    {
      "epoch": 3.6666666666666665,
      "grad_norm": 8.124380111694336,
      "learning_rate": 3.522727272727273e-05,
      "loss": 0.1665,
      "step": 33
    },
    {
      "epoch": 3.6666666666666665,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.7838079929351807,
      "eval_runtime": 1.4902,
      "eval_samples_per_second": 48.315,
      "eval_steps_per_second": 3.355,
      "step": 33
    },
    {
      "epoch": 3.7777777777777777,
      "grad_norm": 5.12769079208374,
      "learning_rate": 3.465909090909091e-05,
      "loss": 0.0566,
      "step": 34
    },
    {
      "epoch": 3.7777777777777777,
      "eval_accuracy": 0.7222222222222222,
      "eval_loss": 0.823827862739563,
      "eval_runtime": 1.4808,
      "eval_samples_per_second": 48.623,
      "eval_steps_per_second": 3.377,
      "step": 34
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 4.054475784301758,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.0416,
      "step": 35
    },
    {
      "epoch": 3.888888888888889,
      "eval_accuracy": 0.7083333333333334,
      "eval_loss": 0.8069868087768555,
      "eval_runtime": 1.4641,
      "eval_samples_per_second": 49.176,
      "eval_steps_per_second": 3.415,
      "step": 35
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.9879519939422607,
      "learning_rate": 3.352272727272727e-05,
      "loss": 0.082,
      "step": 36
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 0.9339464902877808,
      "eval_runtime": 1.475,
      "eval_samples_per_second": 48.815,
      "eval_steps_per_second": 3.39,
      "step": 36
    },
    {
      "epoch": 4.111111111111111,
      "grad_norm": 5.574306488037109,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.0587,
      "step": 37
    },
    {
      "epoch": 4.111111111111111,
      "eval_accuracy": 0.75,
      "eval_loss": 0.898453414440155,
      "eval_runtime": 1.4751,
      "eval_samples_per_second": 48.809,
      "eval_steps_per_second": 3.39,
      "step": 37
    },
    {
      "epoch": 4.222222222222222,
      "grad_norm": 4.23170280456543,
      "learning_rate": 3.238636363636364e-05,
      "loss": 0.1462,
      "step": 38
    },
    {
      "epoch": 4.222222222222222,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.8970211744308472,
      "eval_runtime": 1.496,
      "eval_samples_per_second": 48.127,
      "eval_steps_per_second": 3.342,
      "step": 38
    },
    {
      "epoch": 4.333333333333333,
      "grad_norm": 4.057853698730469,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.0474,
      "step": 39
    },
    {
      "epoch": 4.333333333333333,
      "eval_accuracy": 0.7083333333333334,
      "eval_loss": 0.9456102848052979,
      "eval_runtime": 1.4756,
      "eval_samples_per_second": 48.794,
      "eval_steps_per_second": 3.388,
      "step": 39
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 1.9089412689208984,
      "learning_rate": 3.125e-05,
      "loss": 0.0247,
      "step": 40
    },
    {
      "epoch": 4.444444444444445,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 1.0376629829406738,
      "eval_runtime": 1.4767,
      "eval_samples_per_second": 48.757,
      "eval_steps_per_second": 3.386,
      "step": 40
    },
    {
      "epoch": 4.555555555555555,
      "grad_norm": 8.005538940429688,
      "learning_rate": 3.068181818181818e-05,
      "loss": 0.0492,
      "step": 41
    },
    {
      "epoch": 4.555555555555555,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 1.09256112575531,
      "eval_runtime": 1.4742,
      "eval_samples_per_second": 48.839,
      "eval_steps_per_second": 3.392,
      "step": 41
    },
    {
      "epoch": 4.666666666666667,
      "grad_norm": 7.321335792541504,
      "learning_rate": 3.0113636363636365e-05,
      "loss": 0.0541,
      "step": 42
    },
    {
      "epoch": 4.666666666666667,
      "eval_accuracy": 0.75,
      "eval_loss": 1.060943365097046,
      "eval_runtime": 1.4853,
      "eval_samples_per_second": 48.474,
      "eval_steps_per_second": 3.366,
      "step": 42
    },
    {
      "epoch": 4.777777777777778,
      "grad_norm": 3.932873487472534,
      "learning_rate": 2.954545454545455e-05,
      "loss": 0.0222,
      "step": 43
    },
    {
      "epoch": 4.777777777777778,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 1.047785997390747,
      "eval_runtime": 1.49,
      "eval_samples_per_second": 48.321,
      "eval_steps_per_second": 3.356,
      "step": 43
    },
    {
      "epoch": 4.888888888888889,
      "grad_norm": 0.5964356660842896,
      "learning_rate": 2.8977272727272732e-05,
      "loss": 0.007,
      "step": 44
    },
    {
      "epoch": 4.888888888888889,
      "eval_accuracy": 0.7222222222222222,
      "eval_loss": 0.9296979904174805,
      "eval_runtime": 1.476,
      "eval_samples_per_second": 48.781,
      "eval_steps_per_second": 3.388,
      "step": 44
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.5752363204956055,
      "learning_rate": 2.8409090909090912e-05,
      "loss": 0.0461,
      "step": 45
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8986176252365112,
      "eval_runtime": 1.5427,
      "eval_samples_per_second": 46.672,
      "eval_steps_per_second": 3.241,
      "step": 45
    },
    {
      "epoch": 5.111111111111111,
      "grad_norm": 0.39924749732017517,
      "learning_rate": 2.784090909090909e-05,
      "loss": 0.004,
      "step": 46
    },
    {
      "epoch": 5.111111111111111,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 0.8829209804534912,
      "eval_runtime": 1.5029,
      "eval_samples_per_second": 47.908,
      "eval_steps_per_second": 3.327,
      "step": 46
    },
    {
      "epoch": 5.222222222222222,
      "grad_norm": 1.3180583715438843,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.0085,
      "step": 47
    },
    {
      "epoch": 5.222222222222222,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 0.9565764665603638,
      "eval_runtime": 1.4989,
      "eval_samples_per_second": 48.036,
      "eval_steps_per_second": 3.336,
      "step": 47
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 0.30337390303611755,
      "learning_rate": 2.6704545454545453e-05,
      "loss": 0.0025,
      "step": 48
    },
    {
      "epoch": 5.333333333333333,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 0.9712399244308472,
      "eval_runtime": 1.4897,
      "eval_samples_per_second": 48.33,
      "eval_steps_per_second": 3.356,
      "step": 48
    },
    {
      "epoch": 5.444444444444445,
      "grad_norm": 1.2145888805389404,
      "learning_rate": 2.6136363636363637e-05,
      "loss": 0.0115,
      "step": 49
    },
    {
      "epoch": 5.444444444444445,
      "eval_accuracy": 0.75,
      "eval_loss": 1.0557817220687866,
      "eval_runtime": 1.5821,
      "eval_samples_per_second": 45.51,
      "eval_steps_per_second": 3.16,
      "step": 49
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 0.5874105095863342,
      "learning_rate": 2.5568181818181817e-05,
      "loss": 0.0063,
      "step": 50
    },
    {
      "epoch": 5.555555555555555,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.0096588134765625,
      "eval_runtime": 1.5483,
      "eval_samples_per_second": 46.502,
      "eval_steps_per_second": 3.229,
      "step": 50
    },
    {
      "epoch": 5.666666666666667,
      "grad_norm": 1.8710988759994507,
      "learning_rate": 2.5e-05,
      "loss": 0.0082,
      "step": 51
    },
    {
      "epoch": 5.666666666666667,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 1.1002558469772339,
      "eval_runtime": 1.5485,
      "eval_samples_per_second": 46.496,
      "eval_steps_per_second": 3.229,
      "step": 51
    },
    {
      "epoch": 5.777777777777778,
      "grad_norm": 0.08881735056638718,
      "learning_rate": 2.4431818181818185e-05,
      "loss": 0.0005,
      "step": 52
    },
    {
      "epoch": 5.777777777777778,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 1.111068844795227,
      "eval_runtime": 1.4791,
      "eval_samples_per_second": 48.677,
      "eval_steps_per_second": 3.38,
      "step": 52
    },
    {
      "epoch": 5.888888888888889,
      "grad_norm": 0.055561624467372894,
      "learning_rate": 2.3863636363636365e-05,
      "loss": 0.0004,
      "step": 53
    },
    {
      "epoch": 5.888888888888889,
      "eval_accuracy": 0.75,
      "eval_loss": 1.1046401262283325,
      "eval_runtime": 1.5013,
      "eval_samples_per_second": 47.957,
      "eval_steps_per_second": 3.33,
      "step": 53
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.051147185266017914,
      "learning_rate": 2.3295454545454546e-05,
      "loss": 0.0004,
      "step": 54
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.1201319694519043,
      "eval_runtime": 1.4825,
      "eval_samples_per_second": 48.566,
      "eval_steps_per_second": 3.373,
      "step": 54
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 0.03255375474691391,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.0003,
      "step": 55
    },
    {
      "epoch": 6.111111111111111,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.1359165906906128,
      "eval_runtime": 1.4977,
      "eval_samples_per_second": 48.075,
      "eval_steps_per_second": 3.339,
      "step": 55
    },
    {
      "epoch": 6.222222222222222,
      "grad_norm": 0.0207963977009058,
      "learning_rate": 2.215909090909091e-05,
      "loss": 0.0001,
      "step": 56
    },
    {
      "epoch": 6.222222222222222,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.2502617835998535,
      "eval_runtime": 1.4856,
      "eval_samples_per_second": 48.464,
      "eval_steps_per_second": 3.366,
      "step": 56
    },
    {
      "epoch": 6.333333333333333,
      "grad_norm": 0.14410509169101715,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.0006,
      "step": 57
    },
    {
      "epoch": 6.333333333333333,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.1601958274841309,
      "eval_runtime": 1.4859,
      "eval_samples_per_second": 48.455,
      "eval_steps_per_second": 3.365,
      "step": 57
    },
    {
      "epoch": 6.444444444444445,
      "grad_norm": 0.054281365126371384,
      "learning_rate": 2.1022727272727274e-05,
      "loss": 0.0003,
      "step": 58
    },
    {
      "epoch": 6.444444444444445,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.2148370742797852,
      "eval_runtime": 1.4936,
      "eval_samples_per_second": 48.206,
      "eval_steps_per_second": 3.348,
      "step": 58
    },
    {
      "epoch": 6.555555555555555,
      "grad_norm": 0.02057286538183689,
      "learning_rate": 2.0454545454545457e-05,
      "loss": 0.0002,
      "step": 59
    },
    {
      "epoch": 6.555555555555555,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.3522080183029175,
      "eval_runtime": 1.4763,
      "eval_samples_per_second": 48.769,
      "eval_steps_per_second": 3.387,
      "step": 59
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 0.0014641749439761043,
      "learning_rate": 1.9886363636363638e-05,
      "loss": 0.0,
      "step": 60
    },
    {
      "epoch": 6.666666666666667,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.2812823057174683,
      "eval_runtime": 1.4521,
      "eval_samples_per_second": 49.584,
      "eval_steps_per_second": 3.443,
      "step": 60
    },
    {
      "epoch": 6.777777777777778,
      "grad_norm": 0.018109608441591263,
      "learning_rate": 1.9318181818181818e-05,
      "loss": 0.0001,
      "step": 61
    },
    {
      "epoch": 6.777777777777778,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.253478765487671,
      "eval_runtime": 1.4921,
      "eval_samples_per_second": 48.254,
      "eval_steps_per_second": 3.351,
      "step": 61
    },
    {
      "epoch": 6.888888888888889,
      "grad_norm": 0.05381924659013748,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0002,
      "step": 62
    },
    {
      "epoch": 6.888888888888889,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3584959506988525,
      "eval_runtime": 1.4846,
      "eval_samples_per_second": 48.499,
      "eval_steps_per_second": 3.368,
      "step": 62
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.002476187190040946,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.0,
      "step": 63
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.364111304283142,
      "eval_runtime": 1.4702,
      "eval_samples_per_second": 48.973,
      "eval_steps_per_second": 3.401,
      "step": 63
    },
    {
      "epoch": 7.111111111111111,
      "grad_norm": 0.014680763706564903,
      "learning_rate": 1.7613636363636366e-05,
      "loss": 0.0001,
      "step": 64
    },
    {
      "epoch": 7.111111111111111,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3435033559799194,
      "eval_runtime": 1.4658,
      "eval_samples_per_second": 49.118,
      "eval_steps_per_second": 3.411,
      "step": 64
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 0.010304883122444153,
      "learning_rate": 1.7045454545454546e-05,
      "loss": 0.0,
      "step": 65
    },
    {
      "epoch": 7.222222222222222,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3150172233581543,
      "eval_runtime": 1.4673,
      "eval_samples_per_second": 49.071,
      "eval_steps_per_second": 3.408,
      "step": 65
    },
    {
      "epoch": 7.333333333333333,
      "grad_norm": 0.014999941922724247,
      "learning_rate": 1.6477272727272726e-05,
      "loss": 0.0001,
      "step": 66
    },
    {
      "epoch": 7.333333333333333,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.396562099456787,
      "eval_runtime": 1.4754,
      "eval_samples_per_second": 48.799,
      "eval_steps_per_second": 3.389,
      "step": 66
    },
    {
      "epoch": 7.444444444444445,
      "grad_norm": 0.021909315139055252,
      "learning_rate": 1.590909090909091e-05,
      "loss": 0.0001,
      "step": 67
    },
    {
      "epoch": 7.444444444444445,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3331305980682373,
      "eval_runtime": 1.4793,
      "eval_samples_per_second": 48.672,
      "eval_steps_per_second": 3.38,
      "step": 67
    },
    {
      "epoch": 7.555555555555555,
      "grad_norm": 0.0006957874284125865,
      "learning_rate": 1.534090909090909e-05,
      "loss": 0.0,
      "step": 68
    },
    {
      "epoch": 7.555555555555555,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3853394985198975,
      "eval_runtime": 1.4743,
      "eval_samples_per_second": 48.838,
      "eval_steps_per_second": 3.392,
      "step": 68
    },
    {
      "epoch": 7.666666666666667,
      "grad_norm": 0.016373885795474052,
      "learning_rate": 1.4772727272727274e-05,
      "loss": 0.0001,
      "step": 69
    },
    {
      "epoch": 7.666666666666667,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.342584490776062,
      "eval_runtime": 1.5274,
      "eval_samples_per_second": 47.14,
      "eval_steps_per_second": 3.274,
      "step": 69
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 0.09788516908884048,
      "learning_rate": 1.4204545454545456e-05,
      "loss": 0.0002,
      "step": 70
    },
    {
      "epoch": 7.777777777777778,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.3891209363937378,
      "eval_runtime": 1.4849,
      "eval_samples_per_second": 48.488,
      "eval_steps_per_second": 3.367,
      "step": 70
    },
    {
      "epoch": 7.888888888888889,
      "grad_norm": 0.004993033595383167,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.0,
      "step": 71
    },
    {
      "epoch": 7.888888888888889,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.461037039756775,
      "eval_runtime": 1.489,
      "eval_samples_per_second": 48.354,
      "eval_steps_per_second": 3.358,
      "step": 71
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.00147097313310951,
      "learning_rate": 1.3068181818181819e-05,
      "loss": 0.0,
      "step": 72
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8194444444444444,
      "eval_loss": 1.3761588335037231,
      "eval_runtime": 1.4965,
      "eval_samples_per_second": 48.111,
      "eval_steps_per_second": 3.341,
      "step": 72
    },
    {
      "epoch": 8.11111111111111,
      "grad_norm": 0.011906428262591362,
      "learning_rate": 1.25e-05,
      "loss": 0.0001,
      "step": 73
    },
    {
      "epoch": 8.11111111111111,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.4502863883972168,
      "eval_runtime": 1.5145,
      "eval_samples_per_second": 47.539,
      "eval_steps_per_second": 3.301,
      "step": 73
    },
    {
      "epoch": 8.222222222222221,
      "grad_norm": 0.0011542687425389886,
      "learning_rate": 1.1931818181818183e-05,
      "loss": 0.0,
      "step": 74
    },
    {
      "epoch": 8.222222222222221,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4473508596420288,
      "eval_runtime": 1.4978,
      "eval_samples_per_second": 48.069,
      "eval_steps_per_second": 3.338,
      "step": 74
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 0.0013727850746363401,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 0.0,
      "step": 75
    },
    {
      "epoch": 8.333333333333334,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.4718106985092163,
      "eval_runtime": 1.4975,
      "eval_samples_per_second": 48.08,
      "eval_steps_per_second": 3.339,
      "step": 75
    },
    {
      "epoch": 8.444444444444445,
      "grad_norm": 0.005262918770313263,
      "learning_rate": 1.0795454545454547e-05,
      "loss": 0.0,
      "step": 76
    },
    {
      "epoch": 8.444444444444445,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3964359760284424,
      "eval_runtime": 1.4988,
      "eval_samples_per_second": 48.037,
      "eval_steps_per_second": 3.336,
      "step": 76
    },
    {
      "epoch": 8.555555555555555,
      "grad_norm": 0.011616203002631664,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.0,
      "step": 77
    },
    {
      "epoch": 8.555555555555555,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.365304708480835,
      "eval_runtime": 1.5036,
      "eval_samples_per_second": 47.884,
      "eval_steps_per_second": 3.325,
      "step": 77
    },
    {
      "epoch": 8.666666666666666,
      "grad_norm": 0.0016829015221446753,
      "learning_rate": 9.659090909090909e-06,
      "loss": 0.0,
      "step": 78
    },
    {
      "epoch": 8.666666666666666,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3907890319824219,
      "eval_runtime": 1.5386,
      "eval_samples_per_second": 46.797,
      "eval_steps_per_second": 3.25,
      "step": 78
    },
    {
      "epoch": 8.777777777777779,
      "grad_norm": 0.007697275839745998,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.0,
      "step": 79
    },
    {
      "epoch": 8.777777777777779,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.2962956428527832,
      "eval_runtime": 1.5653,
      "eval_samples_per_second": 45.997,
      "eval_steps_per_second": 3.194,
      "step": 79
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 0.0010299325222149491,
      "learning_rate": 8.522727272727273e-06,
      "loss": 0.0,
      "step": 80
    },
    {
      "epoch": 8.88888888888889,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.383674144744873,
      "eval_runtime": 1.5762,
      "eval_samples_per_second": 45.681,
      "eval_steps_per_second": 3.172,
      "step": 80
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.005102647002786398,
      "learning_rate": 7.954545454545455e-06,
      "loss": 0.0,
      "step": 81
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3699817657470703,
      "eval_runtime": 1.4837,
      "eval_samples_per_second": 48.527,
      "eval_steps_per_second": 3.37,
      "step": 81
    },
    {
      "epoch": 9.11111111111111,
      "grad_norm": 0.0035217173863202333,
      "learning_rate": 7.386363636363637e-06,
      "loss": 0.0,
      "step": 82
    },
    {
      "epoch": 9.11111111111111,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.4067063331604004,
      "eval_runtime": 1.5387,
      "eval_samples_per_second": 46.792,
      "eval_steps_per_second": 3.249,
      "step": 82
    },
    {
      "epoch": 9.222222222222221,
      "grad_norm": 0.0014651613309979439,
      "learning_rate": 6.818181818181818e-06,
      "loss": 0.0,
      "step": 83
    },
    {
      "epoch": 9.222222222222221,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3422017097473145,
      "eval_runtime": 1.4993,
      "eval_samples_per_second": 48.023,
      "eval_steps_per_second": 3.335,
      "step": 83
    },
    {
      "epoch": 9.333333333333334,
      "grad_norm": 0.002838965505361557,
      "learning_rate": 6.25e-06,
      "loss": 0.0,
      "step": 84
    },
    {
      "epoch": 9.333333333333334,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4219313859939575,
      "eval_runtime": 1.5163,
      "eval_samples_per_second": 47.485,
      "eval_steps_per_second": 3.298,
      "step": 84
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 0.010845585726201534,
      "learning_rate": 5.681818181818182e-06,
      "loss": 0.0,
      "step": 85
    },
    {
      "epoch": 9.444444444444445,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.5043329000473022,
      "eval_runtime": 1.5102,
      "eval_samples_per_second": 47.675,
      "eval_steps_per_second": 3.311,
      "step": 85
    },
    {
      "epoch": 9.555555555555555,
      "grad_norm": 0.0013982527889311314,
      "learning_rate": 5.113636363636364e-06,
      "loss": 0.0,
      "step": 86
    },
    {
      "epoch": 9.555555555555555,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4191205501556396,
      "eval_runtime": 1.537,
      "eval_samples_per_second": 46.846,
      "eval_steps_per_second": 3.253,
      "step": 86
    },
    {
      "epoch": 9.666666666666666,
      "grad_norm": 0.0016395855927839875,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.0,
      "step": 87
    },
    {
      "epoch": 9.666666666666666,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.4259694814682007,
      "eval_runtime": 1.4834,
      "eval_samples_per_second": 48.536,
      "eval_steps_per_second": 3.371,
      "step": 87
    },
    {
      "epoch": 9.777777777777779,
      "grad_norm": 0.0018825188744813204,
      "learning_rate": 3.9772727272727275e-06,
      "loss": 0.0,
      "step": 88
    },
    {
      "epoch": 9.777777777777779,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.4664974212646484,
      "eval_runtime": 1.4871,
      "eval_samples_per_second": 48.416,
      "eval_steps_per_second": 3.362,
      "step": 88
    },
    {
      "epoch": 9.88888888888889,
      "grad_norm": 0.00835789367556572,
      "learning_rate": 3.409090909090909e-06,
      "loss": 0.0,
      "step": 89
    },
    {
      "epoch": 9.88888888888889,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.4490399360656738,
      "eval_runtime": 1.5036,
      "eval_samples_per_second": 47.886,
      "eval_steps_per_second": 3.325,
      "step": 89
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.001027365098707378,
      "learning_rate": 2.840909090909091e-06,
      "loss": 0.0,
      "step": 90
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.3632524013519287,
      "eval_runtime": 1.484,
      "eval_samples_per_second": 48.518,
      "eval_steps_per_second": 3.369,
      "step": 90
    },
    {
      "epoch": 10.0,
      "step": 90,
      "total_flos": 5016736558481408.0,
      "train_loss": 0.20236469821797476,
      "train_runtime": 343.5318,
      "train_samples_per_second": 8.267,
      "train_steps_per_second": 0.262
    }
  ],
  "logging_steps": 1,
  "max_steps": 90,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 5016736558481408.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}