{
  "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": 34.7608757019043,
      "learning_rate": 2.5e-05,
      "loss": 1.2054,
      "step": 1
    },
    {
      "epoch": 0.1111111111111111,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 1.0094876289367676,
      "eval_runtime": 0.7089,
      "eval_samples_per_second": 101.565,
      "eval_steps_per_second": 7.053,
      "step": 1
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": 30.633609771728516,
      "learning_rate": 5e-05,
      "loss": 1.1567,
      "step": 2
    },
    {
      "epoch": 0.2222222222222222,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.9987521767616272,
      "eval_runtime": 0.603,
      "eval_samples_per_second": 119.406,
      "eval_steps_per_second": 8.292,
      "step": 2
    },
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 25.385608673095703,
      "learning_rate": 4.943181818181818e-05,
      "loss": 1.0806,
      "step": 3
    },
    {
      "epoch": 0.3333333333333333,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.9628159999847412,
      "eval_runtime": 0.7075,
      "eval_samples_per_second": 101.767,
      "eval_steps_per_second": 7.067,
      "step": 3
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 22.26402473449707,
      "learning_rate": 4.886363636363637e-05,
      "loss": 0.9883,
      "step": 4
    },
    {
      "epoch": 0.4444444444444444,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.9259067177772522,
      "eval_runtime": 0.7031,
      "eval_samples_per_second": 102.402,
      "eval_steps_per_second": 7.111,
      "step": 4
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 26.522153854370117,
      "learning_rate": 4.829545454545455e-05,
      "loss": 0.9736,
      "step": 5
    },
    {
      "epoch": 0.5555555555555556,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.8975830078125,
      "eval_runtime": 0.7036,
      "eval_samples_per_second": 102.336,
      "eval_steps_per_second": 7.107,
      "step": 5
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 18.14297866821289,
      "learning_rate": 4.772727272727273e-05,
      "loss": 0.8267,
      "step": 6
    },
    {
      "epoch": 0.6666666666666666,
      "eval_accuracy": 0.4583333333333333,
      "eval_loss": 0.8745185136795044,
      "eval_runtime": 0.6043,
      "eval_samples_per_second": 119.149,
      "eval_steps_per_second": 8.274,
      "step": 6
    },
    {
      "epoch": 0.7777777777777778,
      "grad_norm": 10.767016410827637,
      "learning_rate": 4.715909090909091e-05,
      "loss": 0.8047,
      "step": 7
    },
    {
      "epoch": 0.7777777777777778,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8637288212776184,
      "eval_runtime": 0.7076,
      "eval_samples_per_second": 101.757,
      "eval_steps_per_second": 7.066,
      "step": 7
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 9.212442398071289,
      "learning_rate": 4.659090909090909e-05,
      "loss": 0.6807,
      "step": 8
    },
    {
      "epoch": 0.8888888888888888,
      "eval_accuracy": 0.4027777777777778,
      "eval_loss": 0.8712022304534912,
      "eval_runtime": 0.7061,
      "eval_samples_per_second": 101.97,
      "eval_steps_per_second": 7.081,
      "step": 8
    },
    {
      "epoch": 1.0,
      "grad_norm": 10.467320442199707,
      "learning_rate": 4.602272727272727e-05,
      "loss": 0.7425,
      "step": 9
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.8966064453125,
      "eval_runtime": 0.7054,
      "eval_samples_per_second": 102.068,
      "eval_steps_per_second": 7.088,
      "step": 9
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 8.758502960205078,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.7941,
      "step": 10
    },
    {
      "epoch": 1.1111111111111112,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.896728515625,
      "eval_runtime": 0.7073,
      "eval_samples_per_second": 101.79,
      "eval_steps_per_second": 7.069,
      "step": 10
    },
    {
      "epoch": 1.2222222222222223,
      "grad_norm": 2.704101800918579,
      "learning_rate": 4.488636363636364e-05,
      "loss": 0.6092,
      "step": 11
    },
    {
      "epoch": 1.2222222222222223,
      "eval_accuracy": 0.3333333333333333,
      "eval_loss": 0.8883192539215088,
      "eval_runtime": 0.603,
      "eval_samples_per_second": 119.401,
      "eval_steps_per_second": 8.292,
      "step": 11
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 3.9579689502716064,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.7372,
      "step": 12
    },
    {
      "epoch": 1.3333333333333333,
      "eval_accuracy": 0.3333333333333333,
      "eval_loss": 0.8713921308517456,
      "eval_runtime": 0.6557,
      "eval_samples_per_second": 109.8,
      "eval_steps_per_second": 7.625,
      "step": 12
    },
    {
      "epoch": 1.4444444444444444,
      "grad_norm": 4.962625980377197,
      "learning_rate": 4.375e-05,
      "loss": 0.6553,
      "step": 13
    },
    {
      "epoch": 1.4444444444444444,
      "eval_accuracy": 0.3194444444444444,
      "eval_loss": 0.8686930537223816,
      "eval_runtime": 0.7054,
      "eval_samples_per_second": 102.073,
      "eval_steps_per_second": 7.088,
      "step": 13
    },
    {
      "epoch": 1.5555555555555556,
      "grad_norm": 2.9537837505340576,
      "learning_rate": 4.318181818181819e-05,
      "loss": 0.7266,
      "step": 14
    },
    {
      "epoch": 1.5555555555555556,
      "eval_accuracy": 0.2777777777777778,
      "eval_loss": 0.8620063066482544,
      "eval_runtime": 0.7075,
      "eval_samples_per_second": 101.766,
      "eval_steps_per_second": 7.067,
      "step": 14
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 4.937398910522461,
      "learning_rate": 4.261363636363637e-05,
      "loss": 0.7462,
      "step": 15
    },
    {
      "epoch": 1.6666666666666665,
      "eval_accuracy": 0.3333333333333333,
      "eval_loss": 0.8434787392616272,
      "eval_runtime": 0.7043,
      "eval_samples_per_second": 102.225,
      "eval_steps_per_second": 7.099,
      "step": 15
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 8.647956848144531,
      "learning_rate": 4.204545454545455e-05,
      "loss": 0.66,
      "step": 16
    },
    {
      "epoch": 1.7777777777777777,
      "eval_accuracy": 0.375,
      "eval_loss": 0.8198378086090088,
      "eval_runtime": 0.7049,
      "eval_samples_per_second": 102.141,
      "eval_steps_per_second": 7.093,
      "step": 16
    },
    {
      "epoch": 1.8888888888888888,
      "grad_norm": 6.325115203857422,
      "learning_rate": 4.1477272727272734e-05,
      "loss": 0.6908,
      "step": 17
    },
    {
      "epoch": 1.8888888888888888,
      "eval_accuracy": 0.4444444444444444,
      "eval_loss": 0.8034396767616272,
      "eval_runtime": 0.6506,
      "eval_samples_per_second": 110.674,
      "eval_steps_per_second": 7.686,
      "step": 17
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.925394058227539,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.6993,
      "step": 18
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.7982177734375,
      "eval_runtime": 0.7049,
      "eval_samples_per_second": 102.145,
      "eval_steps_per_second": 7.093,
      "step": 18
    },
    {
      "epoch": 2.111111111111111,
      "grad_norm": 4.028661251068115,
      "learning_rate": 4.034090909090909e-05,
      "loss": 0.7788,
      "step": 19
    },
    {
      "epoch": 2.111111111111111,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.787353515625,
      "eval_runtime": 0.7038,
      "eval_samples_per_second": 102.296,
      "eval_steps_per_second": 7.104,
      "step": 19
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 3.346914529800415,
      "learning_rate": 3.9772727272727275e-05,
      "loss": 0.6594,
      "step": 20
    },
    {
      "epoch": 2.2222222222222223,
      "eval_accuracy": 0.5138888888888888,
      "eval_loss": 0.7767741084098816,
      "eval_runtime": 0.7031,
      "eval_samples_per_second": 102.403,
      "eval_steps_per_second": 7.111,
      "step": 20
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 5.400989532470703,
      "learning_rate": 3.9204545454545456e-05,
      "loss": 0.7303,
      "step": 21
    },
    {
      "epoch": 2.3333333333333335,
      "eval_accuracy": 0.5,
      "eval_loss": 0.7682834267616272,
      "eval_runtime": 0.6498,
      "eval_samples_per_second": 110.797,
      "eval_steps_per_second": 7.694,
      "step": 21
    },
    {
      "epoch": 2.4444444444444446,
      "grad_norm": 2.919682025909424,
      "learning_rate": 3.8636363636363636e-05,
      "loss": 0.6688,
      "step": 22
    },
    {
      "epoch": 2.4444444444444446,
      "eval_accuracy": 0.4861111111111111,
      "eval_loss": 0.7653401494026184,
      "eval_runtime": 0.551,
      "eval_samples_per_second": 130.66,
      "eval_steps_per_second": 9.074,
      "step": 22
    },
    {
      "epoch": 2.5555555555555554,
      "grad_norm": 9.212122917175293,
      "learning_rate": 3.8068181818181816e-05,
      "loss": 0.7337,
      "step": 23
    },
    {
      "epoch": 2.5555555555555554,
      "eval_accuracy": 0.5,
      "eval_loss": 0.7591145634651184,
      "eval_runtime": 0.7017,
      "eval_samples_per_second": 102.602,
      "eval_steps_per_second": 7.125,
      "step": 23
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 8.884961128234863,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.6648,
      "step": 24
    },
    {
      "epoch": 2.6666666666666665,
      "eval_accuracy": 0.5,
      "eval_loss": 0.7524074912071228,
      "eval_runtime": 0.7032,
      "eval_samples_per_second": 102.392,
      "eval_steps_per_second": 7.111,
      "step": 24
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 5.935830116271973,
      "learning_rate": 3.6931818181818184e-05,
      "loss": 0.6589,
      "step": 25
    },
    {
      "epoch": 2.7777777777777777,
      "eval_accuracy": 0.5138888888888888,
      "eval_loss": 0.7492743730545044,
      "eval_runtime": 0.6514,
      "eval_samples_per_second": 110.527,
      "eval_steps_per_second": 7.675,
      "step": 25
    },
    {
      "epoch": 2.888888888888889,
      "grad_norm": 7.1309027671813965,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 0.8163,
      "step": 26
    },
    {
      "epoch": 2.888888888888889,
      "eval_accuracy": 0.5416666666666666,
      "eval_loss": 0.7473212480545044,
      "eval_runtime": 0.706,
      "eval_samples_per_second": 101.99,
      "eval_steps_per_second": 7.083,
      "step": 26
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.935413837432861,
      "learning_rate": 3.579545454545455e-05,
      "loss": 0.6456,
      "step": 27
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5694444444444444,
      "eval_loss": 0.74639892578125,
      "eval_runtime": 0.6008,
      "eval_samples_per_second": 119.848,
      "eval_steps_per_second": 8.323,
      "step": 27
    },
    {
      "epoch": 3.111111111111111,
      "grad_norm": 4.931116104125977,
      "learning_rate": 3.522727272727273e-05,
      "loss": 0.7576,
      "step": 28
    },
    {
      "epoch": 3.111111111111111,
      "eval_accuracy": 0.5555555555555556,
      "eval_loss": 0.7495456337928772,
      "eval_runtime": 0.6526,
      "eval_samples_per_second": 110.329,
      "eval_steps_per_second": 7.662,
      "step": 28
    },
    {
      "epoch": 3.2222222222222223,
      "grad_norm": 8.242729187011719,
      "learning_rate": 3.465909090909091e-05,
      "loss": 0.6763,
      "step": 29
    },
    {
      "epoch": 3.2222222222222223,
      "eval_accuracy": 0.5555555555555556,
      "eval_loss": 0.7501423954963684,
      "eval_runtime": 0.7058,
      "eval_samples_per_second": 102.009,
      "eval_steps_per_second": 7.084,
      "step": 29
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 4.411537170410156,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.7565,
      "step": 30
    },
    {
      "epoch": 3.3333333333333335,
      "eval_accuracy": 0.5555555555555556,
      "eval_loss": 0.7513563632965088,
      "eval_runtime": 0.7022,
      "eval_samples_per_second": 102.536,
      "eval_steps_per_second": 7.121,
      "step": 30
    },
    {
      "epoch": 3.4444444444444446,
      "grad_norm": 5.082301139831543,
      "learning_rate": 3.352272727272727e-05,
      "loss": 0.7888,
      "step": 31
    },
    {
      "epoch": 3.4444444444444446,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.7555338740348816,
      "eval_runtime": 0.7031,
      "eval_samples_per_second": 102.41,
      "eval_steps_per_second": 7.112,
      "step": 31
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 3.5483062267303467,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.6769,
      "step": 32
    },
    {
      "epoch": 3.5555555555555554,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.7550930380821228,
      "eval_runtime": 0.6042,
      "eval_samples_per_second": 119.164,
      "eval_steps_per_second": 8.275,
      "step": 32
    },
    {
      "epoch": 3.6666666666666665,
      "grad_norm": 3.9409377574920654,
      "learning_rate": 3.238636363636364e-05,
      "loss": 0.718,
      "step": 33
    },
    {
      "epoch": 3.6666666666666665,
      "eval_accuracy": 0.5277777777777778,
      "eval_loss": 0.7593722939491272,
      "eval_runtime": 0.7024,
      "eval_samples_per_second": 102.508,
      "eval_steps_per_second": 7.119,
      "step": 33
    },
    {
      "epoch": 3.7777777777777777,
      "grad_norm": 11.279867172241211,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.6424,
      "step": 34
    },
    {
      "epoch": 3.7777777777777777,
      "eval_accuracy": 0.5138888888888888,
      "eval_loss": 0.7629123330116272,
      "eval_runtime": 0.7053,
      "eval_samples_per_second": 102.077,
      "eval_steps_per_second": 7.089,
      "step": 34
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 8.519107818603516,
      "learning_rate": 3.125e-05,
      "loss": 0.6471,
      "step": 35
    },
    {
      "epoch": 3.888888888888889,
      "eval_accuracy": 0.4861111111111111,
      "eval_loss": 0.7659505009651184,
      "eval_runtime": 0.703,
      "eval_samples_per_second": 102.418,
      "eval_steps_per_second": 7.112,
      "step": 35
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.580870628356934,
      "learning_rate": 3.068181818181818e-05,
      "loss": 0.727,
      "step": 36
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4027777777777778,
      "eval_loss": 0.7695041298866272,
      "eval_runtime": 0.6543,
      "eval_samples_per_second": 110.049,
      "eval_steps_per_second": 7.642,
      "step": 36
    },
    {
      "epoch": 4.111111111111111,
      "grad_norm": 11.691934585571289,
      "learning_rate": 3.0113636363636365e-05,
      "loss": 0.73,
      "step": 37
    },
    {
      "epoch": 4.111111111111111,
      "eval_accuracy": 0.4027777777777778,
      "eval_loss": 0.7756076455116272,
      "eval_runtime": 0.7028,
      "eval_samples_per_second": 102.442,
      "eval_steps_per_second": 7.114,
      "step": 37
    },
    {
      "epoch": 4.222222222222222,
      "grad_norm": 13.515392303466797,
      "learning_rate": 2.954545454545455e-05,
      "loss": 0.687,
      "step": 38
    },
    {
      "epoch": 4.222222222222222,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.7799343466758728,
      "eval_runtime": 0.705,
      "eval_samples_per_second": 102.128,
      "eval_steps_per_second": 7.092,
      "step": 38
    },
    {
      "epoch": 4.333333333333333,
      "grad_norm": 3.4968955516815186,
      "learning_rate": 2.8977272727272732e-05,
      "loss": 0.6573,
      "step": 39
    },
    {
      "epoch": 4.333333333333333,
      "eval_accuracy": 0.4583333333333333,
      "eval_loss": 0.7805582880973816,
      "eval_runtime": 0.7046,
      "eval_samples_per_second": 102.181,
      "eval_steps_per_second": 7.096,
      "step": 39
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 6.2186713218688965,
      "learning_rate": 2.8409090909090912e-05,
      "loss": 0.6691,
      "step": 40
    },
    {
      "epoch": 4.444444444444445,
      "eval_accuracy": 0.4861111111111111,
      "eval_loss": 0.7847764492034912,
      "eval_runtime": 0.7058,
      "eval_samples_per_second": 102.009,
      "eval_steps_per_second": 7.084,
      "step": 40
    },
    {
      "epoch": 4.555555555555555,
      "grad_norm": 9.075811386108398,
      "learning_rate": 2.784090909090909e-05,
      "loss": 0.8023,
      "step": 41
    },
    {
      "epoch": 4.555555555555555,
      "eval_accuracy": 0.4583333333333333,
      "eval_loss": 0.78857421875,
      "eval_runtime": 0.7051,
      "eval_samples_per_second": 102.113,
      "eval_steps_per_second": 7.091,
      "step": 41
    },
    {
      "epoch": 4.666666666666667,
      "grad_norm": 5.013810157775879,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.6703,
      "step": 42
    },
    {
      "epoch": 4.666666666666667,
      "eval_accuracy": 0.4861111111111111,
      "eval_loss": 0.7918837070465088,
      "eval_runtime": 0.655,
      "eval_samples_per_second": 109.915,
      "eval_steps_per_second": 7.633,
      "step": 42
    },
    {
      "epoch": 4.777777777777778,
      "grad_norm": 5.304051876068115,
      "learning_rate": 2.6704545454545453e-05,
      "loss": 0.7019,
      "step": 43
    },
    {
      "epoch": 4.777777777777778,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.7928195595741272,
      "eval_runtime": 0.6557,
      "eval_samples_per_second": 109.8,
      "eval_steps_per_second": 7.625,
      "step": 43
    },
    {
      "epoch": 4.888888888888889,
      "grad_norm": 7.697951316833496,
      "learning_rate": 2.6136363636363637e-05,
      "loss": 0.6973,
      "step": 44
    },
    {
      "epoch": 4.888888888888889,
      "eval_accuracy": 0.4861111111111111,
      "eval_loss": 0.8020155429840088,
      "eval_runtime": 0.6545,
      "eval_samples_per_second": 110.008,
      "eval_steps_per_second": 7.639,
      "step": 44
    },
    {
      "epoch": 5.0,
      "grad_norm": 18.014299392700195,
      "learning_rate": 2.5568181818181817e-05,
      "loss": 0.7859,
      "step": 45
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8017985224723816,
      "eval_runtime": 0.702,
      "eval_samples_per_second": 102.565,
      "eval_steps_per_second": 7.123,
      "step": 45
    },
    {
      "epoch": 5.111111111111111,
      "grad_norm": 5.591033458709717,
      "learning_rate": 2.5e-05,
      "loss": 0.6987,
      "step": 46
    },
    {
      "epoch": 5.111111111111111,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8059489130973816,
      "eval_runtime": 0.7047,
      "eval_samples_per_second": 102.172,
      "eval_steps_per_second": 7.095,
      "step": 46
    },
    {
      "epoch": 5.222222222222222,
      "grad_norm": 4.745278835296631,
      "learning_rate": 2.4431818181818185e-05,
      "loss": 0.6566,
      "step": 47
    },
    {
      "epoch": 5.222222222222222,
      "eval_accuracy": 0.4305555555555556,
      "eval_loss": 0.807861328125,
      "eval_runtime": 0.6555,
      "eval_samples_per_second": 109.837,
      "eval_steps_per_second": 7.628,
      "step": 47
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 7.4073486328125,
      "learning_rate": 2.3863636363636365e-05,
      "loss": 0.7671,
      "step": 48
    },
    {
      "epoch": 5.333333333333333,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8123643398284912,
      "eval_runtime": 0.7034,
      "eval_samples_per_second": 102.354,
      "eval_steps_per_second": 7.108,
      "step": 48
    },
    {
      "epoch": 5.444444444444445,
      "grad_norm": 7.266972541809082,
      "learning_rate": 2.3295454545454546e-05,
      "loss": 0.7223,
      "step": 49
    },
    {
      "epoch": 5.444444444444445,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8111029863357544,
      "eval_runtime": 0.7066,
      "eval_samples_per_second": 101.892,
      "eval_steps_per_second": 7.076,
      "step": 49
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 3.803126573562622,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.637,
      "step": 50
    },
    {
      "epoch": 5.555555555555555,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8150363564491272,
      "eval_runtime": 0.5565,
      "eval_samples_per_second": 129.373,
      "eval_steps_per_second": 8.984,
      "step": 50
    },
    {
      "epoch": 5.666666666666667,
      "grad_norm": 6.748312950134277,
      "learning_rate": 2.215909090909091e-05,
      "loss": 0.6406,
      "step": 51
    },
    {
      "epoch": 5.666666666666667,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8139106035232544,
      "eval_runtime": 0.6526,
      "eval_samples_per_second": 110.328,
      "eval_steps_per_second": 7.662,
      "step": 51
    },
    {
      "epoch": 5.777777777777778,
      "grad_norm": 14.67806625366211,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.7293,
      "step": 52
    },
    {
      "epoch": 5.777777777777778,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8112928867340088,
      "eval_runtime": 0.5507,
      "eval_samples_per_second": 130.733,
      "eval_steps_per_second": 9.079,
      "step": 52
    },
    {
      "epoch": 5.888888888888889,
      "grad_norm": 12.766432762145996,
      "learning_rate": 2.1022727272727274e-05,
      "loss": 0.6668,
      "step": 53
    },
    {
      "epoch": 5.888888888888889,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8097330927848816,
      "eval_runtime": 0.6502,
      "eval_samples_per_second": 110.73,
      "eval_steps_per_second": 7.69,
      "step": 53
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.800503730773926,
      "learning_rate": 2.0454545454545457e-05,
      "loss": 0.69,
      "step": 54
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8104926347732544,
      "eval_runtime": 0.5514,
      "eval_samples_per_second": 130.565,
      "eval_steps_per_second": 9.067,
      "step": 54
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 3.1313846111297607,
      "learning_rate": 1.9886363636363638e-05,
      "loss": 0.6828,
      "step": 55
    },
    {
      "epoch": 6.111111111111111,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8099636435508728,
      "eval_runtime": 0.5992,
      "eval_samples_per_second": 120.168,
      "eval_steps_per_second": 8.345,
      "step": 55
    },
    {
      "epoch": 6.222222222222222,
      "grad_norm": 4.510275363922119,
      "learning_rate": 1.9318181818181818e-05,
      "loss": 0.6665,
      "step": 56
    },
    {
      "epoch": 6.222222222222222,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8113335371017456,
      "eval_runtime": 0.5509,
      "eval_samples_per_second": 130.706,
      "eval_steps_per_second": 9.077,
      "step": 56
    },
    {
      "epoch": 6.333333333333333,
      "grad_norm": 8.366003036499023,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.7423,
      "step": 57
    },
    {
      "epoch": 6.333333333333333,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8116319179534912,
      "eval_runtime": 0.5995,
      "eval_samples_per_second": 120.099,
      "eval_steps_per_second": 8.34,
      "step": 57
    },
    {
      "epoch": 6.444444444444445,
      "grad_norm": 2.654489517211914,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.7059,
      "step": 58
    },
    {
      "epoch": 6.444444444444445,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.8082817792892456,
      "eval_runtime": 0.5516,
      "eval_samples_per_second": 130.519,
      "eval_steps_per_second": 9.064,
      "step": 58
    },
    {
      "epoch": 6.555555555555555,
      "grad_norm": 9.320557594299316,
      "learning_rate": 1.7613636363636366e-05,
      "loss": 0.7396,
      "step": 59
    },
    {
      "epoch": 6.555555555555555,
      "eval_accuracy": 0.375,
      "eval_loss": 0.8052436113357544,
      "eval_runtime": 0.5998,
      "eval_samples_per_second": 120.045,
      "eval_steps_per_second": 8.336,
      "step": 59
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 7.24456787109375,
      "learning_rate": 1.7045454545454546e-05,
      "loss": 0.6561,
      "step": 60
    },
    {
      "epoch": 6.666666666666667,
      "eval_accuracy": 0.375,
      "eval_loss": 0.8018798828125,
      "eval_runtime": 0.5995,
      "eval_samples_per_second": 120.107,
      "eval_steps_per_second": 8.341,
      "step": 60
    },
    {
      "epoch": 6.777777777777778,
      "grad_norm": 7.286713123321533,
      "learning_rate": 1.6477272727272726e-05,
      "loss": 0.7355,
      "step": 61
    },
    {
      "epoch": 6.777777777777778,
      "eval_accuracy": 0.375,
      "eval_loss": 0.8011067509651184,
      "eval_runtime": 0.6,
      "eval_samples_per_second": 119.998,
      "eval_steps_per_second": 8.333,
      "step": 61
    },
    {
      "epoch": 6.888888888888889,
      "grad_norm": 4.26874303817749,
      "learning_rate": 1.590909090909091e-05,
      "loss": 0.7353,
      "step": 62
    },
    {
      "epoch": 6.888888888888889,
      "eval_accuracy": 0.375,
      "eval_loss": 0.7994384765625,
      "eval_runtime": 0.5529,
      "eval_samples_per_second": 130.221,
      "eval_steps_per_second": 9.043,
      "step": 62
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.4262943267822266,
      "learning_rate": 1.534090909090909e-05,
      "loss": 0.6377,
      "step": 63
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.375,
      "eval_loss": 0.7936333417892456,
      "eval_runtime": 0.5486,
      "eval_samples_per_second": 131.251,
      "eval_steps_per_second": 9.115,
      "step": 63
    },
    {
      "epoch": 7.111111111111111,
      "grad_norm": 5.528657913208008,
      "learning_rate": 1.4772727272727274e-05,
      "loss": 0.6874,
      "step": 64
    },
    {
      "epoch": 7.111111111111111,
      "eval_accuracy": 0.375,
      "eval_loss": 0.7888047695159912,
      "eval_runtime": 0.6014,
      "eval_samples_per_second": 119.712,
      "eval_steps_per_second": 8.313,
      "step": 64
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 5.771848678588867,
      "learning_rate": 1.4204545454545456e-05,
      "loss": 0.7112,
      "step": 65
    },
    {
      "epoch": 7.222222222222222,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.7859971523284912,
      "eval_runtime": 0.6005,
      "eval_samples_per_second": 119.903,
      "eval_steps_per_second": 8.327,
      "step": 65
    },
    {
      "epoch": 7.333333333333333,
      "grad_norm": 8.485419273376465,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.6282,
      "step": 66
    },
    {
      "epoch": 7.333333333333333,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.7813856601715088,
      "eval_runtime": 0.6515,
      "eval_samples_per_second": 110.513,
      "eval_steps_per_second": 7.675,
      "step": 66
    },
    {
      "epoch": 7.444444444444445,
      "grad_norm": 10.057698249816895,
      "learning_rate": 1.3068181818181819e-05,
      "loss": 0.6793,
      "step": 67
    },
    {
      "epoch": 7.444444444444445,
      "eval_accuracy": 0.375,
      "eval_loss": 0.7774929404258728,
      "eval_runtime": 0.5496,
      "eval_samples_per_second": 131.003,
      "eval_steps_per_second": 9.097,
      "step": 67
    },
    {
      "epoch": 7.555555555555555,
      "grad_norm": 9.190048217773438,
      "learning_rate": 1.25e-05,
      "loss": 0.6707,
      "step": 68
    },
    {
      "epoch": 7.555555555555555,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7748887538909912,
      "eval_runtime": 0.5517,
      "eval_samples_per_second": 130.502,
      "eval_steps_per_second": 9.063,
      "step": 68
    },
    {
      "epoch": 7.666666666666667,
      "grad_norm": 6.497270107269287,
      "learning_rate": 1.1931818181818183e-05,
      "loss": 0.6571,
      "step": 69
    },
    {
      "epoch": 7.666666666666667,
      "eval_accuracy": 0.375,
      "eval_loss": 0.77294921875,
      "eval_runtime": 0.5507,
      "eval_samples_per_second": 130.752,
      "eval_steps_per_second": 9.08,
      "step": 69
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 5.636641025543213,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 0.6539,
      "step": 70
    },
    {
      "epoch": 7.777777777777778,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.7728678584098816,
      "eval_runtime": 0.5491,
      "eval_samples_per_second": 131.117,
      "eval_steps_per_second": 9.105,
      "step": 70
    },
    {
      "epoch": 7.888888888888889,
      "grad_norm": 4.822576522827148,
      "learning_rate": 1.0795454545454547e-05,
      "loss": 0.6953,
      "step": 71
    },
    {
      "epoch": 7.888888888888889,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7734103798866272,
      "eval_runtime": 0.5482,
      "eval_samples_per_second": 131.348,
      "eval_steps_per_second": 9.121,
      "step": 71
    },
    {
      "epoch": 8.0,
      "grad_norm": 7.65333366394043,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.7214,
      "step": 72
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.7741156816482544,
      "eval_runtime": 0.6017,
      "eval_samples_per_second": 119.662,
      "eval_steps_per_second": 8.31,
      "step": 72
    },
    {
      "epoch": 8.11111111111111,
      "grad_norm": 8.882245063781738,
      "learning_rate": 9.659090909090909e-06,
      "loss": 0.722,
      "step": 73
    },
    {
      "epoch": 8.11111111111111,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7755805253982544,
      "eval_runtime": 0.5489,
      "eval_samples_per_second": 131.162,
      "eval_steps_per_second": 9.108,
      "step": 73
    },
    {
      "epoch": 8.222222222222221,
      "grad_norm": 3.2345800399780273,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.6578,
      "step": 74
    },
    {
      "epoch": 8.222222222222221,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.771728515625,
      "eval_runtime": 0.6516,
      "eval_samples_per_second": 110.506,
      "eval_steps_per_second": 7.674,
      "step": 74
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 4.6726484298706055,
      "learning_rate": 8.522727272727273e-06,
      "loss": 0.7094,
      "step": 75
    },
    {
      "epoch": 8.333333333333334,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7747802734375,
      "eval_runtime": 0.5499,
      "eval_samples_per_second": 130.927,
      "eval_steps_per_second": 9.092,
      "step": 75
    },
    {
      "epoch": 8.444444444444445,
      "grad_norm": 16.0637264251709,
      "learning_rate": 7.954545454545455e-06,
      "loss": 0.7181,
      "step": 76
    },
    {
      "epoch": 8.444444444444445,
      "eval_accuracy": 0.375,
      "eval_loss": 0.7764350175857544,
      "eval_runtime": 0.7006,
      "eval_samples_per_second": 102.77,
      "eval_steps_per_second": 7.137,
      "step": 76
    },
    {
      "epoch": 8.555555555555555,
      "grad_norm": 4.427946090698242,
      "learning_rate": 7.386363636363637e-06,
      "loss": 0.6544,
      "step": 77
    },
    {
      "epoch": 8.555555555555555,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7750922441482544,
      "eval_runtime": 0.5503,
      "eval_samples_per_second": 130.846,
      "eval_steps_per_second": 9.086,
      "step": 77
    },
    {
      "epoch": 8.666666666666666,
      "grad_norm": 3.1340301036834717,
      "learning_rate": 6.818181818181818e-06,
      "loss": 0.6482,
      "step": 78
    },
    {
      "epoch": 8.666666666666666,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7764485478401184,
      "eval_runtime": 0.6071,
      "eval_samples_per_second": 118.588,
      "eval_steps_per_second": 8.235,
      "step": 78
    },
    {
      "epoch": 8.777777777777779,
      "grad_norm": 14.458281517028809,
      "learning_rate": 6.25e-06,
      "loss": 0.7313,
      "step": 79
    },
    {
      "epoch": 8.777777777777779,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7769097089767456,
      "eval_runtime": 0.6033,
      "eval_samples_per_second": 119.344,
      "eval_steps_per_second": 8.288,
      "step": 79
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 10.602180480957031,
      "learning_rate": 5.681818181818182e-06,
      "loss": 0.7158,
      "step": 80
    },
    {
      "epoch": 8.88888888888889,
      "eval_accuracy": 0.3611111111111111,
      "eval_loss": 0.7765977382659912,
      "eval_runtime": 0.7058,
      "eval_samples_per_second": 102.018,
      "eval_steps_per_second": 7.085,
      "step": 80
    },
    {
      "epoch": 9.0,
      "grad_norm": 8.127225875854492,
      "learning_rate": 5.113636363636364e-06,
      "loss": 0.6447,
      "step": 81
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.3472222222222222,
      "eval_loss": 0.7755669355392456,
      "eval_runtime": 0.6543,
      "eval_samples_per_second": 110.049,
      "eval_steps_per_second": 7.642,
      "step": 81
    },
    {
      "epoch": 9.11111111111111,
      "grad_norm": 6.457716464996338,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.6601,
      "step": 82
    },
    {
      "epoch": 9.11111111111111,
      "eval_accuracy": 0.3333333333333333,
      "eval_loss": 0.7762587070465088,
      "eval_runtime": 0.6069,
      "eval_samples_per_second": 118.636,
      "eval_steps_per_second": 8.239,
      "step": 82
    },
    {
      "epoch": 9.222222222222221,
      "grad_norm": 3.813633918762207,
      "learning_rate": 3.9772727272727275e-06,
      "loss": 0.7732,
      "step": 83
    },
    {
      "epoch": 9.222222222222221,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7770453691482544,
      "eval_runtime": 0.7076,
      "eval_samples_per_second": 101.748,
      "eval_steps_per_second": 7.066,
      "step": 83
    },
    {
      "epoch": 9.333333333333334,
      "grad_norm": 10.45854377746582,
      "learning_rate": 3.409090909090909e-06,
      "loss": 0.6657,
      "step": 84
    },
    {
      "epoch": 9.333333333333334,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7784559726715088,
      "eval_runtime": 0.7109,
      "eval_samples_per_second": 101.287,
      "eval_steps_per_second": 7.034,
      "step": 84
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 4.11740255355835,
      "learning_rate": 2.840909090909091e-06,
      "loss": 0.716,
      "step": 85
    },
    {
      "epoch": 9.444444444444445,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7778727412223816,
      "eval_runtime": 0.5543,
      "eval_samples_per_second": 129.897,
      "eval_steps_per_second": 9.021,
      "step": 85
    },
    {
      "epoch": 9.555555555555555,
      "grad_norm": 10.107967376708984,
      "learning_rate": 2.2727272727272728e-06,
      "loss": 0.7518,
      "step": 86
    },
    {
      "epoch": 9.555555555555555,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7781982421875,
      "eval_runtime": 0.6493,
      "eval_samples_per_second": 110.892,
      "eval_steps_per_second": 7.701,
      "step": 86
    },
    {
      "epoch": 9.666666666666666,
      "grad_norm": 6.475906848907471,
      "learning_rate": 1.7045454545454546e-06,
      "loss": 0.713,
      "step": 87
    },
    {
      "epoch": 9.666666666666666,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7775471806526184,
      "eval_runtime": 0.5511,
      "eval_samples_per_second": 130.655,
      "eval_steps_per_second": 9.073,
      "step": 87
    },
    {
      "epoch": 9.777777777777779,
      "grad_norm": 11.71182632446289,
      "learning_rate": 1.1363636363636364e-06,
      "loss": 0.6931,
      "step": 88
    },
    {
      "epoch": 9.777777777777779,
      "eval_accuracy": 0.375,
      "eval_loss": 0.7794325351715088,
      "eval_runtime": 0.5536,
      "eval_samples_per_second": 130.057,
      "eval_steps_per_second": 9.032,
      "step": 88
    },
    {
      "epoch": 9.88888888888889,
      "grad_norm": 8.962738037109375,
      "learning_rate": 5.681818181818182e-07,
      "loss": 0.7001,
      "step": 89
    },
    {
      "epoch": 9.88888888888889,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7806532382965088,
      "eval_runtime": 0.6003,
      "eval_samples_per_second": 119.933,
      "eval_steps_per_second": 8.329,
      "step": 89
    },
    {
      "epoch": 10.0,
      "grad_norm": 8.880528450012207,
      "learning_rate": 0.0,
      "loss": 0.6525,
      "step": 90
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.3888888888888889,
      "eval_loss": 0.7818332314491272,
      "eval_runtime": 0.6538,
      "eval_samples_per_second": 110.129,
      "eval_steps_per_second": 7.648,
      "step": 90
    },
    {
      "epoch": 10.0,
      "step": 90,
      "total_flos": 13264513597440.0,
      "train_loss": 0.7229804992675781,
      "train_runtime": 239.59,
      "train_samples_per_second": 11.854,
      "train_steps_per_second": 0.376
    }
  ],
  "logging_steps": 1,
  "max_steps": 90,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 13264513597440.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}