{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.984,
  "eval_steps": 1,
  "global_step": 124,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016,
      "grad_norm": 19.760618209838867,
      "learning_rate": 2.5e-05,
      "loss": 1.0115,
      "step": 1
    },
    {
      "epoch": 0.016,
      "eval_accuracy": 0.388,
      "eval_loss": 1.2842470407485962,
      "eval_runtime": 7.054,
      "eval_samples_per_second": 35.441,
      "eval_steps_per_second": 4.536,
      "step": 1
    },
    {
      "epoch": 0.032,
      "grad_norm": 26.019010543823242,
      "learning_rate": 5e-05,
      "loss": 1.1792,
      "step": 2
    },
    {
      "epoch": 0.032,
      "eval_accuracy": 0.388,
      "eval_loss": 1.2652573585510254,
      "eval_runtime": 7.0551,
      "eval_samples_per_second": 35.435,
      "eval_steps_per_second": 4.536,
      "step": 2
    },
    {
      "epoch": 0.048,
      "grad_norm": 18.010499954223633,
      "learning_rate": 4.959016393442623e-05,
      "loss": 0.9077,
      "step": 3
    },
    {
      "epoch": 0.048,
      "eval_accuracy": 0.396,
      "eval_loss": 1.1932929754257202,
      "eval_runtime": 6.9922,
      "eval_samples_per_second": 35.754,
      "eval_steps_per_second": 4.577,
      "step": 3
    },
    {
      "epoch": 0.064,
      "grad_norm": 35.78538131713867,
      "learning_rate": 4.918032786885246e-05,
      "loss": 1.2982,
      "step": 4
    },
    {
      "epoch": 0.064,
      "eval_accuracy": 0.392,
      "eval_loss": 1.1161845922470093,
      "eval_runtime": 6.9902,
      "eval_samples_per_second": 35.764,
      "eval_steps_per_second": 4.578,
      "step": 4
    },
    {
      "epoch": 0.08,
      "grad_norm": 18.530590057373047,
      "learning_rate": 4.8770491803278687e-05,
      "loss": 1.0213,
      "step": 5
    },
    {
      "epoch": 0.08,
      "eval_accuracy": 0.396,
      "eval_loss": 1.0403300523757935,
      "eval_runtime": 7.052,
      "eval_samples_per_second": 35.451,
      "eval_steps_per_second": 4.538,
      "step": 5
    },
    {
      "epoch": 0.096,
      "grad_norm": 31.981719970703125,
      "learning_rate": 4.836065573770492e-05,
      "loss": 1.0673,
      "step": 6
    },
    {
      "epoch": 0.096,
      "eval_accuracy": 0.412,
      "eval_loss": 0.9646367430686951,
      "eval_runtime": 7.0018,
      "eval_samples_per_second": 35.705,
      "eval_steps_per_second": 4.57,
      "step": 6
    },
    {
      "epoch": 0.112,
      "grad_norm": 17.69211196899414,
      "learning_rate": 4.795081967213115e-05,
      "loss": 0.8041,
      "step": 7
    },
    {
      "epoch": 0.112,
      "eval_accuracy": 0.448,
      "eval_loss": 0.8947343826293945,
      "eval_runtime": 7.0526,
      "eval_samples_per_second": 35.448,
      "eval_steps_per_second": 4.537,
      "step": 7
    },
    {
      "epoch": 0.128,
      "grad_norm": 21.216232299804688,
      "learning_rate": 4.754098360655738e-05,
      "loss": 0.8364,
      "step": 8
    },
    {
      "epoch": 0.128,
      "eval_accuracy": 0.468,
      "eval_loss": 0.8569296598434448,
      "eval_runtime": 7.0532,
      "eval_samples_per_second": 35.445,
      "eval_steps_per_second": 4.537,
      "step": 8
    },
    {
      "epoch": 0.144,
      "grad_norm": 28.313173294067383,
      "learning_rate": 4.713114754098361e-05,
      "loss": 0.9756,
      "step": 9
    },
    {
      "epoch": 0.144,
      "eval_accuracy": 0.468,
      "eval_loss": 0.8217734098434448,
      "eval_runtime": 7.0563,
      "eval_samples_per_second": 35.429,
      "eval_steps_per_second": 4.535,
      "step": 9
    },
    {
      "epoch": 0.16,
      "grad_norm": 20.08592414855957,
      "learning_rate": 4.672131147540984e-05,
      "loss": 0.7215,
      "step": 10
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.476,
      "eval_loss": 0.7888495922088623,
      "eval_runtime": 7.0604,
      "eval_samples_per_second": 35.409,
      "eval_steps_per_second": 4.532,
      "step": 10
    },
    {
      "epoch": 0.176,
      "grad_norm": 22.27912712097168,
      "learning_rate": 4.631147540983607e-05,
      "loss": 0.9865,
      "step": 11
    },
    {
      "epoch": 0.176,
      "eval_accuracy": 0.492,
      "eval_loss": 0.7644336223602295,
      "eval_runtime": 7.0442,
      "eval_samples_per_second": 35.49,
      "eval_steps_per_second": 4.543,
      "step": 11
    },
    {
      "epoch": 0.192,
      "grad_norm": 9.316252708435059,
      "learning_rate": 4.59016393442623e-05,
      "loss": 0.8134,
      "step": 12
    },
    {
      "epoch": 0.192,
      "eval_accuracy": 0.492,
      "eval_loss": 0.7514082193374634,
      "eval_runtime": 7.0626,
      "eval_samples_per_second": 35.398,
      "eval_steps_per_second": 4.531,
      "step": 12
    },
    {
      "epoch": 0.208,
      "grad_norm": 22.068655014038086,
      "learning_rate": 4.549180327868853e-05,
      "loss": 1.0133,
      "step": 13
    },
    {
      "epoch": 0.208,
      "eval_accuracy": 0.504,
      "eval_loss": 0.743643581867218,
      "eval_runtime": 7.0467,
      "eval_samples_per_second": 35.478,
      "eval_steps_per_second": 4.541,
      "step": 13
    },
    {
      "epoch": 0.224,
      "grad_norm": 13.667263984680176,
      "learning_rate": 4.508196721311476e-05,
      "loss": 0.6027,
      "step": 14
    },
    {
      "epoch": 0.224,
      "eval_accuracy": 0.564,
      "eval_loss": 0.7456718683242798,
      "eval_runtime": 7.0525,
      "eval_samples_per_second": 35.448,
      "eval_steps_per_second": 4.537,
      "step": 14
    },
    {
      "epoch": 0.24,
      "grad_norm": 8.404646873474121,
      "learning_rate": 4.467213114754098e-05,
      "loss": 0.6702,
      "step": 15
    },
    {
      "epoch": 0.24,
      "eval_accuracy": 0.632,
      "eval_loss": 0.7482665777206421,
      "eval_runtime": 7.0041,
      "eval_samples_per_second": 35.694,
      "eval_steps_per_second": 4.569,
      "step": 15
    },
    {
      "epoch": 0.256,
      "grad_norm": 10.424461364746094,
      "learning_rate": 4.426229508196721e-05,
      "loss": 0.5514,
      "step": 16
    },
    {
      "epoch": 0.256,
      "eval_accuracy": 0.612,
      "eval_loss": 0.7656826376914978,
      "eval_runtime": 7.0609,
      "eval_samples_per_second": 35.406,
      "eval_steps_per_second": 4.532,
      "step": 16
    },
    {
      "epoch": 0.272,
      "grad_norm": 4.860620498657227,
      "learning_rate": 4.3852459016393444e-05,
      "loss": 0.4876,
      "step": 17
    },
    {
      "epoch": 0.272,
      "eval_accuracy": 0.612,
      "eval_loss": 0.7941005825996399,
      "eval_runtime": 7.0565,
      "eval_samples_per_second": 35.428,
      "eval_steps_per_second": 4.535,
      "step": 17
    },
    {
      "epoch": 0.288,
      "grad_norm": 12.54118824005127,
      "learning_rate": 4.3442622950819674e-05,
      "loss": 0.8596,
      "step": 18
    },
    {
      "epoch": 0.288,
      "eval_accuracy": 0.612,
      "eval_loss": 0.8038740158081055,
      "eval_runtime": 7.0427,
      "eval_samples_per_second": 35.498,
      "eval_steps_per_second": 4.544,
      "step": 18
    },
    {
      "epoch": 0.304,
      "grad_norm": 17.676427841186523,
      "learning_rate": 4.3032786885245904e-05,
      "loss": 0.8936,
      "step": 19
    },
    {
      "epoch": 0.304,
      "eval_accuracy": 0.612,
      "eval_loss": 0.7973695993423462,
      "eval_runtime": 7.0497,
      "eval_samples_per_second": 35.463,
      "eval_steps_per_second": 4.539,
      "step": 19
    },
    {
      "epoch": 0.32,
      "grad_norm": 26.01458740234375,
      "learning_rate": 4.262295081967213e-05,
      "loss": 1.1065,
      "step": 20
    },
    {
      "epoch": 0.32,
      "eval_accuracy": 0.612,
      "eval_loss": 0.7785317301750183,
      "eval_runtime": 7.0561,
      "eval_samples_per_second": 35.43,
      "eval_steps_per_second": 4.535,
      "step": 20
    },
    {
      "epoch": 0.336,
      "grad_norm": 14.344263076782227,
      "learning_rate": 4.2213114754098365e-05,
      "loss": 0.7305,
      "step": 21
    },
    {
      "epoch": 0.336,
      "eval_accuracy": 0.608,
      "eval_loss": 0.7563486099243164,
      "eval_runtime": 7.0512,
      "eval_samples_per_second": 35.455,
      "eval_steps_per_second": 4.538,
      "step": 21
    },
    {
      "epoch": 0.352,
      "grad_norm": 9.17346477508545,
      "learning_rate": 4.1803278688524595e-05,
      "loss": 0.5736,
      "step": 22
    },
    {
      "epoch": 0.352,
      "eval_accuracy": 0.608,
      "eval_loss": 0.7412070035934448,
      "eval_runtime": 7.0471,
      "eval_samples_per_second": 35.476,
      "eval_steps_per_second": 4.541,
      "step": 22
    },
    {
      "epoch": 0.368,
      "grad_norm": 12.143023490905762,
      "learning_rate": 4.1393442622950826e-05,
      "loss": 0.589,
      "step": 23
    },
    {
      "epoch": 0.368,
      "eval_accuracy": 0.608,
      "eval_loss": 0.7287958860397339,
      "eval_runtime": 7.0441,
      "eval_samples_per_second": 35.491,
      "eval_steps_per_second": 4.543,
      "step": 23
    },
    {
      "epoch": 0.384,
      "grad_norm": 18.4329891204834,
      "learning_rate": 4.098360655737705e-05,
      "loss": 0.9443,
      "step": 24
    },
    {
      "epoch": 0.384,
      "eval_accuracy": 0.596,
      "eval_loss": 0.714157223701477,
      "eval_runtime": 6.9925,
      "eval_samples_per_second": 35.753,
      "eval_steps_per_second": 4.576,
      "step": 24
    },
    {
      "epoch": 0.4,
      "grad_norm": 17.655550003051758,
      "learning_rate": 4.057377049180328e-05,
      "loss": 0.6071,
      "step": 25
    },
    {
      "epoch": 0.4,
      "eval_accuracy": 0.592,
      "eval_loss": 0.7043437361717224,
      "eval_runtime": 7.0498,
      "eval_samples_per_second": 35.462,
      "eval_steps_per_second": 4.539,
      "step": 25
    },
    {
      "epoch": 0.416,
      "grad_norm": 5.435880184173584,
      "learning_rate": 4.016393442622951e-05,
      "loss": 0.5661,
      "step": 26
    },
    {
      "epoch": 0.416,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6993535161018372,
      "eval_runtime": 6.9976,
      "eval_samples_per_second": 35.726,
      "eval_steps_per_second": 4.573,
      "step": 26
    },
    {
      "epoch": 0.432,
      "grad_norm": 7.0688395500183105,
      "learning_rate": 3.975409836065574e-05,
      "loss": 0.6571,
      "step": 27
    },
    {
      "epoch": 0.432,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6944531202316284,
      "eval_runtime": 7.0588,
      "eval_samples_per_second": 35.417,
      "eval_steps_per_second": 4.533,
      "step": 27
    },
    {
      "epoch": 0.448,
      "grad_norm": 4.792002201080322,
      "learning_rate": 3.934426229508197e-05,
      "loss": 0.7296,
      "step": 28
    },
    {
      "epoch": 0.448,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6921367049217224,
      "eval_runtime": 6.9708,
      "eval_samples_per_second": 35.864,
      "eval_steps_per_second": 4.591,
      "step": 28
    },
    {
      "epoch": 0.464,
      "grad_norm": 8.166378021240234,
      "learning_rate": 3.89344262295082e-05,
      "loss": 0.6252,
      "step": 29
    },
    {
      "epoch": 0.464,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6932871341705322,
      "eval_runtime": 7.0129,
      "eval_samples_per_second": 35.649,
      "eval_steps_per_second": 4.563,
      "step": 29
    },
    {
      "epoch": 0.48,
      "grad_norm": 13.18557071685791,
      "learning_rate": 3.8524590163934424e-05,
      "loss": 0.7621,
      "step": 30
    },
    {
      "epoch": 0.48,
      "eval_accuracy": 0.564,
      "eval_loss": 0.6983827948570251,
      "eval_runtime": 7.0198,
      "eval_samples_per_second": 35.613,
      "eval_steps_per_second": 4.559,
      "step": 30
    },
    {
      "epoch": 0.496,
      "grad_norm": 5.377532482147217,
      "learning_rate": 3.8114754098360655e-05,
      "loss": 0.6777,
      "step": 31
    },
    {
      "epoch": 0.496,
      "eval_accuracy": 0.564,
      "eval_loss": 0.703988254070282,
      "eval_runtime": 7.0118,
      "eval_samples_per_second": 35.654,
      "eval_steps_per_second": 4.564,
      "step": 31
    },
    {
      "epoch": 0.512,
      "grad_norm": 9.753649711608887,
      "learning_rate": 3.7704918032786885e-05,
      "loss": 0.6802,
      "step": 32
    },
    {
      "epoch": 0.512,
      "eval_accuracy": 0.58,
      "eval_loss": 0.707226574420929,
      "eval_runtime": 7.0122,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.563,
      "step": 32
    },
    {
      "epoch": 0.528,
      "grad_norm": 4.363128662109375,
      "learning_rate": 3.729508196721312e-05,
      "loss": 0.4794,
      "step": 33
    },
    {
      "epoch": 0.528,
      "eval_accuracy": 0.576,
      "eval_loss": 0.7095351815223694,
      "eval_runtime": 6.9626,
      "eval_samples_per_second": 35.906,
      "eval_steps_per_second": 4.596,
      "step": 33
    },
    {
      "epoch": 0.544,
      "grad_norm": 9.72979736328125,
      "learning_rate": 3.6885245901639346e-05,
      "loss": 0.5588,
      "step": 34
    },
    {
      "epoch": 0.544,
      "eval_accuracy": 0.576,
      "eval_loss": 0.709597647190094,
      "eval_runtime": 7.0132,
      "eval_samples_per_second": 35.647,
      "eval_steps_per_second": 4.563,
      "step": 34
    },
    {
      "epoch": 0.56,
      "grad_norm": 8.723620414733887,
      "learning_rate": 3.6475409836065576e-05,
      "loss": 0.7186,
      "step": 35
    },
    {
      "epoch": 0.56,
      "eval_accuracy": 0.572,
      "eval_loss": 0.7086992263793945,
      "eval_runtime": 7.015,
      "eval_samples_per_second": 35.638,
      "eval_steps_per_second": 4.562,
      "step": 35
    },
    {
      "epoch": 0.576,
      "grad_norm": 8.50931453704834,
      "learning_rate": 3.6065573770491806e-05,
      "loss": 0.8171,
      "step": 36
    },
    {
      "epoch": 0.576,
      "eval_accuracy": 0.576,
      "eval_loss": 0.7101796865463257,
      "eval_runtime": 7.0197,
      "eval_samples_per_second": 35.614,
      "eval_steps_per_second": 4.559,
      "step": 36
    },
    {
      "epoch": 0.592,
      "grad_norm": 7.431636333465576,
      "learning_rate": 3.5655737704918037e-05,
      "loss": 0.828,
      "step": 37
    },
    {
      "epoch": 0.592,
      "eval_accuracy": 0.568,
      "eval_loss": 0.7123827934265137,
      "eval_runtime": 7.013,
      "eval_samples_per_second": 35.648,
      "eval_steps_per_second": 4.563,
      "step": 37
    },
    {
      "epoch": 0.608,
      "grad_norm": 14.679248809814453,
      "learning_rate": 3.524590163934427e-05,
      "loss": 0.7014,
      "step": 38
    },
    {
      "epoch": 0.608,
      "eval_accuracy": 0.568,
      "eval_loss": 0.7120000123977661,
      "eval_runtime": 7.0123,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.563,
      "step": 38
    },
    {
      "epoch": 0.624,
      "grad_norm": 22.605974197387695,
      "learning_rate": 3.483606557377049e-05,
      "loss": 0.8848,
      "step": 39
    },
    {
      "epoch": 0.624,
      "eval_accuracy": 0.576,
      "eval_loss": 0.7097578048706055,
      "eval_runtime": 7.0118,
      "eval_samples_per_second": 35.654,
      "eval_steps_per_second": 4.564,
      "step": 39
    },
    {
      "epoch": 0.64,
      "grad_norm": 10.227527618408203,
      "learning_rate": 3.442622950819672e-05,
      "loss": 0.7408,
      "step": 40
    },
    {
      "epoch": 0.64,
      "eval_accuracy": 0.58,
      "eval_loss": 0.7055820226669312,
      "eval_runtime": 7.0124,
      "eval_samples_per_second": 35.651,
      "eval_steps_per_second": 4.563,
      "step": 40
    },
    {
      "epoch": 0.656,
      "grad_norm": 14.50621223449707,
      "learning_rate": 3.401639344262295e-05,
      "loss": 0.741,
      "step": 41
    },
    {
      "epoch": 0.656,
      "eval_accuracy": 0.576,
      "eval_loss": 0.7033867239952087,
      "eval_runtime": 7.0125,
      "eval_samples_per_second": 35.651,
      "eval_steps_per_second": 4.563,
      "step": 41
    },
    {
      "epoch": 0.672,
      "grad_norm": 17.723159790039062,
      "learning_rate": 3.360655737704918e-05,
      "loss": 0.7386,
      "step": 42
    },
    {
      "epoch": 0.672,
      "eval_accuracy": 0.564,
      "eval_loss": 0.7003242373466492,
      "eval_runtime": 7.0141,
      "eval_samples_per_second": 35.643,
      "eval_steps_per_second": 4.562,
      "step": 42
    },
    {
      "epoch": 0.688,
      "grad_norm": 9.643174171447754,
      "learning_rate": 3.319672131147541e-05,
      "loss": 0.806,
      "step": 43
    },
    {
      "epoch": 0.688,
      "eval_accuracy": 0.56,
      "eval_loss": 0.6969140768051147,
      "eval_runtime": 7.0114,
      "eval_samples_per_second": 35.656,
      "eval_steps_per_second": 4.564,
      "step": 43
    },
    {
      "epoch": 0.704,
      "grad_norm": 8.351974487304688,
      "learning_rate": 3.2786885245901635e-05,
      "loss": 0.6752,
      "step": 44
    },
    {
      "epoch": 0.704,
      "eval_accuracy": 0.568,
      "eval_loss": 0.6954609155654907,
      "eval_runtime": 7.0193,
      "eval_samples_per_second": 35.616,
      "eval_steps_per_second": 4.559,
      "step": 44
    },
    {
      "epoch": 0.72,
      "grad_norm": 11.415303230285645,
      "learning_rate": 3.237704918032787e-05,
      "loss": 0.6256,
      "step": 45
    },
    {
      "epoch": 0.72,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6936054825782776,
      "eval_runtime": 6.9694,
      "eval_samples_per_second": 35.871,
      "eval_steps_per_second": 4.592,
      "step": 45
    },
    {
      "epoch": 0.736,
      "grad_norm": 6.098294734954834,
      "learning_rate": 3.19672131147541e-05,
      "loss": 0.6824,
      "step": 46
    },
    {
      "epoch": 0.736,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6936074495315552,
      "eval_runtime": 6.9629,
      "eval_samples_per_second": 35.905,
      "eval_steps_per_second": 4.596,
      "step": 46
    },
    {
      "epoch": 0.752,
      "grad_norm": 10.126526832580566,
      "learning_rate": 3.155737704918033e-05,
      "loss": 0.665,
      "step": 47
    },
    {
      "epoch": 0.752,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6951718926429749,
      "eval_runtime": 6.9629,
      "eval_samples_per_second": 35.905,
      "eval_steps_per_second": 4.596,
      "step": 47
    },
    {
      "epoch": 0.768,
      "grad_norm": 5.928945064544678,
      "learning_rate": 3.114754098360656e-05,
      "loss": 0.7557,
      "step": 48
    },
    {
      "epoch": 0.768,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6969941258430481,
      "eval_runtime": 7.0149,
      "eval_samples_per_second": 35.639,
      "eval_steps_per_second": 4.562,
      "step": 48
    },
    {
      "epoch": 0.784,
      "grad_norm": 10.933755874633789,
      "learning_rate": 3.073770491803279e-05,
      "loss": 0.6899,
      "step": 49
    },
    {
      "epoch": 0.784,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6987070441246033,
      "eval_runtime": 7.0121,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.564,
      "step": 49
    },
    {
      "epoch": 0.8,
      "grad_norm": 12.806472778320312,
      "learning_rate": 3.0327868852459017e-05,
      "loss": 0.7015,
      "step": 50
    },
    {
      "epoch": 0.8,
      "eval_accuracy": 0.596,
      "eval_loss": 0.7015273571014404,
      "eval_runtime": 7.0147,
      "eval_samples_per_second": 35.639,
      "eval_steps_per_second": 4.562,
      "step": 50
    },
    {
      "epoch": 0.816,
      "grad_norm": 6.302614688873291,
      "learning_rate": 2.9918032786885248e-05,
      "loss": 0.6381,
      "step": 51
    },
    {
      "epoch": 0.816,
      "eval_accuracy": 0.588,
      "eval_loss": 0.7052890658378601,
      "eval_runtime": 7.0149,
      "eval_samples_per_second": 35.638,
      "eval_steps_per_second": 4.562,
      "step": 51
    },
    {
      "epoch": 0.832,
      "grad_norm": 16.82048225402832,
      "learning_rate": 2.9508196721311478e-05,
      "loss": 0.7628,
      "step": 52
    },
    {
      "epoch": 0.832,
      "eval_accuracy": 0.584,
      "eval_loss": 0.7072167992591858,
      "eval_runtime": 7.0181,
      "eval_samples_per_second": 35.622,
      "eval_steps_per_second": 4.56,
      "step": 52
    },
    {
      "epoch": 0.848,
      "grad_norm": 13.512633323669434,
      "learning_rate": 2.9098360655737705e-05,
      "loss": 0.6501,
      "step": 53
    },
    {
      "epoch": 0.848,
      "eval_accuracy": 0.588,
      "eval_loss": 0.7064375281333923,
      "eval_runtime": 7.0167,
      "eval_samples_per_second": 35.629,
      "eval_steps_per_second": 4.561,
      "step": 53
    },
    {
      "epoch": 0.864,
      "grad_norm": 13.333196640014648,
      "learning_rate": 2.8688524590163935e-05,
      "loss": 0.6324,
      "step": 54
    },
    {
      "epoch": 0.864,
      "eval_accuracy": 0.592,
      "eval_loss": 0.7091484665870667,
      "eval_runtime": 7.0246,
      "eval_samples_per_second": 35.589,
      "eval_steps_per_second": 4.555,
      "step": 54
    },
    {
      "epoch": 0.88,
      "grad_norm": 18.334705352783203,
      "learning_rate": 2.8278688524590162e-05,
      "loss": 0.8383,
      "step": 55
    },
    {
      "epoch": 0.88,
      "eval_accuracy": 0.592,
      "eval_loss": 0.7081835865974426,
      "eval_runtime": 7.0146,
      "eval_samples_per_second": 35.64,
      "eval_steps_per_second": 4.562,
      "step": 55
    },
    {
      "epoch": 0.896,
      "grad_norm": 18.812061309814453,
      "learning_rate": 2.7868852459016392e-05,
      "loss": 0.8253,
      "step": 56
    },
    {
      "epoch": 0.896,
      "eval_accuracy": 0.592,
      "eval_loss": 0.7055683732032776,
      "eval_runtime": 7.0156,
      "eval_samples_per_second": 35.635,
      "eval_steps_per_second": 4.561,
      "step": 56
    },
    {
      "epoch": 0.912,
      "grad_norm": 17.45163345336914,
      "learning_rate": 2.7459016393442626e-05,
      "loss": 0.8982,
      "step": 57
    },
    {
      "epoch": 0.912,
      "eval_accuracy": 0.592,
      "eval_loss": 0.7019336223602295,
      "eval_runtime": 7.0165,
      "eval_samples_per_second": 35.63,
      "eval_steps_per_second": 4.561,
      "step": 57
    },
    {
      "epoch": 0.928,
      "grad_norm": 9.255583763122559,
      "learning_rate": 2.7049180327868856e-05,
      "loss": 0.8069,
      "step": 58
    },
    {
      "epoch": 0.928,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6979863047599792,
      "eval_runtime": 7.0275,
      "eval_samples_per_second": 35.574,
      "eval_steps_per_second": 4.554,
      "step": 58
    },
    {
      "epoch": 0.944,
      "grad_norm": 6.064142227172852,
      "learning_rate": 2.6639344262295087e-05,
      "loss": 0.4618,
      "step": 59
    },
    {
      "epoch": 0.944,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6955175995826721,
      "eval_runtime": 7.0188,
      "eval_samples_per_second": 35.618,
      "eval_steps_per_second": 4.559,
      "step": 59
    },
    {
      "epoch": 0.96,
      "grad_norm": 9.563672065734863,
      "learning_rate": 2.6229508196721314e-05,
      "loss": 0.7786,
      "step": 60
    },
    {
      "epoch": 0.96,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6935468912124634,
      "eval_runtime": 7.0191,
      "eval_samples_per_second": 35.617,
      "eval_steps_per_second": 4.559,
      "step": 60
    },
    {
      "epoch": 0.976,
      "grad_norm": 4.209216594696045,
      "learning_rate": 2.5819672131147544e-05,
      "loss": 0.5239,
      "step": 61
    },
    {
      "epoch": 0.976,
      "eval_accuracy": 0.604,
      "eval_loss": 0.6912910342216492,
      "eval_runtime": 7.0138,
      "eval_samples_per_second": 35.644,
      "eval_steps_per_second": 4.562,
      "step": 61
    },
    {
      "epoch": 0.992,
      "grad_norm": 4.372480392456055,
      "learning_rate": 2.540983606557377e-05,
      "loss": 0.578,
      "step": 62
    },
    {
      "epoch": 0.992,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6903886795043945,
      "eval_runtime": 7.0123,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.563,
      "step": 62
    },
    {
      "epoch": 1.008,
      "grad_norm": 4.641872882843018,
      "learning_rate": 2.5e-05,
      "loss": 0.6415,
      "step": 63
    },
    {
      "epoch": 1.008,
      "eval_accuracy": 0.6,
      "eval_loss": 0.6886054873466492,
      "eval_runtime": 7.0156,
      "eval_samples_per_second": 35.635,
      "eval_steps_per_second": 4.561,
      "step": 63
    },
    {
      "epoch": 1.024,
      "grad_norm": 9.991393089294434,
      "learning_rate": 2.459016393442623e-05,
      "loss": 0.624,
      "step": 64
    },
    {
      "epoch": 1.024,
      "eval_accuracy": 0.596,
      "eval_loss": 0.6881250143051147,
      "eval_runtime": 7.0155,
      "eval_samples_per_second": 35.636,
      "eval_steps_per_second": 4.561,
      "step": 64
    },
    {
      "epoch": 1.04,
      "grad_norm": 5.3148980140686035,
      "learning_rate": 2.418032786885246e-05,
      "loss": 0.7223,
      "step": 65
    },
    {
      "epoch": 1.04,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6887734532356262,
      "eval_runtime": 7.0118,
      "eval_samples_per_second": 35.654,
      "eval_steps_per_second": 4.564,
      "step": 65
    },
    {
      "epoch": 1.056,
      "grad_norm": 9.664165496826172,
      "learning_rate": 2.377049180327869e-05,
      "loss": 0.8397,
      "step": 66
    },
    {
      "epoch": 1.056,
      "eval_accuracy": 0.592,
      "eval_loss": 0.688064455986023,
      "eval_runtime": 7.0139,
      "eval_samples_per_second": 35.643,
      "eval_steps_per_second": 4.562,
      "step": 66
    },
    {
      "epoch": 1.072,
      "grad_norm": 12.799510955810547,
      "learning_rate": 2.336065573770492e-05,
      "loss": 0.6944,
      "step": 67
    },
    {
      "epoch": 1.072,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6884999871253967,
      "eval_runtime": 7.0162,
      "eval_samples_per_second": 35.632,
      "eval_steps_per_second": 4.561,
      "step": 67
    },
    {
      "epoch": 1.088,
      "grad_norm": 13.912579536437988,
      "learning_rate": 2.295081967213115e-05,
      "loss": 0.8555,
      "step": 68
    },
    {
      "epoch": 1.088,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6879472732543945,
      "eval_runtime": 7.0157,
      "eval_samples_per_second": 35.635,
      "eval_steps_per_second": 4.561,
      "step": 68
    },
    {
      "epoch": 1.104,
      "grad_norm": 5.277914047241211,
      "learning_rate": 2.254098360655738e-05,
      "loss": 0.6251,
      "step": 69
    },
    {
      "epoch": 1.104,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6889882683753967,
      "eval_runtime": 7.0131,
      "eval_samples_per_second": 35.647,
      "eval_steps_per_second": 4.563,
      "step": 69
    },
    {
      "epoch": 1.12,
      "grad_norm": 8.122929573059082,
      "learning_rate": 2.2131147540983607e-05,
      "loss": 0.6682,
      "step": 70
    },
    {
      "epoch": 1.12,
      "eval_accuracy": 0.588,
      "eval_loss": 0.689697265625,
      "eval_runtime": 7.0187,
      "eval_samples_per_second": 35.619,
      "eval_steps_per_second": 4.559,
      "step": 70
    },
    {
      "epoch": 1.1360000000000001,
      "grad_norm": 10.972965240478516,
      "learning_rate": 2.1721311475409837e-05,
      "loss": 0.6869,
      "step": 71
    },
    {
      "epoch": 1.1360000000000001,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6901894807815552,
      "eval_runtime": 7.0124,
      "eval_samples_per_second": 35.651,
      "eval_steps_per_second": 4.563,
      "step": 71
    },
    {
      "epoch": 1.152,
      "grad_norm": 8.138443946838379,
      "learning_rate": 2.1311475409836064e-05,
      "loss": 0.4818,
      "step": 72
    },
    {
      "epoch": 1.152,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6901718974113464,
      "eval_runtime": 7.0104,
      "eval_samples_per_second": 35.661,
      "eval_steps_per_second": 4.565,
      "step": 72
    },
    {
      "epoch": 1.168,
      "grad_norm": 12.203225135803223,
      "learning_rate": 2.0901639344262298e-05,
      "loss": 0.7419,
      "step": 73
    },
    {
      "epoch": 1.168,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6908398270606995,
      "eval_runtime": 7.0172,
      "eval_samples_per_second": 35.627,
      "eval_steps_per_second": 4.56,
      "step": 73
    },
    {
      "epoch": 1.184,
      "grad_norm": 8.825016975402832,
      "learning_rate": 2.0491803278688525e-05,
      "loss": 0.8366,
      "step": 74
    },
    {
      "epoch": 1.184,
      "eval_accuracy": 0.576,
      "eval_loss": 0.689871072769165,
      "eval_runtime": 7.0114,
      "eval_samples_per_second": 35.656,
      "eval_steps_per_second": 4.564,
      "step": 74
    },
    {
      "epoch": 1.2,
      "grad_norm": 8.380965232849121,
      "learning_rate": 2.0081967213114755e-05,
      "loss": 0.728,
      "step": 75
    },
    {
      "epoch": 1.2,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6898242235183716,
      "eval_runtime": 7.0129,
      "eval_samples_per_second": 35.648,
      "eval_steps_per_second": 4.563,
      "step": 75
    },
    {
      "epoch": 1.216,
      "grad_norm": 13.098135948181152,
      "learning_rate": 1.9672131147540985e-05,
      "loss": 0.722,
      "step": 76
    },
    {
      "epoch": 1.216,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6897304654121399,
      "eval_runtime": 7.0172,
      "eval_samples_per_second": 35.627,
      "eval_steps_per_second": 4.56,
      "step": 76
    },
    {
      "epoch": 1.232,
      "grad_norm": 7.279478549957275,
      "learning_rate": 1.9262295081967212e-05,
      "loss": 0.9835,
      "step": 77
    },
    {
      "epoch": 1.232,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6891875267028809,
      "eval_runtime": 7.0202,
      "eval_samples_per_second": 35.612,
      "eval_steps_per_second": 4.558,
      "step": 77
    },
    {
      "epoch": 1.248,
      "grad_norm": 7.378861904144287,
      "learning_rate": 1.8852459016393442e-05,
      "loss": 0.6647,
      "step": 78
    },
    {
      "epoch": 1.248,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6887988448143005,
      "eval_runtime": 7.01,
      "eval_samples_per_second": 35.664,
      "eval_steps_per_second": 4.565,
      "step": 78
    },
    {
      "epoch": 1.264,
      "grad_norm": 5.906654357910156,
      "learning_rate": 1.8442622950819673e-05,
      "loss": 0.7211,
      "step": 79
    },
    {
      "epoch": 1.264,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6894453167915344,
      "eval_runtime": 7.0095,
      "eval_samples_per_second": 35.666,
      "eval_steps_per_second": 4.565,
      "step": 79
    },
    {
      "epoch": 1.28,
      "grad_norm": 7.757331848144531,
      "learning_rate": 1.8032786885245903e-05,
      "loss": 0.7091,
      "step": 80
    },
    {
      "epoch": 1.28,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6894394755363464,
      "eval_runtime": 7.0159,
      "eval_samples_per_second": 35.633,
      "eval_steps_per_second": 4.561,
      "step": 80
    },
    {
      "epoch": 1.296,
      "grad_norm": 19.17230224609375,
      "learning_rate": 1.7622950819672133e-05,
      "loss": 0.843,
      "step": 81
    },
    {
      "epoch": 1.296,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6897109150886536,
      "eval_runtime": 7.016,
      "eval_samples_per_second": 35.633,
      "eval_steps_per_second": 4.561,
      "step": 81
    },
    {
      "epoch": 1.312,
      "grad_norm": 6.863316535949707,
      "learning_rate": 1.721311475409836e-05,
      "loss": 0.8185,
      "step": 82
    },
    {
      "epoch": 1.312,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6896894574165344,
      "eval_runtime": 7.0131,
      "eval_samples_per_second": 35.648,
      "eval_steps_per_second": 4.563,
      "step": 82
    },
    {
      "epoch": 1.328,
      "grad_norm": 6.566874980926514,
      "learning_rate": 1.680327868852459e-05,
      "loss": 0.7932,
      "step": 83
    },
    {
      "epoch": 1.328,
      "eval_accuracy": 0.58,
      "eval_loss": 0.690066397190094,
      "eval_runtime": 7.0194,
      "eval_samples_per_second": 35.616,
      "eval_steps_per_second": 4.559,
      "step": 83
    },
    {
      "epoch": 1.3439999999999999,
      "grad_norm": 7.54443359375,
      "learning_rate": 1.6393442622950818e-05,
      "loss": 0.703,
      "step": 84
    },
    {
      "epoch": 1.3439999999999999,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6897050738334656,
      "eval_runtime": 7.0114,
      "eval_samples_per_second": 35.656,
      "eval_steps_per_second": 4.564,
      "step": 84
    },
    {
      "epoch": 1.3599999999999999,
      "grad_norm": 13.671883583068848,
      "learning_rate": 1.598360655737705e-05,
      "loss": 0.7187,
      "step": 85
    },
    {
      "epoch": 1.3599999999999999,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6892929673194885,
      "eval_runtime": 7.0122,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.563,
      "step": 85
    },
    {
      "epoch": 1.376,
      "grad_norm": 11.296185493469238,
      "learning_rate": 1.557377049180328e-05,
      "loss": 0.8156,
      "step": 86
    },
    {
      "epoch": 1.376,
      "eval_accuracy": 0.58,
      "eval_loss": 0.6897851824760437,
      "eval_runtime": 7.0173,
      "eval_samples_per_second": 35.626,
      "eval_steps_per_second": 4.56,
      "step": 86
    },
    {
      "epoch": 1.392,
      "grad_norm": 12.910079002380371,
      "learning_rate": 1.5163934426229509e-05,
      "loss": 0.7051,
      "step": 87
    },
    {
      "epoch": 1.392,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6891074180603027,
      "eval_runtime": 7.0172,
      "eval_samples_per_second": 35.627,
      "eval_steps_per_second": 4.56,
      "step": 87
    },
    {
      "epoch": 1.408,
      "grad_norm": 19.143896102905273,
      "learning_rate": 1.4754098360655739e-05,
      "loss": 0.5656,
      "step": 88
    },
    {
      "epoch": 1.408,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6875136494636536,
      "eval_runtime": 6.966,
      "eval_samples_per_second": 35.889,
      "eval_steps_per_second": 4.594,
      "step": 88
    },
    {
      "epoch": 1.424,
      "grad_norm": 11.563498497009277,
      "learning_rate": 1.4344262295081968e-05,
      "loss": 0.6649,
      "step": 89
    },
    {
      "epoch": 1.424,
      "eval_accuracy": 0.588,
      "eval_loss": 0.687857449054718,
      "eval_runtime": 7.0159,
      "eval_samples_per_second": 35.633,
      "eval_steps_per_second": 4.561,
      "step": 89
    },
    {
      "epoch": 1.44,
      "grad_norm": 22.282468795776367,
      "learning_rate": 1.3934426229508196e-05,
      "loss": 0.6697,
      "step": 90
    },
    {
      "epoch": 1.44,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6870273351669312,
      "eval_runtime": 7.0118,
      "eval_samples_per_second": 35.654,
      "eval_steps_per_second": 4.564,
      "step": 90
    },
    {
      "epoch": 1.456,
      "grad_norm": 11.898663520812988,
      "learning_rate": 1.3524590163934428e-05,
      "loss": 0.7745,
      "step": 91
    },
    {
      "epoch": 1.456,
      "eval_accuracy": 0.596,
      "eval_loss": 0.6867011785507202,
      "eval_runtime": 7.0125,
      "eval_samples_per_second": 35.651,
      "eval_steps_per_second": 4.563,
      "step": 91
    },
    {
      "epoch": 1.472,
      "grad_norm": 8.677886962890625,
      "learning_rate": 1.3114754098360657e-05,
      "loss": 0.7695,
      "step": 92
    },
    {
      "epoch": 1.472,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6872344017028809,
      "eval_runtime": 7.0145,
      "eval_samples_per_second": 35.64,
      "eval_steps_per_second": 4.562,
      "step": 92
    },
    {
      "epoch": 1.488,
      "grad_norm": 4.995587348937988,
      "learning_rate": 1.2704918032786885e-05,
      "loss": 0.5541,
      "step": 93
    },
    {
      "epoch": 1.488,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6873027086257935,
      "eval_runtime": 7.0143,
      "eval_samples_per_second": 35.641,
      "eval_steps_per_second": 4.562,
      "step": 93
    },
    {
      "epoch": 1.504,
      "grad_norm": 9.183490753173828,
      "learning_rate": 1.2295081967213116e-05,
      "loss": 0.5547,
      "step": 94
    },
    {
      "epoch": 1.504,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6867578029632568,
      "eval_runtime": 7.0101,
      "eval_samples_per_second": 35.663,
      "eval_steps_per_second": 4.565,
      "step": 94
    },
    {
      "epoch": 1.52,
      "grad_norm": 17.053442001342773,
      "learning_rate": 1.1885245901639344e-05,
      "loss": 0.6949,
      "step": 95
    },
    {
      "epoch": 1.52,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6869179606437683,
      "eval_runtime": 7.015,
      "eval_samples_per_second": 35.638,
      "eval_steps_per_second": 4.562,
      "step": 95
    },
    {
      "epoch": 1.536,
      "grad_norm": 11.609779357910156,
      "learning_rate": 1.1475409836065575e-05,
      "loss": 0.7494,
      "step": 96
    },
    {
      "epoch": 1.536,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6875781416893005,
      "eval_runtime": 7.0146,
      "eval_samples_per_second": 35.64,
      "eval_steps_per_second": 4.562,
      "step": 96
    },
    {
      "epoch": 1.552,
      "grad_norm": 7.269466400146484,
      "learning_rate": 1.1065573770491803e-05,
      "loss": 0.7667,
      "step": 97
    },
    {
      "epoch": 1.552,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6860898733139038,
      "eval_runtime": 7.0103,
      "eval_samples_per_second": 35.662,
      "eval_steps_per_second": 4.565,
      "step": 97
    },
    {
      "epoch": 1.568,
      "grad_norm": 7.489114284515381,
      "learning_rate": 1.0655737704918032e-05,
      "loss": 0.6479,
      "step": 98
    },
    {
      "epoch": 1.568,
      "eval_accuracy": 0.592,
      "eval_loss": 0.686941385269165,
      "eval_runtime": 7.0121,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.564,
      "step": 98
    },
    {
      "epoch": 1.584,
      "grad_norm": 12.461353302001953,
      "learning_rate": 1.0245901639344262e-05,
      "loss": 0.5956,
      "step": 99
    },
    {
      "epoch": 1.584,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6870800852775574,
      "eval_runtime": 7.0119,
      "eval_samples_per_second": 35.654,
      "eval_steps_per_second": 4.564,
      "step": 99
    },
    {
      "epoch": 1.6,
      "grad_norm": 10.943467140197754,
      "learning_rate": 9.836065573770493e-06,
      "loss": 0.6405,
      "step": 100
    },
    {
      "epoch": 1.6,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6868281364440918,
      "eval_runtime": 7.0172,
      "eval_samples_per_second": 35.627,
      "eval_steps_per_second": 4.56,
      "step": 100
    },
    {
      "epoch": 1.616,
      "grad_norm": 14.788191795349121,
      "learning_rate": 9.426229508196721e-06,
      "loss": 0.7084,
      "step": 101
    },
    {
      "epoch": 1.616,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6861425638198853,
      "eval_runtime": 7.0188,
      "eval_samples_per_second": 35.619,
      "eval_steps_per_second": 4.559,
      "step": 101
    },
    {
      "epoch": 1.6320000000000001,
      "grad_norm": 8.610825538635254,
      "learning_rate": 9.016393442622952e-06,
      "loss": 0.5918,
      "step": 102
    },
    {
      "epoch": 1.6320000000000001,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6875644326210022,
      "eval_runtime": 7.0178,
      "eval_samples_per_second": 35.624,
      "eval_steps_per_second": 4.56,
      "step": 102
    },
    {
      "epoch": 1.6480000000000001,
      "grad_norm": 8.98721981048584,
      "learning_rate": 8.60655737704918e-06,
      "loss": 0.6359,
      "step": 103
    },
    {
      "epoch": 1.6480000000000001,
      "eval_accuracy": 0.592,
      "eval_loss": 0.687570333480835,
      "eval_runtime": 7.0145,
      "eval_samples_per_second": 35.64,
      "eval_steps_per_second": 4.562,
      "step": 103
    },
    {
      "epoch": 1.6640000000000001,
      "grad_norm": 7.4758992195129395,
      "learning_rate": 8.196721311475409e-06,
      "loss": 0.6349,
      "step": 104
    },
    {
      "epoch": 1.6640000000000001,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6866093873977661,
      "eval_runtime": 7.0157,
      "eval_samples_per_second": 35.635,
      "eval_steps_per_second": 4.561,
      "step": 104
    },
    {
      "epoch": 1.6800000000000002,
      "grad_norm": 5.330641269683838,
      "learning_rate": 7.78688524590164e-06,
      "loss": 0.5346,
      "step": 105
    },
    {
      "epoch": 1.6800000000000002,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6873124837875366,
      "eval_runtime": 7.013,
      "eval_samples_per_second": 35.648,
      "eval_steps_per_second": 4.563,
      "step": 105
    },
    {
      "epoch": 1.696,
      "grad_norm": 7.447720050811768,
      "learning_rate": 7.3770491803278695e-06,
      "loss": 0.7078,
      "step": 106
    },
    {
      "epoch": 1.696,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6875371336936951,
      "eval_runtime": 7.0195,
      "eval_samples_per_second": 35.615,
      "eval_steps_per_second": 4.559,
      "step": 106
    },
    {
      "epoch": 1.712,
      "grad_norm": 6.726109504699707,
      "learning_rate": 6.967213114754098e-06,
      "loss": 0.7211,
      "step": 107
    },
    {
      "epoch": 1.712,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6877539157867432,
      "eval_runtime": 7.0154,
      "eval_samples_per_second": 35.636,
      "eval_steps_per_second": 4.561,
      "step": 107
    },
    {
      "epoch": 1.728,
      "grad_norm": 8.027348518371582,
      "learning_rate": 6.557377049180328e-06,
      "loss": 0.6603,
      "step": 108
    },
    {
      "epoch": 1.728,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6869003772735596,
      "eval_runtime": 7.0216,
      "eval_samples_per_second": 35.605,
      "eval_steps_per_second": 4.557,
      "step": 108
    },
    {
      "epoch": 1.744,
      "grad_norm": 5.765665531158447,
      "learning_rate": 6.147540983606558e-06,
      "loss": 0.6232,
      "step": 109
    },
    {
      "epoch": 1.744,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6870996356010437,
      "eval_runtime": 7.0173,
      "eval_samples_per_second": 35.626,
      "eval_steps_per_second": 4.56,
      "step": 109
    },
    {
      "epoch": 1.76,
      "grad_norm": 7.992295742034912,
      "learning_rate": 5.737704918032787e-06,
      "loss": 0.5974,
      "step": 110
    },
    {
      "epoch": 1.76,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6865136623382568,
      "eval_runtime": 7.015,
      "eval_samples_per_second": 35.638,
      "eval_steps_per_second": 4.562,
      "step": 110
    },
    {
      "epoch": 1.776,
      "grad_norm": 7.537731647491455,
      "learning_rate": 5.327868852459016e-06,
      "loss": 0.7903,
      "step": 111
    },
    {
      "epoch": 1.776,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6870468854904175,
      "eval_runtime": 7.0136,
      "eval_samples_per_second": 35.645,
      "eval_steps_per_second": 4.563,
      "step": 111
    },
    {
      "epoch": 1.792,
      "grad_norm": 10.202421188354492,
      "learning_rate": 4.918032786885246e-06,
      "loss": 0.7646,
      "step": 112
    },
    {
      "epoch": 1.792,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6872812509536743,
      "eval_runtime": 7.014,
      "eval_samples_per_second": 35.643,
      "eval_steps_per_second": 4.562,
      "step": 112
    },
    {
      "epoch": 1.808,
      "grad_norm": 5.8284993171691895,
      "learning_rate": 4.508196721311476e-06,
      "loss": 0.7683,
      "step": 113
    },
    {
      "epoch": 1.808,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6874414086341858,
      "eval_runtime": 7.0182,
      "eval_samples_per_second": 35.622,
      "eval_steps_per_second": 4.56,
      "step": 113
    },
    {
      "epoch": 1.8239999999999998,
      "grad_norm": 14.522705078125,
      "learning_rate": 4.098360655737704e-06,
      "loss": 0.7617,
      "step": 114
    },
    {
      "epoch": 1.8239999999999998,
      "eval_accuracy": 0.588,
      "eval_loss": 0.687423825263977,
      "eval_runtime": 7.0145,
      "eval_samples_per_second": 35.64,
      "eval_steps_per_second": 4.562,
      "step": 114
    },
    {
      "epoch": 1.8399999999999999,
      "grad_norm": 13.447626113891602,
      "learning_rate": 3.6885245901639347e-06,
      "loss": 0.6276,
      "step": 115
    },
    {
      "epoch": 1.8399999999999999,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6862011551856995,
      "eval_runtime": 7.0162,
      "eval_samples_per_second": 35.632,
      "eval_steps_per_second": 4.561,
      "step": 115
    },
    {
      "epoch": 1.8559999999999999,
      "grad_norm": 6.554261684417725,
      "learning_rate": 3.278688524590164e-06,
      "loss": 0.7358,
      "step": 116
    },
    {
      "epoch": 1.8559999999999999,
      "eval_accuracy": 0.588,
      "eval_loss": 0.686761736869812,
      "eval_runtime": 7.0114,
      "eval_samples_per_second": 35.656,
      "eval_steps_per_second": 4.564,
      "step": 116
    },
    {
      "epoch": 1.8719999999999999,
      "grad_norm": 4.159426689147949,
      "learning_rate": 2.8688524590163937e-06,
      "loss": 0.5127,
      "step": 117
    },
    {
      "epoch": 1.8719999999999999,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6867890357971191,
      "eval_runtime": 7.0154,
      "eval_samples_per_second": 35.636,
      "eval_steps_per_second": 4.561,
      "step": 117
    },
    {
      "epoch": 1.888,
      "grad_norm": 8.290789604187012,
      "learning_rate": 2.459016393442623e-06,
      "loss": 0.8021,
      "step": 118
    },
    {
      "epoch": 1.888,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6858710646629333,
      "eval_runtime": 7.0128,
      "eval_samples_per_second": 35.649,
      "eval_steps_per_second": 4.563,
      "step": 118
    },
    {
      "epoch": 1.904,
      "grad_norm": 10.31582260131836,
      "learning_rate": 2.049180327868852e-06,
      "loss": 0.5137,
      "step": 119
    },
    {
      "epoch": 1.904,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6865000128746033,
      "eval_runtime": 7.0153,
      "eval_samples_per_second": 35.637,
      "eval_steps_per_second": 4.561,
      "step": 119
    },
    {
      "epoch": 1.92,
      "grad_norm": 22.238492965698242,
      "learning_rate": 1.639344262295082e-06,
      "loss": 0.8707,
      "step": 120
    },
    {
      "epoch": 1.92,
      "eval_accuracy": 0.588,
      "eval_loss": 0.686654269695282,
      "eval_runtime": 7.0123,
      "eval_samples_per_second": 35.652,
      "eval_steps_per_second": 4.563,
      "step": 120
    },
    {
      "epoch": 1.936,
      "grad_norm": 12.90554141998291,
      "learning_rate": 1.2295081967213116e-06,
      "loss": 0.5289,
      "step": 121
    },
    {
      "epoch": 1.936,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6870156526565552,
      "eval_runtime": 7.0116,
      "eval_samples_per_second": 35.655,
      "eval_steps_per_second": 4.564,
      "step": 121
    },
    {
      "epoch": 1.952,
      "grad_norm": 4.567208766937256,
      "learning_rate": 8.19672131147541e-07,
      "loss": 0.6272,
      "step": 122
    },
    {
      "epoch": 1.952,
      "eval_accuracy": 0.588,
      "eval_loss": 0.6864765882492065,
      "eval_runtime": 7.0146,
      "eval_samples_per_second": 35.64,
      "eval_steps_per_second": 4.562,
      "step": 122
    },
    {
      "epoch": 1.968,
      "grad_norm": 4.912770748138428,
      "learning_rate": 4.098360655737705e-07,
      "loss": 0.7425,
      "step": 123
    },
    {
      "epoch": 1.968,
      "eval_accuracy": 0.584,
      "eval_loss": 0.6869745850563049,
      "eval_runtime": 7.0106,
      "eval_samples_per_second": 35.66,
      "eval_steps_per_second": 4.565,
      "step": 123
    },
    {
      "epoch": 1.984,
      "grad_norm": 10.384734153747559,
      "learning_rate": 0.0,
      "loss": 0.5696,
      "step": 124
    },
    {
      "epoch": 1.984,
      "eval_accuracy": 0.592,
      "eval_loss": 0.6866250038146973,
      "eval_runtime": 7.0175,
      "eval_samples_per_second": 35.625,
      "eval_steps_per_second": 4.56,
      "step": 124
    },
    {
      "epoch": 1.984,
      "step": 124,
      "total_flos": 41199005827072.0,
      "train_loss": 0.7272295798024824,
      "train_runtime": 1167.0627,
      "train_samples_per_second": 1.714,
      "train_steps_per_second": 0.106
    }
  ],
  "logging_steps": 1,
  "max_steps": 124,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 41199005827072.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}