{
  "best_metric": 0.6124338624338624,
  "best_model_checkpoint": "vivit-b-16x2-kinetics400-ft-76388\\checkpoint-4773",
  "epoch": 49.01109090909091,
  "eval_steps": 500,
  "global_step": 5500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0018181818181818182,
      "grad_norm": 5.520795822143555,
      "learning_rate": 9.09090909090909e-07,
      "loss": 1.0986,
      "step": 10
    },
    {
      "epoch": 0.0036363636363636364,
      "grad_norm": 10.853554725646973,
      "learning_rate": 1.818181818181818e-06,
      "loss": 1.1002,
      "step": 20
    },
    {
      "epoch": 0.005454545454545455,
      "grad_norm": 6.47132682800293,
      "learning_rate": 2.7272727272727272e-06,
      "loss": 1.1019,
      "step": 30
    },
    {
      "epoch": 0.007272727272727273,
      "grad_norm": 6.264883041381836,
      "learning_rate": 3.636363636363636e-06,
      "loss": 1.0819,
      "step": 40
    },
    {
      "epoch": 0.00909090909090909,
      "grad_norm": 7.561264991760254,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 1.0847,
      "step": 50
    },
    {
      "epoch": 0.01090909090909091,
      "grad_norm": 4.925744533538818,
      "learning_rate": 5.4545454545454545e-06,
      "loss": 1.097,
      "step": 60
    },
    {
      "epoch": 0.012727272727272728,
      "grad_norm": 13.454752922058105,
      "learning_rate": 6.363636363636363e-06,
      "loss": 1.12,
      "step": 70
    },
    {
      "epoch": 0.014545454545454545,
      "grad_norm": 2.357506036758423,
      "learning_rate": 7.272727272727272e-06,
      "loss": 1.0849,
      "step": 80
    },
    {
      "epoch": 0.016363636363636365,
      "grad_norm": 6.968355655670166,
      "learning_rate": 8.181818181818183e-06,
      "loss": 1.0887,
      "step": 90
    },
    {
      "epoch": 0.01818181818181818,
      "grad_norm": 5.487667083740234,
      "learning_rate": 9.090909090909091e-06,
      "loss": 1.1272,
      "step": 100
    },
    {
      "epoch": 0.02,
      "grad_norm": 6.389057159423828,
      "learning_rate": 1e-05,
      "loss": 1.1083,
      "step": 110
    },
    {
      "epoch": 0.020181818181818183,
      "eval_accuracy": 0.33465608465608465,
      "eval_loss": 1.1111931800842285,
      "eval_runtime": 148.2258,
      "eval_samples_per_second": 5.1,
      "eval_steps_per_second": 0.641,
      "step": 111
    },
    {
      "epoch": 1.0016363636363637,
      "grad_norm": 5.377053260803223,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 1.0977,
      "step": 120
    },
    {
      "epoch": 1.0034545454545454,
      "grad_norm": 6.0129170417785645,
      "learning_rate": 1.1818181818181819e-05,
      "loss": 1.1504,
      "step": 130
    },
    {
      "epoch": 1.0052727272727273,
      "grad_norm": 7.416619777679443,
      "learning_rate": 1.2727272727272727e-05,
      "loss": 1.072,
      "step": 140
    },
    {
      "epoch": 1.007090909090909,
      "grad_norm": 4.585965156555176,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 1.0834,
      "step": 150
    },
    {
      "epoch": 1.008909090909091,
      "grad_norm": 9.934368133544922,
      "learning_rate": 1.4545454545454545e-05,
      "loss": 1.1117,
      "step": 160
    },
    {
      "epoch": 1.0107272727272727,
      "grad_norm": 7.8346381187438965,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 1.1222,
      "step": 170
    },
    {
      "epoch": 1.0125454545454546,
      "grad_norm": 4.306520462036133,
      "learning_rate": 1.6363636363636366e-05,
      "loss": 1.0952,
      "step": 180
    },
    {
      "epoch": 1.0143636363636364,
      "grad_norm": 7.647275924682617,
      "learning_rate": 1.7272727272727274e-05,
      "loss": 1.1341,
      "step": 190
    },
    {
      "epoch": 1.016181818181818,
      "grad_norm": 5.643113613128662,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 1.0344,
      "step": 200
    },
    {
      "epoch": 1.018,
      "grad_norm": 6.641160011291504,
      "learning_rate": 1.9090909090909094e-05,
      "loss": 1.1053,
      "step": 210
    },
    {
      "epoch": 1.0198181818181817,
      "grad_norm": 7.255860805511475,
      "learning_rate": 2e-05,
      "loss": 1.0789,
      "step": 220
    },
    {
      "epoch": 1.020181818181818,
      "eval_accuracy": 0.42592592592592593,
      "eval_loss": 1.0576354265213013,
      "eval_runtime": 149.5382,
      "eval_samples_per_second": 5.056,
      "eval_steps_per_second": 0.635,
      "step": 222
    },
    {
      "epoch": 2.0014545454545454,
      "grad_norm": 6.65429162979126,
      "learning_rate": 2.090909090909091e-05,
      "loss": 1.091,
      "step": 230
    },
    {
      "epoch": 2.0032727272727273,
      "grad_norm": 3.869105815887451,
      "learning_rate": 2.1818181818181818e-05,
      "loss": 1.0371,
      "step": 240
    },
    {
      "epoch": 2.0050909090909093,
      "grad_norm": 8.261011123657227,
      "learning_rate": 2.272727272727273e-05,
      "loss": 1.0778,
      "step": 250
    },
    {
      "epoch": 2.0069090909090908,
      "grad_norm": 8.152657508850098,
      "learning_rate": 2.3636363636363637e-05,
      "loss": 1.0327,
      "step": 260
    },
    {
      "epoch": 2.0087272727272727,
      "grad_norm": 6.811939239501953,
      "learning_rate": 2.4545454545454545e-05,
      "loss": 1.094,
      "step": 270
    },
    {
      "epoch": 2.0105454545454546,
      "grad_norm": 4.7158942222595215,
      "learning_rate": 2.5454545454545454e-05,
      "loss": 1.1247,
      "step": 280
    },
    {
      "epoch": 2.0123636363636366,
      "grad_norm": 4.587483882904053,
      "learning_rate": 2.636363636363636e-05,
      "loss": 1.1225,
      "step": 290
    },
    {
      "epoch": 2.014181818181818,
      "grad_norm": 3.0075366497039795,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 1.1186,
      "step": 300
    },
    {
      "epoch": 2.016,
      "grad_norm": 6.268433094024658,
      "learning_rate": 2.818181818181818e-05,
      "loss": 1.1272,
      "step": 310
    },
    {
      "epoch": 2.017818181818182,
      "grad_norm": 8.382287979125977,
      "learning_rate": 2.909090909090909e-05,
      "loss": 1.087,
      "step": 320
    },
    {
      "epoch": 2.0196363636363635,
      "grad_norm": 4.22382926940918,
      "learning_rate": 3e-05,
      "loss": 1.0767,
      "step": 330
    },
    {
      "epoch": 2.0201818181818183,
      "eval_accuracy": 0.4246031746031746,
      "eval_loss": 1.0863010883331299,
      "eval_runtime": 153.7648,
      "eval_samples_per_second": 4.917,
      "eval_steps_per_second": 0.618,
      "step": 333
    },
    {
      "epoch": 3.001272727272727,
      "grad_norm": 5.407969951629639,
      "learning_rate": 3.090909090909091e-05,
      "loss": 1.0193,
      "step": 340
    },
    {
      "epoch": 3.003090909090909,
      "grad_norm": 9.856935501098633,
      "learning_rate": 3.181818181818182e-05,
      "loss": 1.0889,
      "step": 350
    },
    {
      "epoch": 3.004909090909091,
      "grad_norm": 9.323800086975098,
      "learning_rate": 3.272727272727273e-05,
      "loss": 1.0824,
      "step": 360
    },
    {
      "epoch": 3.006727272727273,
      "grad_norm": 13.862144470214844,
      "learning_rate": 3.3636363636363636e-05,
      "loss": 1.1013,
      "step": 370
    },
    {
      "epoch": 3.0085454545454544,
      "grad_norm": 7.345916748046875,
      "learning_rate": 3.454545454545455e-05,
      "loss": 1.0071,
      "step": 380
    },
    {
      "epoch": 3.0103636363636364,
      "grad_norm": 3.808772087097168,
      "learning_rate": 3.545454545454546e-05,
      "loss": 1.0615,
      "step": 390
    },
    {
      "epoch": 3.0121818181818183,
      "grad_norm": 5.705800533294678,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 1.0663,
      "step": 400
    },
    {
      "epoch": 3.014,
      "grad_norm": 9.873875617980957,
      "learning_rate": 3.7272727272727276e-05,
      "loss": 1.0465,
      "step": 410
    },
    {
      "epoch": 3.0158181818181817,
      "grad_norm": 2.8121228218078613,
      "learning_rate": 3.818181818181819e-05,
      "loss": 1.1368,
      "step": 420
    },
    {
      "epoch": 3.0176363636363637,
      "grad_norm": 10.855435371398926,
      "learning_rate": 3.909090909090909e-05,
      "loss": 1.1093,
      "step": 430
    },
    {
      "epoch": 3.0194545454545456,
      "grad_norm": 4.415251731872559,
      "learning_rate": 4e-05,
      "loss": 1.1114,
      "step": 440
    },
    {
      "epoch": 3.0201818181818183,
      "eval_accuracy": 0.37037037037037035,
      "eval_loss": 1.1061029434204102,
      "eval_runtime": 149.7962,
      "eval_samples_per_second": 5.047,
      "eval_steps_per_second": 0.634,
      "step": 444
    },
    {
      "epoch": 4.001090909090909,
      "grad_norm": 2.8874096870422363,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 1.159,
      "step": 450
    },
    {
      "epoch": 4.002909090909091,
      "grad_norm": 15.52538776397705,
      "learning_rate": 4.181818181818182e-05,
      "loss": 1.0661,
      "step": 460
    },
    {
      "epoch": 4.004727272727273,
      "grad_norm": 8.193879127502441,
      "learning_rate": 4.2727272727272724e-05,
      "loss": 1.0345,
      "step": 470
    },
    {
      "epoch": 4.006545454545455,
      "grad_norm": 8.885357856750488,
      "learning_rate": 4.3636363636363636e-05,
      "loss": 1.01,
      "step": 480
    },
    {
      "epoch": 4.008363636363637,
      "grad_norm": 9.789897918701172,
      "learning_rate": 4.454545454545455e-05,
      "loss": 1.1213,
      "step": 490
    },
    {
      "epoch": 4.0101818181818185,
      "grad_norm": 11.676102638244629,
      "learning_rate": 4.545454545454546e-05,
      "loss": 1.111,
      "step": 500
    },
    {
      "epoch": 4.012,
      "grad_norm": 6.093684196472168,
      "learning_rate": 4.636363636363636e-05,
      "loss": 1.2359,
      "step": 510
    },
    {
      "epoch": 4.0138181818181815,
      "grad_norm": 8.584346771240234,
      "learning_rate": 4.7272727272727275e-05,
      "loss": 1.0954,
      "step": 520
    },
    {
      "epoch": 4.0156363636363634,
      "grad_norm": 5.42387580871582,
      "learning_rate": 4.8181818181818186e-05,
      "loss": 1.0584,
      "step": 530
    },
    {
      "epoch": 4.017454545454545,
      "grad_norm": 9.476812362670898,
      "learning_rate": 4.909090909090909e-05,
      "loss": 1.069,
      "step": 540
    },
    {
      "epoch": 4.019272727272727,
      "grad_norm": 8.928860664367676,
      "learning_rate": 5e-05,
      "loss": 1.0832,
      "step": 550
    },
    {
      "epoch": 4.020181818181818,
      "eval_accuracy": 0.4193121693121693,
      "eval_loss": 1.053645372390747,
      "eval_runtime": 148.4018,
      "eval_samples_per_second": 5.094,
      "eval_steps_per_second": 0.64,
      "step": 555
    },
    {
      "epoch": 5.000909090909091,
      "grad_norm": 3.7041380405426025,
      "learning_rate": 4.98989898989899e-05,
      "loss": 1.0869,
      "step": 560
    },
    {
      "epoch": 5.002727272727273,
      "grad_norm": 9.805282592773438,
      "learning_rate": 4.97979797979798e-05,
      "loss": 1.0385,
      "step": 570
    },
    {
      "epoch": 5.004545454545455,
      "grad_norm": 7.139551162719727,
      "learning_rate": 4.9696969696969694e-05,
      "loss": 1.1415,
      "step": 580
    },
    {
      "epoch": 5.006363636363637,
      "grad_norm": 5.1117682456970215,
      "learning_rate": 4.9595959595959594e-05,
      "loss": 1.0708,
      "step": 590
    },
    {
      "epoch": 5.008181818181818,
      "grad_norm": 5.95142126083374,
      "learning_rate": 4.94949494949495e-05,
      "loss": 1.1131,
      "step": 600
    },
    {
      "epoch": 5.01,
      "grad_norm": 4.644528388977051,
      "learning_rate": 4.93939393939394e-05,
      "loss": 1.0531,
      "step": 610
    },
    {
      "epoch": 5.011818181818182,
      "grad_norm": 5.309731483459473,
      "learning_rate": 4.92929292929293e-05,
      "loss": 1.0557,
      "step": 620
    },
    {
      "epoch": 5.013636363636364,
      "grad_norm": 2.6945815086364746,
      "learning_rate": 4.919191919191919e-05,
      "loss": 1.0923,
      "step": 630
    },
    {
      "epoch": 5.015454545454546,
      "grad_norm": 6.136772155761719,
      "learning_rate": 4.909090909090909e-05,
      "loss": 1.0675,
      "step": 640
    },
    {
      "epoch": 5.0172727272727276,
      "grad_norm": 3.571549415588379,
      "learning_rate": 4.898989898989899e-05,
      "loss": 1.0437,
      "step": 650
    },
    {
      "epoch": 5.0190909090909095,
      "grad_norm": 3.366074323654175,
      "learning_rate": 4.888888888888889e-05,
      "loss": 1.0622,
      "step": 660
    },
    {
      "epoch": 5.020181818181818,
      "eval_accuracy": 0.4576719576719577,
      "eval_loss": 1.0719990730285645,
      "eval_runtime": 148.7515,
      "eval_samples_per_second": 5.082,
      "eval_steps_per_second": 0.639,
      "step": 666
    },
    {
      "epoch": 6.000727272727273,
      "grad_norm": 8.250535011291504,
      "learning_rate": 4.878787878787879e-05,
      "loss": 0.9758,
      "step": 670
    },
    {
      "epoch": 6.002545454545454,
      "grad_norm": 3.9611146450042725,
      "learning_rate": 4.868686868686869e-05,
      "loss": 1.1303,
      "step": 680
    },
    {
      "epoch": 6.004363636363636,
      "grad_norm": 6.4248504638671875,
      "learning_rate": 4.858585858585859e-05,
      "loss": 0.9938,
      "step": 690
    },
    {
      "epoch": 6.006181818181818,
      "grad_norm": 6.866003513336182,
      "learning_rate": 4.848484848484849e-05,
      "loss": 0.9641,
      "step": 700
    },
    {
      "epoch": 6.008,
      "grad_norm": 11.4672269821167,
      "learning_rate": 4.838383838383839e-05,
      "loss": 1.0336,
      "step": 710
    },
    {
      "epoch": 6.009818181818182,
      "grad_norm": 16.061107635498047,
      "learning_rate": 4.828282828282829e-05,
      "loss": 1.0764,
      "step": 720
    },
    {
      "epoch": 6.011636363636364,
      "grad_norm": 4.682388782501221,
      "learning_rate": 4.8181818181818186e-05,
      "loss": 1.0844,
      "step": 730
    },
    {
      "epoch": 6.013454545454546,
      "grad_norm": 7.241218566894531,
      "learning_rate": 4.808080808080808e-05,
      "loss": 1.0297,
      "step": 740
    },
    {
      "epoch": 6.015272727272727,
      "grad_norm": 7.628449440002441,
      "learning_rate": 4.797979797979798e-05,
      "loss": 1.1384,
      "step": 750
    },
    {
      "epoch": 6.017090909090909,
      "grad_norm": 6.4355878829956055,
      "learning_rate": 4.787878787878788e-05,
      "loss": 1.0197,
      "step": 760
    },
    {
      "epoch": 6.018909090909091,
      "grad_norm": 4.381698131561279,
      "learning_rate": 4.7777777777777784e-05,
      "loss": 1.0874,
      "step": 770
    },
    {
      "epoch": 6.020181818181818,
      "eval_accuracy": 0.4708994708994709,
      "eval_loss": 1.0304068326950073,
      "eval_runtime": 148.4186,
      "eval_samples_per_second": 5.094,
      "eval_steps_per_second": 0.64,
      "step": 777
    },
    {
      "epoch": 7.000545454545454,
      "grad_norm": 8.861906051635742,
      "learning_rate": 4.7676767676767684e-05,
      "loss": 1.0022,
      "step": 780
    },
    {
      "epoch": 7.002363636363636,
      "grad_norm": 5.51254940032959,
      "learning_rate": 4.7575757575757576e-05,
      "loss": 1.0364,
      "step": 790
    },
    {
      "epoch": 7.004181818181818,
      "grad_norm": 7.339147090911865,
      "learning_rate": 4.7474747474747476e-05,
      "loss": 1.0236,
      "step": 800
    },
    {
      "epoch": 7.006,
      "grad_norm": 5.729611396789551,
      "learning_rate": 4.7373737373737375e-05,
      "loss": 1.0856,
      "step": 810
    },
    {
      "epoch": 7.007818181818182,
      "grad_norm": 11.61246109008789,
      "learning_rate": 4.7272727272727275e-05,
      "loss": 1.0058,
      "step": 820
    },
    {
      "epoch": 7.009636363636363,
      "grad_norm": 7.409841060638428,
      "learning_rate": 4.7171717171717174e-05,
      "loss": 0.9701,
      "step": 830
    },
    {
      "epoch": 7.011454545454545,
      "grad_norm": 9.707467079162598,
      "learning_rate": 4.7070707070707074e-05,
      "loss": 0.9544,
      "step": 840
    },
    {
      "epoch": 7.013272727272727,
      "grad_norm": 4.250189781188965,
      "learning_rate": 4.696969696969697e-05,
      "loss": 1.0981,
      "step": 850
    },
    {
      "epoch": 7.015090909090909,
      "grad_norm": 5.50283670425415,
      "learning_rate": 4.686868686868687e-05,
      "loss": 1.0752,
      "step": 860
    },
    {
      "epoch": 7.016909090909091,
      "grad_norm": 10.252998352050781,
      "learning_rate": 4.676767676767677e-05,
      "loss": 1.0213,
      "step": 870
    },
    {
      "epoch": 7.018727272727273,
      "grad_norm": 9.177423477172852,
      "learning_rate": 4.666666666666667e-05,
      "loss": 0.9742,
      "step": 880
    },
    {
      "epoch": 7.020181818181818,
      "eval_accuracy": 0.45105820105820105,
      "eval_loss": 1.0340412855148315,
      "eval_runtime": 148.8072,
      "eval_samples_per_second": 5.08,
      "eval_steps_per_second": 0.638,
      "step": 888
    },
    {
      "epoch": 8.000363636363636,
      "grad_norm": 7.860219478607178,
      "learning_rate": 4.656565656565657e-05,
      "loss": 0.9891,
      "step": 890
    },
    {
      "epoch": 8.002181818181818,
      "grad_norm": 11.450358390808105,
      "learning_rate": 4.6464646464646464e-05,
      "loss": 0.966,
      "step": 900
    },
    {
      "epoch": 8.004,
      "grad_norm": 7.019264221191406,
      "learning_rate": 4.636363636363636e-05,
      "loss": 1.0707,
      "step": 910
    },
    {
      "epoch": 8.005818181818182,
      "grad_norm": 6.124785900115967,
      "learning_rate": 4.626262626262626e-05,
      "loss": 0.8674,
      "step": 920
    },
    {
      "epoch": 8.007636363636363,
      "grad_norm": 5.532207012176514,
      "learning_rate": 4.616161616161616e-05,
      "loss": 1.1394,
      "step": 930
    },
    {
      "epoch": 8.009454545454545,
      "grad_norm": 2.409095287322998,
      "learning_rate": 4.606060606060607e-05,
      "loss": 1.0774,
      "step": 940
    },
    {
      "epoch": 8.011272727272727,
      "grad_norm": 5.863190174102783,
      "learning_rate": 4.595959595959596e-05,
      "loss": 1.0996,
      "step": 950
    },
    {
      "epoch": 8.01309090909091,
      "grad_norm": 5.189179420471191,
      "learning_rate": 4.585858585858586e-05,
      "loss": 1.0476,
      "step": 960
    },
    {
      "epoch": 8.014909090909091,
      "grad_norm": 6.774593353271484,
      "learning_rate": 4.575757575757576e-05,
      "loss": 1.0299,
      "step": 970
    },
    {
      "epoch": 8.016727272727273,
      "grad_norm": 9.079130172729492,
      "learning_rate": 4.565656565656566e-05,
      "loss": 1.0262,
      "step": 980
    },
    {
      "epoch": 8.018545454545455,
      "grad_norm": 6.1006760597229,
      "learning_rate": 4.555555555555556e-05,
      "loss": 0.9848,
      "step": 990
    },
    {
      "epoch": 8.020181818181818,
      "eval_accuracy": 0.4669312169312169,
      "eval_loss": 1.036673903465271,
      "eval_runtime": 146.7468,
      "eval_samples_per_second": 5.152,
      "eval_steps_per_second": 0.647,
      "step": 999
    },
    {
      "epoch": 9.000181818181819,
      "grad_norm": 6.422980308532715,
      "learning_rate": 4.545454545454546e-05,
      "loss": 1.1866,
      "step": 1000
    },
    {
      "epoch": 9.002,
      "grad_norm": 5.095742702484131,
      "learning_rate": 4.535353535353535e-05,
      "loss": 1.0583,
      "step": 1010
    },
    {
      "epoch": 9.003818181818183,
      "grad_norm": 6.119103908538818,
      "learning_rate": 4.525252525252526e-05,
      "loss": 1.0929,
      "step": 1020
    },
    {
      "epoch": 9.005636363636363,
      "grad_norm": 4.790987968444824,
      "learning_rate": 4.515151515151516e-05,
      "loss": 1.0684,
      "step": 1030
    },
    {
      "epoch": 9.007454545454545,
      "grad_norm": 5.97383975982666,
      "learning_rate": 4.5050505050505056e-05,
      "loss": 1.0431,
      "step": 1040
    },
    {
      "epoch": 9.009272727272727,
      "grad_norm": 3.85739803314209,
      "learning_rate": 4.494949494949495e-05,
      "loss": 0.934,
      "step": 1050
    },
    {
      "epoch": 9.011090909090909,
      "grad_norm": 3.361600875854492,
      "learning_rate": 4.484848484848485e-05,
      "loss": 0.9329,
      "step": 1060
    },
    {
      "epoch": 9.01290909090909,
      "grad_norm": 7.449879169464111,
      "learning_rate": 4.474747474747475e-05,
      "loss": 1.0629,
      "step": 1070
    },
    {
      "epoch": 9.014727272727272,
      "grad_norm": 7.561132431030273,
      "learning_rate": 4.464646464646465e-05,
      "loss": 1.058,
      "step": 1080
    },
    {
      "epoch": 9.016545454545454,
      "grad_norm": 12.090225219726562,
      "learning_rate": 4.454545454545455e-05,
      "loss": 1.1276,
      "step": 1090
    },
    {
      "epoch": 9.018363636363636,
      "grad_norm": 5.2706708908081055,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.9764,
      "step": 1100
    },
    {
      "epoch": 9.020181818181818,
      "grad_norm": 14.38776969909668,
      "learning_rate": 4.4343434343434346e-05,
      "loss": 1.12,
      "step": 1110
    },
    {
      "epoch": 9.020181818181818,
      "eval_accuracy": 0.4193121693121693,
      "eval_loss": 1.026868224143982,
      "eval_runtime": 149.067,
      "eval_samples_per_second": 5.072,
      "eval_steps_per_second": 0.637,
      "step": 1110
    },
    {
      "epoch": 10.001818181818182,
      "grad_norm": 4.231562614440918,
      "learning_rate": 4.4242424242424246e-05,
      "loss": 1.0368,
      "step": 1120
    },
    {
      "epoch": 10.003636363636364,
      "grad_norm": 6.167897701263428,
      "learning_rate": 4.4141414141414145e-05,
      "loss": 1.0565,
      "step": 1130
    },
    {
      "epoch": 10.005454545454546,
      "grad_norm": 5.133641719818115,
      "learning_rate": 4.4040404040404044e-05,
      "loss": 1.0345,
      "step": 1140
    },
    {
      "epoch": 10.007272727272728,
      "grad_norm": 4.132250785827637,
      "learning_rate": 4.3939393939393944e-05,
      "loss": 0.9971,
      "step": 1150
    },
    {
      "epoch": 10.00909090909091,
      "grad_norm": 4.209298610687256,
      "learning_rate": 4.383838383838384e-05,
      "loss": 1.0099,
      "step": 1160
    },
    {
      "epoch": 10.010909090909092,
      "grad_norm": 6.258595943450928,
      "learning_rate": 4.3737373737373736e-05,
      "loss": 1.0037,
      "step": 1170
    },
    {
      "epoch": 10.012727272727274,
      "grad_norm": 9.644258499145508,
      "learning_rate": 4.3636363636363636e-05,
      "loss": 0.9807,
      "step": 1180
    },
    {
      "epoch": 10.014545454545454,
      "grad_norm": 6.8741912841796875,
      "learning_rate": 4.3535353535353535e-05,
      "loss": 0.9534,
      "step": 1190
    },
    {
      "epoch": 10.016363636363636,
      "grad_norm": 4.356405735015869,
      "learning_rate": 4.343434343434344e-05,
      "loss": 1.0285,
      "step": 1200
    },
    {
      "epoch": 10.018181818181818,
      "grad_norm": 4.656445503234863,
      "learning_rate": 4.3333333333333334e-05,
      "loss": 1.0168,
      "step": 1210
    },
    {
      "epoch": 10.02,
      "grad_norm": 6.54704475402832,
      "learning_rate": 4.3232323232323234e-05,
      "loss": 1.0484,
      "step": 1220
    },
    {
      "epoch": 10.020181818181818,
      "eval_accuracy": 0.45105820105820105,
      "eval_loss": 1.0105139017105103,
      "eval_runtime": 146.4072,
      "eval_samples_per_second": 5.164,
      "eval_steps_per_second": 0.649,
      "step": 1221
    },
    {
      "epoch": 11.001636363636363,
      "grad_norm": 3.6094846725463867,
      "learning_rate": 4.313131313131313e-05,
      "loss": 0.9776,
      "step": 1230
    },
    {
      "epoch": 11.003454545454545,
      "grad_norm": 4.734577178955078,
      "learning_rate": 4.303030303030303e-05,
      "loss": 0.971,
      "step": 1240
    },
    {
      "epoch": 11.005272727272727,
      "grad_norm": 4.031460285186768,
      "learning_rate": 4.292929292929293e-05,
      "loss": 0.9972,
      "step": 1250
    },
    {
      "epoch": 11.007090909090909,
      "grad_norm": 11.494680404663086,
      "learning_rate": 4.282828282828283e-05,
      "loss": 0.9916,
      "step": 1260
    },
    {
      "epoch": 11.008909090909091,
      "grad_norm": 6.08104944229126,
      "learning_rate": 4.2727272727272724e-05,
      "loss": 1.0135,
      "step": 1270
    },
    {
      "epoch": 11.010727272727273,
      "grad_norm": 4.234505653381348,
      "learning_rate": 4.262626262626263e-05,
      "loss": 0.9287,
      "step": 1280
    },
    {
      "epoch": 11.012545454545455,
      "grad_norm": 6.952130317687988,
      "learning_rate": 4.252525252525253e-05,
      "loss": 0.9476,
      "step": 1290
    },
    {
      "epoch": 11.014363636363637,
      "grad_norm": 7.385903358459473,
      "learning_rate": 4.242424242424243e-05,
      "loss": 0.9915,
      "step": 1300
    },
    {
      "epoch": 11.016181818181819,
      "grad_norm": 6.446252822875977,
      "learning_rate": 4.232323232323233e-05,
      "loss": 1.0603,
      "step": 1310
    },
    {
      "epoch": 11.018,
      "grad_norm": 5.797521591186523,
      "learning_rate": 4.222222222222222e-05,
      "loss": 1.1218,
      "step": 1320
    },
    {
      "epoch": 11.019818181818183,
      "grad_norm": 6.516293048858643,
      "learning_rate": 4.212121212121212e-05,
      "loss": 0.9445,
      "step": 1330
    },
    {
      "epoch": 11.020181818181818,
      "eval_accuracy": 0.4880952380952381,
      "eval_loss": 1.0051753520965576,
      "eval_runtime": 152.1617,
      "eval_samples_per_second": 4.968,
      "eval_steps_per_second": 0.624,
      "step": 1332
    },
    {
      "epoch": 12.001454545454546,
      "grad_norm": 7.457939624786377,
      "learning_rate": 4.202020202020202e-05,
      "loss": 1.0431,
      "step": 1340
    },
    {
      "epoch": 12.003272727272726,
      "grad_norm": 6.6593217849731445,
      "learning_rate": 4.191919191919192e-05,
      "loss": 1.0327,
      "step": 1350
    },
    {
      "epoch": 12.005090909090908,
      "grad_norm": 4.458651542663574,
      "learning_rate": 4.181818181818182e-05,
      "loss": 0.9756,
      "step": 1360
    },
    {
      "epoch": 12.00690909090909,
      "grad_norm": 5.859511852264404,
      "learning_rate": 4.171717171717172e-05,
      "loss": 1.039,
      "step": 1370
    },
    {
      "epoch": 12.008727272727272,
      "grad_norm": 9.763144493103027,
      "learning_rate": 4.161616161616162e-05,
      "loss": 1.0068,
      "step": 1380
    },
    {
      "epoch": 12.010545454545454,
      "grad_norm": 4.887930870056152,
      "learning_rate": 4.151515151515152e-05,
      "loss": 0.9022,
      "step": 1390
    },
    {
      "epoch": 12.012363636363636,
      "grad_norm": 6.166642665863037,
      "learning_rate": 4.141414141414142e-05,
      "loss": 0.9438,
      "step": 1400
    },
    {
      "epoch": 12.014181818181818,
      "grad_norm": 13.519564628601074,
      "learning_rate": 4.131313131313132e-05,
      "loss": 1.05,
      "step": 1410
    },
    {
      "epoch": 12.016,
      "grad_norm": 8.109920501708984,
      "learning_rate": 4.1212121212121216e-05,
      "loss": 0.9257,
      "step": 1420
    },
    {
      "epoch": 12.017818181818182,
      "grad_norm": 6.31077241897583,
      "learning_rate": 4.111111111111111e-05,
      "loss": 0.8967,
      "step": 1430
    },
    {
      "epoch": 12.019636363636364,
      "grad_norm": 8.095528602600098,
      "learning_rate": 4.101010101010101e-05,
      "loss": 1.032,
      "step": 1440
    },
    {
      "epoch": 12.020181818181818,
      "eval_accuracy": 0.4523809523809524,
      "eval_loss": 1.0365078449249268,
      "eval_runtime": 148.6389,
      "eval_samples_per_second": 5.086,
      "eval_steps_per_second": 0.639,
      "step": 1443
    },
    {
      "epoch": 13.001272727272728,
      "grad_norm": 4.897077560424805,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.954,
      "step": 1450
    },
    {
      "epoch": 13.00309090909091,
      "grad_norm": 5.455181121826172,
      "learning_rate": 4.0808080808080814e-05,
      "loss": 0.9739,
      "step": 1460
    },
    {
      "epoch": 13.004909090909091,
      "grad_norm": 6.275787830352783,
      "learning_rate": 4.070707070707071e-05,
      "loss": 1.0453,
      "step": 1470
    },
    {
      "epoch": 13.006727272727273,
      "grad_norm": 4.212862491607666,
      "learning_rate": 4.0606060606060606e-05,
      "loss": 1.1058,
      "step": 1480
    },
    {
      "epoch": 13.008545454545455,
      "grad_norm": 3.4009814262390137,
      "learning_rate": 4.0505050505050506e-05,
      "loss": 1.0064,
      "step": 1490
    },
    {
      "epoch": 13.010363636363637,
      "grad_norm": 5.143698692321777,
      "learning_rate": 4.0404040404040405e-05,
      "loss": 0.9478,
      "step": 1500
    },
    {
      "epoch": 13.012181818181817,
      "grad_norm": 5.192939758300781,
      "learning_rate": 4.0303030303030305e-05,
      "loss": 0.9567,
      "step": 1510
    },
    {
      "epoch": 13.014,
      "grad_norm": 4.058863162994385,
      "learning_rate": 4.0202020202020204e-05,
      "loss": 0.9702,
      "step": 1520
    },
    {
      "epoch": 13.015818181818181,
      "grad_norm": 4.788520336151123,
      "learning_rate": 4.01010101010101e-05,
      "loss": 0.8961,
      "step": 1530
    },
    {
      "epoch": 13.017636363636363,
      "grad_norm": 8.213604927062988,
      "learning_rate": 4e-05,
      "loss": 1.0142,
      "step": 1540
    },
    {
      "epoch": 13.019454545454545,
      "grad_norm": 4.203430652618408,
      "learning_rate": 3.98989898989899e-05,
      "loss": 0.987,
      "step": 1550
    },
    {
      "epoch": 13.020181818181818,
      "eval_accuracy": 0.5105820105820106,
      "eval_loss": 1.001942753791809,
      "eval_runtime": 146.3151,
      "eval_samples_per_second": 5.167,
      "eval_steps_per_second": 0.649,
      "step": 1554
    },
    {
      "epoch": 14.001090909090909,
      "grad_norm": 5.082915782928467,
      "learning_rate": 3.97979797979798e-05,
      "loss": 0.9439,
      "step": 1560
    },
    {
      "epoch": 14.00290909090909,
      "grad_norm": 5.0494256019592285,
      "learning_rate": 3.96969696969697e-05,
      "loss": 0.918,
      "step": 1570
    },
    {
      "epoch": 14.004727272727273,
      "grad_norm": 5.370521068572998,
      "learning_rate": 3.9595959595959594e-05,
      "loss": 1.0225,
      "step": 1580
    },
    {
      "epoch": 14.006545454545455,
      "grad_norm": 8.859224319458008,
      "learning_rate": 3.9494949494949494e-05,
      "loss": 0.9746,
      "step": 1590
    },
    {
      "epoch": 14.008363636363637,
      "grad_norm": 6.678452491760254,
      "learning_rate": 3.939393939393939e-05,
      "loss": 1.0394,
      "step": 1600
    },
    {
      "epoch": 14.010181818181819,
      "grad_norm": 9.348724365234375,
      "learning_rate": 3.929292929292929e-05,
      "loss": 1.0252,
      "step": 1610
    },
    {
      "epoch": 14.012,
      "grad_norm": 9.378219604492188,
      "learning_rate": 3.91919191919192e-05,
      "loss": 1.0869,
      "step": 1620
    },
    {
      "epoch": 14.013818181818182,
      "grad_norm": 7.069434642791748,
      "learning_rate": 3.909090909090909e-05,
      "loss": 0.8797,
      "step": 1630
    },
    {
      "epoch": 14.015636363636364,
      "grad_norm": 4.11262845993042,
      "learning_rate": 3.898989898989899e-05,
      "loss": 0.9745,
      "step": 1640
    },
    {
      "epoch": 14.017454545454546,
      "grad_norm": 4.320169448852539,
      "learning_rate": 3.888888888888889e-05,
      "loss": 1.0043,
      "step": 1650
    },
    {
      "epoch": 14.019272727272726,
      "grad_norm": 6.942636489868164,
      "learning_rate": 3.878787878787879e-05,
      "loss": 1.0797,
      "step": 1660
    },
    {
      "epoch": 14.020181818181818,
      "eval_accuracy": 0.4656084656084656,
      "eval_loss": 1.0128320455551147,
      "eval_runtime": 150.7976,
      "eval_samples_per_second": 5.013,
      "eval_steps_per_second": 0.63,
      "step": 1665
    },
    {
      "epoch": 15.00090909090909,
      "grad_norm": 5.204150199890137,
      "learning_rate": 3.868686868686869e-05,
      "loss": 0.9386,
      "step": 1670
    },
    {
      "epoch": 15.002727272727272,
      "grad_norm": 5.085378646850586,
      "learning_rate": 3.858585858585859e-05,
      "loss": 1.0595,
      "step": 1680
    },
    {
      "epoch": 15.004545454545454,
      "grad_norm": 9.999809265136719,
      "learning_rate": 3.848484848484848e-05,
      "loss": 1.0415,
      "step": 1690
    },
    {
      "epoch": 15.006363636363636,
      "grad_norm": 2.8342270851135254,
      "learning_rate": 3.838383838383838e-05,
      "loss": 0.9376,
      "step": 1700
    },
    {
      "epoch": 15.008181818181818,
      "grad_norm": 6.779669761657715,
      "learning_rate": 3.828282828282829e-05,
      "loss": 0.971,
      "step": 1710
    },
    {
      "epoch": 15.01,
      "grad_norm": 8.835999488830566,
      "learning_rate": 3.818181818181819e-05,
      "loss": 0.9513,
      "step": 1720
    },
    {
      "epoch": 15.011818181818182,
      "grad_norm": 7.420837879180908,
      "learning_rate": 3.8080808080808087e-05,
      "loss": 0.9304,
      "step": 1730
    },
    {
      "epoch": 15.013636363636364,
      "grad_norm": 4.993125915527344,
      "learning_rate": 3.797979797979798e-05,
      "loss": 0.9871,
      "step": 1740
    },
    {
      "epoch": 15.015454545454546,
      "grad_norm": 8.37247371673584,
      "learning_rate": 3.787878787878788e-05,
      "loss": 0.9466,
      "step": 1750
    },
    {
      "epoch": 15.017272727272728,
      "grad_norm": 7.0074663162231445,
      "learning_rate": 3.777777777777778e-05,
      "loss": 0.8956,
      "step": 1760
    },
    {
      "epoch": 15.01909090909091,
      "grad_norm": 9.646268844604492,
      "learning_rate": 3.767676767676768e-05,
      "loss": 0.9196,
      "step": 1770
    },
    {
      "epoch": 15.020181818181818,
      "eval_accuracy": 0.5013227513227513,
      "eval_loss": 1.0431362390518188,
      "eval_runtime": 147.3058,
      "eval_samples_per_second": 5.132,
      "eval_steps_per_second": 0.645,
      "step": 1776
    },
    {
      "epoch": 16.00072727272727,
      "grad_norm": 7.494339942932129,
      "learning_rate": 3.757575757575758e-05,
      "loss": 0.9682,
      "step": 1780
    },
    {
      "epoch": 16.002545454545455,
      "grad_norm": 10.007247924804688,
      "learning_rate": 3.747474747474748e-05,
      "loss": 1.0289,
      "step": 1790
    },
    {
      "epoch": 16.004363636363635,
      "grad_norm": 5.851564884185791,
      "learning_rate": 3.7373737373737376e-05,
      "loss": 0.9664,
      "step": 1800
    },
    {
      "epoch": 16.00618181818182,
      "grad_norm": 7.351463794708252,
      "learning_rate": 3.7272727272727276e-05,
      "loss": 0.9844,
      "step": 1810
    },
    {
      "epoch": 16.008,
      "grad_norm": 5.768230438232422,
      "learning_rate": 3.7171717171717175e-05,
      "loss": 1.0384,
      "step": 1820
    },
    {
      "epoch": 16.009818181818183,
      "grad_norm": 6.959085941314697,
      "learning_rate": 3.7070707070707075e-05,
      "loss": 0.9584,
      "step": 1830
    },
    {
      "epoch": 16.011636363636363,
      "grad_norm": 8.661413192749023,
      "learning_rate": 3.6969696969696974e-05,
      "loss": 1.001,
      "step": 1840
    },
    {
      "epoch": 16.013454545454547,
      "grad_norm": 4.314298152923584,
      "learning_rate": 3.686868686868687e-05,
      "loss": 0.9738,
      "step": 1850
    },
    {
      "epoch": 16.015272727272727,
      "grad_norm": 6.815008163452148,
      "learning_rate": 3.6767676767676766e-05,
      "loss": 0.9342,
      "step": 1860
    },
    {
      "epoch": 16.01709090909091,
      "grad_norm": 10.533506393432617,
      "learning_rate": 3.6666666666666666e-05,
      "loss": 0.9174,
      "step": 1870
    },
    {
      "epoch": 16.01890909090909,
      "grad_norm": 13.072206497192383,
      "learning_rate": 3.656565656565657e-05,
      "loss": 1.0727,
      "step": 1880
    },
    {
      "epoch": 16.02018181818182,
      "eval_accuracy": 0.5343915343915344,
      "eval_loss": 1.001629114151001,
      "eval_runtime": 149.4378,
      "eval_samples_per_second": 5.059,
      "eval_steps_per_second": 0.636,
      "step": 1887
    },
    {
      "epoch": 17.000545454545456,
      "grad_norm": 9.216334342956543,
      "learning_rate": 3.6464646464646465e-05,
      "loss": 0.9356,
      "step": 1890
    },
    {
      "epoch": 17.002363636363636,
      "grad_norm": 12.995749473571777,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 1.0007,
      "step": 1900
    },
    {
      "epoch": 17.004181818181817,
      "grad_norm": 10.831470489501953,
      "learning_rate": 3.6262626262626264e-05,
      "loss": 1.0075,
      "step": 1910
    },
    {
      "epoch": 17.006,
      "grad_norm": 6.675658226013184,
      "learning_rate": 3.616161616161616e-05,
      "loss": 1.0298,
      "step": 1920
    },
    {
      "epoch": 17.00781818181818,
      "grad_norm": 8.875894546508789,
      "learning_rate": 3.606060606060606e-05,
      "loss": 0.8819,
      "step": 1930
    },
    {
      "epoch": 17.009636363636364,
      "grad_norm": 3.7380897998809814,
      "learning_rate": 3.595959595959596e-05,
      "loss": 1.0307,
      "step": 1940
    },
    {
      "epoch": 17.011454545454544,
      "grad_norm": 5.976683139801025,
      "learning_rate": 3.5858585858585855e-05,
      "loss": 0.9766,
      "step": 1950
    },
    {
      "epoch": 17.013272727272728,
      "grad_norm": 4.93250846862793,
      "learning_rate": 3.575757575757576e-05,
      "loss": 0.9317,
      "step": 1960
    },
    {
      "epoch": 17.015090909090908,
      "grad_norm": 4.194036960601807,
      "learning_rate": 3.565656565656566e-05,
      "loss": 0.9038,
      "step": 1970
    },
    {
      "epoch": 17.016909090909092,
      "grad_norm": 7.000123977661133,
      "learning_rate": 3.555555555555556e-05,
      "loss": 0.9665,
      "step": 1980
    },
    {
      "epoch": 17.018727272727272,
      "grad_norm": 9.221128463745117,
      "learning_rate": 3.545454545454546e-05,
      "loss": 0.9481,
      "step": 1990
    },
    {
      "epoch": 17.02018181818182,
      "eval_accuracy": 0.5264550264550265,
      "eval_loss": 0.9982506632804871,
      "eval_runtime": 149.0509,
      "eval_samples_per_second": 5.072,
      "eval_steps_per_second": 0.637,
      "step": 1998
    },
    {
      "epoch": 18.000363636363637,
      "grad_norm": 5.551551818847656,
      "learning_rate": 3.535353535353535e-05,
      "loss": 0.9912,
      "step": 2000
    },
    {
      "epoch": 18.002181818181818,
      "grad_norm": 8.715899467468262,
      "learning_rate": 3.525252525252525e-05,
      "loss": 1.0457,
      "step": 2010
    },
    {
      "epoch": 18.004,
      "grad_norm": 11.81405258178711,
      "learning_rate": 3.515151515151515e-05,
      "loss": 0.9465,
      "step": 2020
    },
    {
      "epoch": 18.00581818181818,
      "grad_norm": 5.644428730010986,
      "learning_rate": 3.505050505050505e-05,
      "loss": 1.0424,
      "step": 2030
    },
    {
      "epoch": 18.007636363636365,
      "grad_norm": 5.840436935424805,
      "learning_rate": 3.494949494949495e-05,
      "loss": 0.9281,
      "step": 2040
    },
    {
      "epoch": 18.009454545454545,
      "grad_norm": 6.476779460906982,
      "learning_rate": 3.484848484848485e-05,
      "loss": 0.9568,
      "step": 2050
    },
    {
      "epoch": 18.011272727272726,
      "grad_norm": 7.143520355224609,
      "learning_rate": 3.474747474747475e-05,
      "loss": 0.947,
      "step": 2060
    },
    {
      "epoch": 18.01309090909091,
      "grad_norm": 6.404818058013916,
      "learning_rate": 3.464646464646465e-05,
      "loss": 0.8925,
      "step": 2070
    },
    {
      "epoch": 18.01490909090909,
      "grad_norm": 8.545957565307617,
      "learning_rate": 3.454545454545455e-05,
      "loss": 0.9917,
      "step": 2080
    },
    {
      "epoch": 18.016727272727273,
      "grad_norm": 6.831506729125977,
      "learning_rate": 3.444444444444445e-05,
      "loss": 0.8973,
      "step": 2090
    },
    {
      "epoch": 18.018545454545453,
      "grad_norm": 7.68071985244751,
      "learning_rate": 3.434343434343435e-05,
      "loss": 0.9034,
      "step": 2100
    },
    {
      "epoch": 18.02018181818182,
      "eval_accuracy": 0.5013227513227513,
      "eval_loss": 1.0221006870269775,
      "eval_runtime": 144.419,
      "eval_samples_per_second": 5.235,
      "eval_steps_per_second": 0.658,
      "step": 2109
    },
    {
      "epoch": 19.00018181818182,
      "grad_norm": 7.987392902374268,
      "learning_rate": 3.424242424242424e-05,
      "loss": 1.057,
      "step": 2110
    },
    {
      "epoch": 19.002,
      "grad_norm": 5.622376918792725,
      "learning_rate": 3.414141414141414e-05,
      "loss": 0.9968,
      "step": 2120
    },
    {
      "epoch": 19.003818181818183,
      "grad_norm": 4.796281814575195,
      "learning_rate": 3.4040404040404045e-05,
      "loss": 0.9697,
      "step": 2130
    },
    {
      "epoch": 19.005636363636363,
      "grad_norm": 6.3387908935546875,
      "learning_rate": 3.3939393939393945e-05,
      "loss": 0.9431,
      "step": 2140
    },
    {
      "epoch": 19.007454545454546,
      "grad_norm": 8.202577590942383,
      "learning_rate": 3.3838383838383844e-05,
      "loss": 1.0198,
      "step": 2150
    },
    {
      "epoch": 19.009272727272727,
      "grad_norm": 12.490442276000977,
      "learning_rate": 3.373737373737374e-05,
      "loss": 0.9727,
      "step": 2160
    },
    {
      "epoch": 19.01109090909091,
      "grad_norm": 9.358773231506348,
      "learning_rate": 3.3636363636363636e-05,
      "loss": 0.9366,
      "step": 2170
    },
    {
      "epoch": 19.01290909090909,
      "grad_norm": 6.616375923156738,
      "learning_rate": 3.3535353535353536e-05,
      "loss": 1.0193,
      "step": 2180
    },
    {
      "epoch": 19.014727272727274,
      "grad_norm": 5.817307949066162,
      "learning_rate": 3.3434343434343435e-05,
      "loss": 1.0252,
      "step": 2190
    },
    {
      "epoch": 19.016545454545454,
      "grad_norm": 6.090243339538574,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.9156,
      "step": 2200
    },
    {
      "epoch": 19.018363636363638,
      "grad_norm": 7.476328372955322,
      "learning_rate": 3.3232323232323234e-05,
      "loss": 0.917,
      "step": 2210
    },
    {
      "epoch": 19.02018181818182,
      "grad_norm": 10.545258522033691,
      "learning_rate": 3.3131313131313134e-05,
      "loss": 0.8569,
      "step": 2220
    },
    {
      "epoch": 19.02018181818182,
      "eval_accuracy": 0.5264550264550265,
      "eval_loss": 0.9825001955032349,
      "eval_runtime": 151.5161,
      "eval_samples_per_second": 4.99,
      "eval_steps_per_second": 0.627,
      "step": 2220
    },
    {
      "epoch": 20.00181818181818,
      "grad_norm": 4.553400039672852,
      "learning_rate": 3.303030303030303e-05,
      "loss": 1.0637,
      "step": 2230
    },
    {
      "epoch": 20.003636363636364,
      "grad_norm": 4.245593070983887,
      "learning_rate": 3.292929292929293e-05,
      "loss": 0.954,
      "step": 2240
    },
    {
      "epoch": 20.005454545454544,
      "grad_norm": 6.2897210121154785,
      "learning_rate": 3.282828282828283e-05,
      "loss": 1.0215,
      "step": 2250
    },
    {
      "epoch": 20.007272727272728,
      "grad_norm": 8.956347465515137,
      "learning_rate": 3.272727272727273e-05,
      "loss": 0.8645,
      "step": 2260
    },
    {
      "epoch": 20.009090909090908,
      "grad_norm": 8.199272155761719,
      "learning_rate": 3.2626262626262624e-05,
      "loss": 0.9678,
      "step": 2270
    },
    {
      "epoch": 20.01090909090909,
      "grad_norm": 4.4115681648254395,
      "learning_rate": 3.2525252525252524e-05,
      "loss": 0.9508,
      "step": 2280
    },
    {
      "epoch": 20.012727272727272,
      "grad_norm": 5.1302714347839355,
      "learning_rate": 3.2424242424242423e-05,
      "loss": 0.8871,
      "step": 2290
    },
    {
      "epoch": 20.014545454545456,
      "grad_norm": 5.004090785980225,
      "learning_rate": 3.232323232323233e-05,
      "loss": 0.887,
      "step": 2300
    },
    {
      "epoch": 20.016363636363636,
      "grad_norm": 13.2914400100708,
      "learning_rate": 3.222222222222223e-05,
      "loss": 0.9051,
      "step": 2310
    },
    {
      "epoch": 20.01818181818182,
      "grad_norm": 9.987497329711914,
      "learning_rate": 3.212121212121212e-05,
      "loss": 0.9421,
      "step": 2320
    },
    {
      "epoch": 20.02,
      "grad_norm": 12.255473136901855,
      "learning_rate": 3.202020202020202e-05,
      "loss": 0.9256,
      "step": 2330
    },
    {
      "epoch": 20.02018181818182,
      "eval_accuracy": 0.5396825396825397,
      "eval_loss": 0.9677516222000122,
      "eval_runtime": 148.6642,
      "eval_samples_per_second": 5.085,
      "eval_steps_per_second": 0.639,
      "step": 2331
    },
    {
      "epoch": 21.001636363636365,
      "grad_norm": 5.143698692321777,
      "learning_rate": 3.191919191919192e-05,
      "loss": 0.9137,
      "step": 2340
    },
    {
      "epoch": 21.003454545454545,
      "grad_norm": 7.78878927230835,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.9939,
      "step": 2350
    },
    {
      "epoch": 21.00527272727273,
      "grad_norm": 6.338476657867432,
      "learning_rate": 3.171717171717172e-05,
      "loss": 0.8976,
      "step": 2360
    },
    {
      "epoch": 21.00709090909091,
      "grad_norm": 6.025676250457764,
      "learning_rate": 3.161616161616161e-05,
      "loss": 0.9612,
      "step": 2370
    },
    {
      "epoch": 21.00890909090909,
      "grad_norm": 6.251794338226318,
      "learning_rate": 3.151515151515151e-05,
      "loss": 0.9476,
      "step": 2380
    },
    {
      "epoch": 21.010727272727273,
      "grad_norm": 7.739345550537109,
      "learning_rate": 3.141414141414142e-05,
      "loss": 0.9162,
      "step": 2390
    },
    {
      "epoch": 21.012545454545453,
      "grad_norm": 6.820745944976807,
      "learning_rate": 3.131313131313132e-05,
      "loss": 0.8919,
      "step": 2400
    },
    {
      "epoch": 21.014363636363637,
      "grad_norm": 7.606575012207031,
      "learning_rate": 3.121212121212122e-05,
      "loss": 0.9543,
      "step": 2410
    },
    {
      "epoch": 21.016181818181817,
      "grad_norm": 5.678408145904541,
      "learning_rate": 3.111111111111111e-05,
      "loss": 0.9386,
      "step": 2420
    },
    {
      "epoch": 21.018,
      "grad_norm": 6.3955206871032715,
      "learning_rate": 3.101010101010101e-05,
      "loss": 0.9258,
      "step": 2430
    },
    {
      "epoch": 21.01981818181818,
      "grad_norm": 6.42524528503418,
      "learning_rate": 3.090909090909091e-05,
      "loss": 1.0311,
      "step": 2440
    },
    {
      "epoch": 21.02018181818182,
      "eval_accuracy": 0.5105820105820106,
      "eval_loss": 0.9574453234672546,
      "eval_runtime": 145.6583,
      "eval_samples_per_second": 5.19,
      "eval_steps_per_second": 0.652,
      "step": 2442
    },
    {
      "epoch": 22.001454545454546,
      "grad_norm": 5.330824375152588,
      "learning_rate": 3.080808080808081e-05,
      "loss": 0.9399,
      "step": 2450
    },
    {
      "epoch": 22.003272727272726,
      "grad_norm": 4.4678192138671875,
      "learning_rate": 3.070707070707071e-05,
      "loss": 0.7764,
      "step": 2460
    },
    {
      "epoch": 22.00509090909091,
      "grad_norm": 6.995047092437744,
      "learning_rate": 3.060606060606061e-05,
      "loss": 0.9542,
      "step": 2470
    },
    {
      "epoch": 22.00690909090909,
      "grad_norm": 9.586326599121094,
      "learning_rate": 3.050505050505051e-05,
      "loss": 1.0099,
      "step": 2480
    },
    {
      "epoch": 22.008727272727274,
      "grad_norm": 6.698061466217041,
      "learning_rate": 3.0404040404040406e-05,
      "loss": 0.9805,
      "step": 2490
    },
    {
      "epoch": 22.010545454545454,
      "grad_norm": 9.36473274230957,
      "learning_rate": 3.0303030303030306e-05,
      "loss": 0.8718,
      "step": 2500
    },
    {
      "epoch": 22.012363636363638,
      "grad_norm": 3.693310260772705,
      "learning_rate": 3.0202020202020205e-05,
      "loss": 0.9963,
      "step": 2510
    },
    {
      "epoch": 22.014181818181818,
      "grad_norm": 10.124568939208984,
      "learning_rate": 3.01010101010101e-05,
      "loss": 0.8962,
      "step": 2520
    },
    {
      "epoch": 22.016,
      "grad_norm": 5.137106895446777,
      "learning_rate": 3e-05,
      "loss": 0.9473,
      "step": 2530
    },
    {
      "epoch": 22.017818181818182,
      "grad_norm": 5.98214054107666,
      "learning_rate": 2.98989898989899e-05,
      "loss": 0.8896,
      "step": 2540
    },
    {
      "epoch": 22.019636363636362,
      "grad_norm": 10.731734275817871,
      "learning_rate": 2.9797979797979796e-05,
      "loss": 0.8651,
      "step": 2550
    },
    {
      "epoch": 22.02018181818182,
      "eval_accuracy": 0.49867724867724866,
      "eval_loss": 1.0047566890716553,
      "eval_runtime": 158.3545,
      "eval_samples_per_second": 4.774,
      "eval_steps_per_second": 0.6,
      "step": 2553
    },
    {
      "epoch": 23.001272727272728,
      "grad_norm": 5.062934398651123,
      "learning_rate": 2.96969696969697e-05,
      "loss": 0.9075,
      "step": 2560
    },
    {
      "epoch": 23.003090909090908,
      "grad_norm": 9.365435600280762,
      "learning_rate": 2.95959595959596e-05,
      "loss": 0.9693,
      "step": 2570
    },
    {
      "epoch": 23.00490909090909,
      "grad_norm": 8.738523483276367,
      "learning_rate": 2.9494949494949498e-05,
      "loss": 0.891,
      "step": 2580
    },
    {
      "epoch": 23.00672727272727,
      "grad_norm": 9.805744171142578,
      "learning_rate": 2.9393939393939394e-05,
      "loss": 0.8262,
      "step": 2590
    },
    {
      "epoch": 23.008545454545455,
      "grad_norm": 9.829437255859375,
      "learning_rate": 2.9292929292929294e-05,
      "loss": 0.9529,
      "step": 2600
    },
    {
      "epoch": 23.010363636363635,
      "grad_norm": 8.564093589782715,
      "learning_rate": 2.9191919191919193e-05,
      "loss": 0.9536,
      "step": 2610
    },
    {
      "epoch": 23.01218181818182,
      "grad_norm": 7.534347057342529,
      "learning_rate": 2.909090909090909e-05,
      "loss": 0.9625,
      "step": 2620
    },
    {
      "epoch": 23.014,
      "grad_norm": 5.0384039878845215,
      "learning_rate": 2.898989898989899e-05,
      "loss": 0.9672,
      "step": 2630
    },
    {
      "epoch": 23.015818181818183,
      "grad_norm": 3.9534647464752197,
      "learning_rate": 2.8888888888888888e-05,
      "loss": 1.0424,
      "step": 2640
    },
    {
      "epoch": 23.017636363636363,
      "grad_norm": 4.473760604858398,
      "learning_rate": 2.878787878787879e-05,
      "loss": 0.8712,
      "step": 2650
    },
    {
      "epoch": 23.019454545454547,
      "grad_norm": 7.2100348472595215,
      "learning_rate": 2.868686868686869e-05,
      "loss": 0.9384,
      "step": 2660
    },
    {
      "epoch": 23.02018181818182,
      "eval_accuracy": 0.5224867724867724,
      "eval_loss": 0.9716671109199524,
      "eval_runtime": 154.7299,
      "eval_samples_per_second": 4.886,
      "eval_steps_per_second": 0.614,
      "step": 2664
    },
    {
      "epoch": 24.00109090909091,
      "grad_norm": 13.06188678741455,
      "learning_rate": 2.8585858585858587e-05,
      "loss": 0.9131,
      "step": 2670
    },
    {
      "epoch": 24.002909090909093,
      "grad_norm": 13.319742202758789,
      "learning_rate": 2.8484848484848486e-05,
      "loss": 1.0963,
      "step": 2680
    },
    {
      "epoch": 24.004727272727273,
      "grad_norm": 8.775079727172852,
      "learning_rate": 2.8383838383838386e-05,
      "loss": 0.8914,
      "step": 2690
    },
    {
      "epoch": 24.006545454545453,
      "grad_norm": 7.000812530517578,
      "learning_rate": 2.8282828282828282e-05,
      "loss": 0.8983,
      "step": 2700
    },
    {
      "epoch": 24.008363636363637,
      "grad_norm": 5.142882823944092,
      "learning_rate": 2.818181818181818e-05,
      "loss": 0.9501,
      "step": 2710
    },
    {
      "epoch": 24.010181818181817,
      "grad_norm": 7.493001937866211,
      "learning_rate": 2.808080808080808e-05,
      "loss": 0.802,
      "step": 2720
    },
    {
      "epoch": 24.012,
      "grad_norm": 5.429752349853516,
      "learning_rate": 2.7979797979797984e-05,
      "loss": 0.8444,
      "step": 2730
    },
    {
      "epoch": 24.01381818181818,
      "grad_norm": 7.175288200378418,
      "learning_rate": 2.7878787878787883e-05,
      "loss": 0.9787,
      "step": 2740
    },
    {
      "epoch": 24.015636363636364,
      "grad_norm": 12.047754287719727,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.9811,
      "step": 2750
    },
    {
      "epoch": 24.017454545454545,
      "grad_norm": 10.042250633239746,
      "learning_rate": 2.767676767676768e-05,
      "loss": 0.9755,
      "step": 2760
    },
    {
      "epoch": 24.019272727272728,
      "grad_norm": 7.948033332824707,
      "learning_rate": 2.7575757575757578e-05,
      "loss": 0.9545,
      "step": 2770
    },
    {
      "epoch": 24.02018181818182,
      "eval_accuracy": 0.5171957671957672,
      "eval_loss": 0.9763391017913818,
      "eval_runtime": 147.5827,
      "eval_samples_per_second": 5.123,
      "eval_steps_per_second": 0.644,
      "step": 2775
    },
    {
      "epoch": 25.00090909090909,
      "grad_norm": 6.587329387664795,
      "learning_rate": 2.7474747474747474e-05,
      "loss": 1.0554,
      "step": 2780
    },
    {
      "epoch": 25.002727272727274,
      "grad_norm": 10.31185531616211,
      "learning_rate": 2.7373737373737374e-05,
      "loss": 0.8739,
      "step": 2790
    },
    {
      "epoch": 25.004545454545454,
      "grad_norm": 3.5984859466552734,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.8191,
      "step": 2800
    },
    {
      "epoch": 25.006363636363638,
      "grad_norm": 5.685831546783447,
      "learning_rate": 2.717171717171717e-05,
      "loss": 0.9804,
      "step": 2810
    },
    {
      "epoch": 25.008181818181818,
      "grad_norm": 8.268026351928711,
      "learning_rate": 2.7070707070707075e-05,
      "loss": 0.8067,
      "step": 2820
    },
    {
      "epoch": 25.01,
      "grad_norm": 5.9249773025512695,
      "learning_rate": 2.696969696969697e-05,
      "loss": 0.8634,
      "step": 2830
    },
    {
      "epoch": 25.01181818181818,
      "grad_norm": 5.94825553894043,
      "learning_rate": 2.686868686868687e-05,
      "loss": 0.8773,
      "step": 2840
    },
    {
      "epoch": 25.013636363636362,
      "grad_norm": 8.133403778076172,
      "learning_rate": 2.676767676767677e-05,
      "loss": 0.9,
      "step": 2850
    },
    {
      "epoch": 25.015454545454546,
      "grad_norm": 9.554986953735352,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 0.9591,
      "step": 2860
    },
    {
      "epoch": 25.017272727272726,
      "grad_norm": 10.977254867553711,
      "learning_rate": 2.6565656565656566e-05,
      "loss": 0.8604,
      "step": 2870
    },
    {
      "epoch": 25.01909090909091,
      "grad_norm": 7.258336067199707,
      "learning_rate": 2.6464646464646466e-05,
      "loss": 0.9187,
      "step": 2880
    },
    {
      "epoch": 25.02018181818182,
      "eval_accuracy": 0.5211640211640212,
      "eval_loss": 0.962836503982544,
      "eval_runtime": 146.3506,
      "eval_samples_per_second": 5.166,
      "eval_steps_per_second": 0.649,
      "step": 2886
    },
    {
      "epoch": 26.00072727272727,
      "grad_norm": 8.66103458404541,
      "learning_rate": 2.636363636363636e-05,
      "loss": 0.8217,
      "step": 2890
    },
    {
      "epoch": 26.002545454545455,
      "grad_norm": 8.29440689086914,
      "learning_rate": 2.6262626262626268e-05,
      "loss": 0.8624,
      "step": 2900
    },
    {
      "epoch": 26.004363636363635,
      "grad_norm": 8.019746780395508,
      "learning_rate": 2.6161616161616164e-05,
      "loss": 0.7612,
      "step": 2910
    },
    {
      "epoch": 26.00618181818182,
      "grad_norm": 13.38365364074707,
      "learning_rate": 2.6060606060606063e-05,
      "loss": 0.9354,
      "step": 2920
    },
    {
      "epoch": 26.008,
      "grad_norm": 9.972203254699707,
      "learning_rate": 2.5959595959595963e-05,
      "loss": 0.7413,
      "step": 2930
    },
    {
      "epoch": 26.009818181818183,
      "grad_norm": 5.007082462310791,
      "learning_rate": 2.585858585858586e-05,
      "loss": 0.9473,
      "step": 2940
    },
    {
      "epoch": 26.011636363636363,
      "grad_norm": 7.641317367553711,
      "learning_rate": 2.575757575757576e-05,
      "loss": 0.9514,
      "step": 2950
    },
    {
      "epoch": 26.013454545454547,
      "grad_norm": 6.921894073486328,
      "learning_rate": 2.5656565656565658e-05,
      "loss": 0.8694,
      "step": 2960
    },
    {
      "epoch": 26.015272727272727,
      "grad_norm": 14.923357963562012,
      "learning_rate": 2.5555555555555554e-05,
      "loss": 0.8904,
      "step": 2970
    },
    {
      "epoch": 26.01709090909091,
      "grad_norm": 9.429475784301758,
      "learning_rate": 2.5454545454545454e-05,
      "loss": 0.8831,
      "step": 2980
    },
    {
      "epoch": 26.01890909090909,
      "grad_norm": 6.4121527671813965,
      "learning_rate": 2.5353535353535356e-05,
      "loss": 0.7953,
      "step": 2990
    },
    {
      "epoch": 26.02018181818182,
      "eval_accuracy": 0.5264550264550265,
      "eval_loss": 0.9523358941078186,
      "eval_runtime": 150.2976,
      "eval_samples_per_second": 5.03,
      "eval_steps_per_second": 0.632,
      "step": 2997
    },
    {
      "epoch": 27.000545454545456,
      "grad_norm": 7.7547101974487305,
      "learning_rate": 2.5252525252525256e-05,
      "loss": 0.9809,
      "step": 3000
    },
    {
      "epoch": 27.002363636363636,
      "grad_norm": 9.476554870605469,
      "learning_rate": 2.5151515151515155e-05,
      "loss": 0.767,
      "step": 3010
    },
    {
      "epoch": 27.004181818181817,
      "grad_norm": 9.019051551818848,
      "learning_rate": 2.505050505050505e-05,
      "loss": 0.9555,
      "step": 3020
    },
    {
      "epoch": 27.006,
      "grad_norm": 7.9360833168029785,
      "learning_rate": 2.494949494949495e-05,
      "loss": 0.8657,
      "step": 3030
    },
    {
      "epoch": 27.00781818181818,
      "grad_norm": 7.8032612800598145,
      "learning_rate": 2.4848484848484847e-05,
      "loss": 0.7659,
      "step": 3040
    },
    {
      "epoch": 27.009636363636364,
      "grad_norm": 8.02381706237793,
      "learning_rate": 2.474747474747475e-05,
      "loss": 0.9237,
      "step": 3050
    },
    {
      "epoch": 27.011454545454544,
      "grad_norm": 6.776451587677002,
      "learning_rate": 2.464646464646465e-05,
      "loss": 1.0109,
      "step": 3060
    },
    {
      "epoch": 27.013272727272728,
      "grad_norm": 9.373581886291504,
      "learning_rate": 2.4545454545454545e-05,
      "loss": 0.8007,
      "step": 3070
    },
    {
      "epoch": 27.015090909090908,
      "grad_norm": 13.388670921325684,
      "learning_rate": 2.4444444444444445e-05,
      "loss": 0.7814,
      "step": 3080
    },
    {
      "epoch": 27.016909090909092,
      "grad_norm": 4.5681352615356445,
      "learning_rate": 2.4343434343434344e-05,
      "loss": 0.8697,
      "step": 3090
    },
    {
      "epoch": 27.018727272727272,
      "grad_norm": 3.7407443523406982,
      "learning_rate": 2.4242424242424244e-05,
      "loss": 0.8793,
      "step": 3100
    },
    {
      "epoch": 27.02018181818182,
      "eval_accuracy": 0.5370370370370371,
      "eval_loss": 0.9977483153343201,
      "eval_runtime": 147.1287,
      "eval_samples_per_second": 5.138,
      "eval_steps_per_second": 0.646,
      "step": 3108
    },
    {
      "epoch": 28.000363636363637,
      "grad_norm": 8.104622840881348,
      "learning_rate": 2.4141414141414143e-05,
      "loss": 0.8612,
      "step": 3110
    },
    {
      "epoch": 28.002181818181818,
      "grad_norm": 7.6057562828063965,
      "learning_rate": 2.404040404040404e-05,
      "loss": 0.9138,
      "step": 3120
    },
    {
      "epoch": 28.004,
      "grad_norm": 9.901479721069336,
      "learning_rate": 2.393939393939394e-05,
      "loss": 0.8845,
      "step": 3130
    },
    {
      "epoch": 28.00581818181818,
      "grad_norm": 10.897790908813477,
      "learning_rate": 2.3838383838383842e-05,
      "loss": 0.7653,
      "step": 3140
    },
    {
      "epoch": 28.007636363636365,
      "grad_norm": 6.1693267822265625,
      "learning_rate": 2.3737373737373738e-05,
      "loss": 0.8565,
      "step": 3150
    },
    {
      "epoch": 28.009454545454545,
      "grad_norm": 8.267416954040527,
      "learning_rate": 2.3636363636363637e-05,
      "loss": 0.8691,
      "step": 3160
    },
    {
      "epoch": 28.011272727272726,
      "grad_norm": 6.09727144241333,
      "learning_rate": 2.3535353535353537e-05,
      "loss": 0.8473,
      "step": 3170
    },
    {
      "epoch": 28.01309090909091,
      "grad_norm": 6.62100887298584,
      "learning_rate": 2.3434343434343436e-05,
      "loss": 0.7564,
      "step": 3180
    },
    {
      "epoch": 28.01490909090909,
      "grad_norm": 9.10604476928711,
      "learning_rate": 2.3333333333333336e-05,
      "loss": 0.7724,
      "step": 3190
    },
    {
      "epoch": 28.016727272727273,
      "grad_norm": 7.155385494232178,
      "learning_rate": 2.3232323232323232e-05,
      "loss": 0.823,
      "step": 3200
    },
    {
      "epoch": 28.018545454545453,
      "grad_norm": 9.220155715942383,
      "learning_rate": 2.313131313131313e-05,
      "loss": 0.7897,
      "step": 3210
    },
    {
      "epoch": 28.02018181818182,
      "eval_accuracy": 0.5317460317460317,
      "eval_loss": 0.9964798092842102,
      "eval_runtime": 147.9044,
      "eval_samples_per_second": 5.111,
      "eval_steps_per_second": 0.642,
      "step": 3219
    },
    {
      "epoch": 29.00018181818182,
      "grad_norm": 10.702564239501953,
      "learning_rate": 2.3030303030303034e-05,
      "loss": 0.893,
      "step": 3220
    },
    {
      "epoch": 29.002,
      "grad_norm": 11.652132034301758,
      "learning_rate": 2.292929292929293e-05,
      "loss": 0.9462,
      "step": 3230
    },
    {
      "epoch": 29.003818181818183,
      "grad_norm": 5.475378036499023,
      "learning_rate": 2.282828282828283e-05,
      "loss": 0.745,
      "step": 3240
    },
    {
      "epoch": 29.005636363636363,
      "grad_norm": 6.293100833892822,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.7957,
      "step": 3250
    },
    {
      "epoch": 29.007454545454546,
      "grad_norm": 9.700437545776367,
      "learning_rate": 2.262626262626263e-05,
      "loss": 0.9725,
      "step": 3260
    },
    {
      "epoch": 29.009272727272727,
      "grad_norm": 8.564444541931152,
      "learning_rate": 2.2525252525252528e-05,
      "loss": 0.9211,
      "step": 3270
    },
    {
      "epoch": 29.01109090909091,
      "grad_norm": 7.833463191986084,
      "learning_rate": 2.2424242424242424e-05,
      "loss": 0.926,
      "step": 3280
    },
    {
      "epoch": 29.01290909090909,
      "grad_norm": 9.064501762390137,
      "learning_rate": 2.2323232323232324e-05,
      "loss": 1.0597,
      "step": 3290
    },
    {
      "epoch": 29.014727272727274,
      "grad_norm": 9.065460205078125,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.8668,
      "step": 3300
    },
    {
      "epoch": 29.016545454545454,
      "grad_norm": 8.223165512084961,
      "learning_rate": 2.2121212121212123e-05,
      "loss": 0.878,
      "step": 3310
    },
    {
      "epoch": 29.018363636363638,
      "grad_norm": 7.36085319519043,
      "learning_rate": 2.2020202020202022e-05,
      "loss": 0.8666,
      "step": 3320
    },
    {
      "epoch": 29.02018181818182,
      "grad_norm": 13.20446491241455,
      "learning_rate": 2.191919191919192e-05,
      "loss": 0.8034,
      "step": 3330
    },
    {
      "epoch": 29.02018181818182,
      "eval_accuracy": 0.5462962962962963,
      "eval_loss": 0.9271609783172607,
      "eval_runtime": 150.9049,
      "eval_samples_per_second": 5.01,
      "eval_steps_per_second": 0.63,
      "step": 3330
    },
    {
      "epoch": 30.00181818181818,
      "grad_norm": 8.620079040527344,
      "learning_rate": 2.1818181818181818e-05,
      "loss": 0.8551,
      "step": 3340
    },
    {
      "epoch": 30.003636363636364,
      "grad_norm": 10.10771656036377,
      "learning_rate": 2.171717171717172e-05,
      "loss": 0.8917,
      "step": 3350
    },
    {
      "epoch": 30.005454545454544,
      "grad_norm": 8.590020179748535,
      "learning_rate": 2.1616161616161617e-05,
      "loss": 0.7187,
      "step": 3360
    },
    {
      "epoch": 30.007272727272728,
      "grad_norm": 10.648637771606445,
      "learning_rate": 2.1515151515151516e-05,
      "loss": 0.8967,
      "step": 3370
    },
    {
      "epoch": 30.009090909090908,
      "grad_norm": 6.126356601715088,
      "learning_rate": 2.1414141414141416e-05,
      "loss": 0.7979,
      "step": 3380
    },
    {
      "epoch": 30.01090909090909,
      "grad_norm": 6.431183815002441,
      "learning_rate": 2.1313131313131315e-05,
      "loss": 0.7723,
      "step": 3390
    },
    {
      "epoch": 30.012727272727272,
      "grad_norm": 7.046969413757324,
      "learning_rate": 2.1212121212121215e-05,
      "loss": 0.8287,
      "step": 3400
    },
    {
      "epoch": 30.014545454545456,
      "grad_norm": 10.070467948913574,
      "learning_rate": 2.111111111111111e-05,
      "loss": 0.9068,
      "step": 3410
    },
    {
      "epoch": 30.016363636363636,
      "grad_norm": 8.365399360656738,
      "learning_rate": 2.101010101010101e-05,
      "loss": 0.8802,
      "step": 3420
    },
    {
      "epoch": 30.01818181818182,
      "grad_norm": 6.84988260269165,
      "learning_rate": 2.090909090909091e-05,
      "loss": 0.9197,
      "step": 3430
    },
    {
      "epoch": 30.02,
      "grad_norm": 8.448094367980957,
      "learning_rate": 2.080808080808081e-05,
      "loss": 0.8469,
      "step": 3440
    },
    {
      "epoch": 30.02018181818182,
      "eval_accuracy": 0.5383597883597884,
      "eval_loss": 0.9230695962905884,
      "eval_runtime": 148.3769,
      "eval_samples_per_second": 5.095,
      "eval_steps_per_second": 0.64,
      "step": 3441
    },
    {
      "epoch": 31.001636363636365,
      "grad_norm": 9.007742881774902,
      "learning_rate": 2.070707070707071e-05,
      "loss": 0.9572,
      "step": 3450
    },
    {
      "epoch": 31.003454545454545,
      "grad_norm": 11.113166809082031,
      "learning_rate": 2.0606060606060608e-05,
      "loss": 0.8089,
      "step": 3460
    },
    {
      "epoch": 31.00527272727273,
      "grad_norm": 6.533670425415039,
      "learning_rate": 2.0505050505050504e-05,
      "loss": 0.9231,
      "step": 3470
    },
    {
      "epoch": 31.00709090909091,
      "grad_norm": 9.098650932312012,
      "learning_rate": 2.0404040404040407e-05,
      "loss": 0.868,
      "step": 3480
    },
    {
      "epoch": 31.00890909090909,
      "grad_norm": 8.53820514678955,
      "learning_rate": 2.0303030303030303e-05,
      "loss": 0.8957,
      "step": 3490
    },
    {
      "epoch": 31.010727272727273,
      "grad_norm": 6.1428422927856445,
      "learning_rate": 2.0202020202020203e-05,
      "loss": 0.7182,
      "step": 3500
    },
    {
      "epoch": 31.012545454545453,
      "grad_norm": 5.717990875244141,
      "learning_rate": 2.0101010101010102e-05,
      "loss": 0.829,
      "step": 3510
    },
    {
      "epoch": 31.014363636363637,
      "grad_norm": 7.748228073120117,
      "learning_rate": 2e-05,
      "loss": 0.8067,
      "step": 3520
    },
    {
      "epoch": 31.016181818181817,
      "grad_norm": 11.068036079406738,
      "learning_rate": 1.98989898989899e-05,
      "loss": 0.8186,
      "step": 3530
    },
    {
      "epoch": 31.018,
      "grad_norm": 8.458252906799316,
      "learning_rate": 1.9797979797979797e-05,
      "loss": 0.8751,
      "step": 3540
    },
    {
      "epoch": 31.01981818181818,
      "grad_norm": 6.225341796875,
      "learning_rate": 1.9696969696969697e-05,
      "loss": 0.79,
      "step": 3550
    },
    {
      "epoch": 31.02018181818182,
      "eval_accuracy": 0.5727513227513228,
      "eval_loss": 0.9281449913978577,
      "eval_runtime": 150.2805,
      "eval_samples_per_second": 5.031,
      "eval_steps_per_second": 0.632,
      "step": 3552
    },
    {
      "epoch": 32.00145454545454,
      "grad_norm": 10.177803993225098,
      "learning_rate": 1.95959595959596e-05,
      "loss": 0.9101,
      "step": 3560
    },
    {
      "epoch": 32.00327272727273,
      "grad_norm": 10.928609848022461,
      "learning_rate": 1.9494949494949496e-05,
      "loss": 0.8019,
      "step": 3570
    },
    {
      "epoch": 32.00509090909091,
      "grad_norm": 8.741167068481445,
      "learning_rate": 1.9393939393939395e-05,
      "loss": 0.9915,
      "step": 3580
    },
    {
      "epoch": 32.00690909090909,
      "grad_norm": 4.667022228240967,
      "learning_rate": 1.9292929292929295e-05,
      "loss": 0.7823,
      "step": 3590
    },
    {
      "epoch": 32.00872727272727,
      "grad_norm": 8.407429695129395,
      "learning_rate": 1.919191919191919e-05,
      "loss": 0.9603,
      "step": 3600
    },
    {
      "epoch": 32.01054545454546,
      "grad_norm": 8.828394889831543,
      "learning_rate": 1.9090909090909094e-05,
      "loss": 0.8425,
      "step": 3610
    },
    {
      "epoch": 32.01236363636364,
      "grad_norm": 7.08673620223999,
      "learning_rate": 1.898989898989899e-05,
      "loss": 0.9535,
      "step": 3620
    },
    {
      "epoch": 32.01418181818182,
      "grad_norm": 5.232740879058838,
      "learning_rate": 1.888888888888889e-05,
      "loss": 0.6998,
      "step": 3630
    },
    {
      "epoch": 32.016,
      "grad_norm": 8.102818489074707,
      "learning_rate": 1.878787878787879e-05,
      "loss": 0.8431,
      "step": 3640
    },
    {
      "epoch": 32.01781818181818,
      "grad_norm": 9.122153282165527,
      "learning_rate": 1.8686868686868688e-05,
      "loss": 0.755,
      "step": 3650
    },
    {
      "epoch": 32.019636363636366,
      "grad_norm": 14.438915252685547,
      "learning_rate": 1.8585858585858588e-05,
      "loss": 0.8516,
      "step": 3660
    },
    {
      "epoch": 32.02018181818182,
      "eval_accuracy": 0.5568783068783069,
      "eval_loss": 0.9310464262962341,
      "eval_runtime": 149.3465,
      "eval_samples_per_second": 5.062,
      "eval_steps_per_second": 0.636,
      "step": 3663
    },
    {
      "epoch": 33.00127272727273,
      "grad_norm": 8.778589248657227,
      "learning_rate": 1.8484848484848487e-05,
      "loss": 0.7724,
      "step": 3670
    },
    {
      "epoch": 33.00309090909091,
      "grad_norm": 26.57788848876953,
      "learning_rate": 1.8383838383838383e-05,
      "loss": 0.9285,
      "step": 3680
    },
    {
      "epoch": 33.00490909090909,
      "grad_norm": 7.8436689376831055,
      "learning_rate": 1.8282828282828286e-05,
      "loss": 0.7533,
      "step": 3690
    },
    {
      "epoch": 33.006727272727275,
      "grad_norm": 8.450164794921875,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.7167,
      "step": 3700
    },
    {
      "epoch": 33.008545454545455,
      "grad_norm": 6.676935195922852,
      "learning_rate": 1.808080808080808e-05,
      "loss": 0.8161,
      "step": 3710
    },
    {
      "epoch": 33.010363636363635,
      "grad_norm": 9.457823753356934,
      "learning_rate": 1.797979797979798e-05,
      "loss": 0.7213,
      "step": 3720
    },
    {
      "epoch": 33.012181818181816,
      "grad_norm": 6.458945274353027,
      "learning_rate": 1.787878787878788e-05,
      "loss": 0.791,
      "step": 3730
    },
    {
      "epoch": 33.014,
      "grad_norm": 10.989433288574219,
      "learning_rate": 1.777777777777778e-05,
      "loss": 0.9043,
      "step": 3740
    },
    {
      "epoch": 33.01581818181818,
      "grad_norm": 9.787355422973633,
      "learning_rate": 1.7676767676767676e-05,
      "loss": 0.8733,
      "step": 3750
    },
    {
      "epoch": 33.01763636363636,
      "grad_norm": 13.30040168762207,
      "learning_rate": 1.7575757575757576e-05,
      "loss": 0.7123,
      "step": 3760
    },
    {
      "epoch": 33.01945454545454,
      "grad_norm": 11.642248153686523,
      "learning_rate": 1.7474747474747475e-05,
      "loss": 0.8138,
      "step": 3770
    },
    {
      "epoch": 33.02018181818182,
      "eval_accuracy": 0.5674603174603174,
      "eval_loss": 0.9582036733627319,
      "eval_runtime": 147.1713,
      "eval_samples_per_second": 5.137,
      "eval_steps_per_second": 0.646,
      "step": 3774
    },
    {
      "epoch": 34.00109090909091,
      "grad_norm": 9.875617027282715,
      "learning_rate": 1.7373737373737375e-05,
      "loss": 0.7208,
      "step": 3780
    },
    {
      "epoch": 34.00290909090909,
      "grad_norm": 19.283206939697266,
      "learning_rate": 1.7272727272727274e-05,
      "loss": 0.7816,
      "step": 3790
    },
    {
      "epoch": 34.00472727272727,
      "grad_norm": 14.294538497924805,
      "learning_rate": 1.7171717171717173e-05,
      "loss": 0.9556,
      "step": 3800
    },
    {
      "epoch": 34.00654545454545,
      "grad_norm": 13.240488052368164,
      "learning_rate": 1.707070707070707e-05,
      "loss": 0.9312,
      "step": 3810
    },
    {
      "epoch": 34.00836363636363,
      "grad_norm": 11.781719207763672,
      "learning_rate": 1.6969696969696972e-05,
      "loss": 0.8563,
      "step": 3820
    },
    {
      "epoch": 34.01018181818182,
      "grad_norm": 6.2492218017578125,
      "learning_rate": 1.686868686868687e-05,
      "loss": 0.6618,
      "step": 3830
    },
    {
      "epoch": 34.012,
      "grad_norm": 11.331971168518066,
      "learning_rate": 1.6767676767676768e-05,
      "loss": 0.7962,
      "step": 3840
    },
    {
      "epoch": 34.01381818181818,
      "grad_norm": 8.892802238464355,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.858,
      "step": 3850
    },
    {
      "epoch": 34.01563636363636,
      "grad_norm": 9.716336250305176,
      "learning_rate": 1.6565656565656567e-05,
      "loss": 0.7353,
      "step": 3860
    },
    {
      "epoch": 34.01745454545455,
      "grad_norm": 14.95606517791748,
      "learning_rate": 1.6464646464646466e-05,
      "loss": 0.6582,
      "step": 3870
    },
    {
      "epoch": 34.01927272727273,
      "grad_norm": 12.510953903198242,
      "learning_rate": 1.6363636363636366e-05,
      "loss": 0.8322,
      "step": 3880
    },
    {
      "epoch": 34.02018181818182,
      "eval_accuracy": 0.5621693121693122,
      "eval_loss": 0.9740582704544067,
      "eval_runtime": 146.7236,
      "eval_samples_per_second": 5.153,
      "eval_steps_per_second": 0.647,
      "step": 3885
    },
    {
      "epoch": 35.00090909090909,
      "grad_norm": 10.233559608459473,
      "learning_rate": 1.6262626262626262e-05,
      "loss": 0.7694,
      "step": 3890
    },
    {
      "epoch": 35.00272727272727,
      "grad_norm": 12.976431846618652,
      "learning_rate": 1.6161616161616165e-05,
      "loss": 0.7612,
      "step": 3900
    },
    {
      "epoch": 35.00454545454546,
      "grad_norm": 11.562703132629395,
      "learning_rate": 1.606060606060606e-05,
      "loss": 0.8862,
      "step": 3910
    },
    {
      "epoch": 35.00636363636364,
      "grad_norm": 9.613603591918945,
      "learning_rate": 1.595959595959596e-05,
      "loss": 0.6993,
      "step": 3920
    },
    {
      "epoch": 35.00818181818182,
      "grad_norm": 12.087981224060059,
      "learning_rate": 1.585858585858586e-05,
      "loss": 0.9855,
      "step": 3930
    },
    {
      "epoch": 35.01,
      "grad_norm": 7.05510139465332,
      "learning_rate": 1.5757575757575756e-05,
      "loss": 0.8415,
      "step": 3940
    },
    {
      "epoch": 35.011818181818185,
      "grad_norm": 8.785040855407715,
      "learning_rate": 1.565656565656566e-05,
      "loss": 0.773,
      "step": 3950
    },
    {
      "epoch": 35.013636363636365,
      "grad_norm": 10.331352233886719,
      "learning_rate": 1.5555555555555555e-05,
      "loss": 0.7655,
      "step": 3960
    },
    {
      "epoch": 35.015454545454546,
      "grad_norm": 5.654623031616211,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 0.7274,
      "step": 3970
    },
    {
      "epoch": 35.017272727272726,
      "grad_norm": 12.63278579711914,
      "learning_rate": 1.5353535353535354e-05,
      "loss": 0.9074,
      "step": 3980
    },
    {
      "epoch": 35.019090909090906,
      "grad_norm": 7.924473285675049,
      "learning_rate": 1.5252525252525255e-05,
      "loss": 0.8064,
      "step": 3990
    },
    {
      "epoch": 35.02018181818182,
      "eval_accuracy": 0.5753968253968254,
      "eval_loss": 0.9572974443435669,
      "eval_runtime": 150.5167,
      "eval_samples_per_second": 5.023,
      "eval_steps_per_second": 0.631,
      "step": 3996
    },
    {
      "epoch": 36.000727272727275,
      "grad_norm": 11.757354736328125,
      "learning_rate": 1.5151515151515153e-05,
      "loss": 0.8435,
      "step": 4000
    },
    {
      "epoch": 36.002545454545455,
      "grad_norm": 13.549497604370117,
      "learning_rate": 1.505050505050505e-05,
      "loss": 0.7881,
      "step": 4010
    },
    {
      "epoch": 36.004363636363635,
      "grad_norm": 11.874682426452637,
      "learning_rate": 1.494949494949495e-05,
      "loss": 0.8561,
      "step": 4020
    },
    {
      "epoch": 36.006181818181815,
      "grad_norm": 11.986287117004395,
      "learning_rate": 1.484848484848485e-05,
      "loss": 0.8678,
      "step": 4030
    },
    {
      "epoch": 36.008,
      "grad_norm": 9.599827766418457,
      "learning_rate": 1.4747474747474749e-05,
      "loss": 0.6111,
      "step": 4040
    },
    {
      "epoch": 36.00981818181818,
      "grad_norm": 12.461723327636719,
      "learning_rate": 1.4646464646464647e-05,
      "loss": 0.8555,
      "step": 4050
    },
    {
      "epoch": 36.01163636363636,
      "grad_norm": 10.647710800170898,
      "learning_rate": 1.4545454545454545e-05,
      "loss": 0.8921,
      "step": 4060
    },
    {
      "epoch": 36.01345454545454,
      "grad_norm": 7.4742350578308105,
      "learning_rate": 1.4444444444444444e-05,
      "loss": 0.7072,
      "step": 4070
    },
    {
      "epoch": 36.01527272727273,
      "grad_norm": 10.952604293823242,
      "learning_rate": 1.4343434343434345e-05,
      "loss": 0.8367,
      "step": 4080
    },
    {
      "epoch": 36.01709090909091,
      "grad_norm": 13.392800331115723,
      "learning_rate": 1.4242424242424243e-05,
      "loss": 0.762,
      "step": 4090
    },
    {
      "epoch": 36.01890909090909,
      "grad_norm": 8.420698165893555,
      "learning_rate": 1.4141414141414141e-05,
      "loss": 0.8767,
      "step": 4100
    },
    {
      "epoch": 36.02018181818182,
      "eval_accuracy": 0.5714285714285714,
      "eval_loss": 0.9290463924407959,
      "eval_runtime": 148.5027,
      "eval_samples_per_second": 5.091,
      "eval_steps_per_second": 0.64,
      "step": 4107
    },
    {
      "epoch": 37.00054545454545,
      "grad_norm": 14.011470794677734,
      "learning_rate": 1.404040404040404e-05,
      "loss": 0.8077,
      "step": 4110
    },
    {
      "epoch": 37.00236363636364,
      "grad_norm": 5.852056980133057,
      "learning_rate": 1.3939393939393942e-05,
      "loss": 0.8372,
      "step": 4120
    },
    {
      "epoch": 37.00418181818182,
      "grad_norm": 14.729412078857422,
      "learning_rate": 1.383838383838384e-05,
      "loss": 0.7565,
      "step": 4130
    },
    {
      "epoch": 37.006,
      "grad_norm": 15.823266983032227,
      "learning_rate": 1.3737373737373737e-05,
      "loss": 0.7657,
      "step": 4140
    },
    {
      "epoch": 37.00781818181818,
      "grad_norm": 8.077898025512695,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.8377,
      "step": 4150
    },
    {
      "epoch": 37.00963636363636,
      "grad_norm": 19.177709579467773,
      "learning_rate": 1.3535353535353538e-05,
      "loss": 0.9433,
      "step": 4160
    },
    {
      "epoch": 37.01145454545455,
      "grad_norm": 10.09786605834961,
      "learning_rate": 1.3434343434343436e-05,
      "loss": 0.7564,
      "step": 4170
    },
    {
      "epoch": 37.01327272727273,
      "grad_norm": 7.137022972106934,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.8067,
      "step": 4180
    },
    {
      "epoch": 37.01509090909091,
      "grad_norm": 6.618945121765137,
      "learning_rate": 1.3232323232323233e-05,
      "loss": 0.803,
      "step": 4190
    },
    {
      "epoch": 37.01690909090909,
      "grad_norm": 15.50954532623291,
      "learning_rate": 1.3131313131313134e-05,
      "loss": 0.8677,
      "step": 4200
    },
    {
      "epoch": 37.018727272727276,
      "grad_norm": 7.7649030685424805,
      "learning_rate": 1.3030303030303032e-05,
      "loss": 0.7978,
      "step": 4210
    },
    {
      "epoch": 37.02018181818182,
      "eval_accuracy": 0.5727513227513228,
      "eval_loss": 0.9448675513267517,
      "eval_runtime": 147.0248,
      "eval_samples_per_second": 5.142,
      "eval_steps_per_second": 0.646,
      "step": 4218
    },
    {
      "epoch": 38.00036363636364,
      "grad_norm": 9.288775444030762,
      "learning_rate": 1.292929292929293e-05,
      "loss": 0.6643,
      "step": 4220
    },
    {
      "epoch": 38.00218181818182,
      "grad_norm": 14.049256324768066,
      "learning_rate": 1.2828282828282829e-05,
      "loss": 0.8048,
      "step": 4230
    },
    {
      "epoch": 38.004,
      "grad_norm": 6.196603775024414,
      "learning_rate": 1.2727272727272727e-05,
      "loss": 0.6936,
      "step": 4240
    },
    {
      "epoch": 38.005818181818185,
      "grad_norm": 16.842363357543945,
      "learning_rate": 1.2626262626262628e-05,
      "loss": 0.9143,
      "step": 4250
    },
    {
      "epoch": 38.007636363636365,
      "grad_norm": 10.34643840789795,
      "learning_rate": 1.2525252525252526e-05,
      "loss": 0.7569,
      "step": 4260
    },
    {
      "epoch": 38.009454545454545,
      "grad_norm": 5.936524868011475,
      "learning_rate": 1.2424242424242424e-05,
      "loss": 0.7721,
      "step": 4270
    },
    {
      "epoch": 38.011272727272726,
      "grad_norm": 8.305501937866211,
      "learning_rate": 1.2323232323232325e-05,
      "loss": 0.7388,
      "step": 4280
    },
    {
      "epoch": 38.013090909090906,
      "grad_norm": 10.99225902557373,
      "learning_rate": 1.2222222222222222e-05,
      "loss": 0.6871,
      "step": 4290
    },
    {
      "epoch": 38.01490909090909,
      "grad_norm": 11.384906768798828,
      "learning_rate": 1.2121212121212122e-05,
      "loss": 0.8106,
      "step": 4300
    },
    {
      "epoch": 38.01672727272727,
      "grad_norm": 10.89456844329834,
      "learning_rate": 1.202020202020202e-05,
      "loss": 0.7482,
      "step": 4310
    },
    {
      "epoch": 38.01854545454545,
      "grad_norm": 7.169947147369385,
      "learning_rate": 1.1919191919191921e-05,
      "loss": 0.8113,
      "step": 4320
    },
    {
      "epoch": 38.02018181818182,
      "eval_accuracy": 0.578042328042328,
      "eval_loss": 0.949256956577301,
      "eval_runtime": 148.4198,
      "eval_samples_per_second": 5.094,
      "eval_steps_per_second": 0.64,
      "step": 4329
    },
    {
      "epoch": 39.000181818181815,
      "grad_norm": 6.558440208435059,
      "learning_rate": 1.1818181818181819e-05,
      "loss": 0.6821,
      "step": 4330
    },
    {
      "epoch": 39.002,
      "grad_norm": 9.286921501159668,
      "learning_rate": 1.1717171717171718e-05,
      "loss": 0.868,
      "step": 4340
    },
    {
      "epoch": 39.00381818181818,
      "grad_norm": 11.597467422485352,
      "learning_rate": 1.1616161616161616e-05,
      "loss": 0.6401,
      "step": 4350
    },
    {
      "epoch": 39.00563636363636,
      "grad_norm": 16.700910568237305,
      "learning_rate": 1.1515151515151517e-05,
      "loss": 0.7055,
      "step": 4360
    },
    {
      "epoch": 39.00745454545454,
      "grad_norm": 11.440078735351562,
      "learning_rate": 1.1414141414141415e-05,
      "loss": 0.7684,
      "step": 4370
    },
    {
      "epoch": 39.00927272727273,
      "grad_norm": 9.321911811828613,
      "learning_rate": 1.1313131313131314e-05,
      "loss": 0.7973,
      "step": 4380
    },
    {
      "epoch": 39.01109090909091,
      "grad_norm": 8.111784934997559,
      "learning_rate": 1.1212121212121212e-05,
      "loss": 0.7334,
      "step": 4390
    },
    {
      "epoch": 39.01290909090909,
      "grad_norm": 7.027799606323242,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 0.6968,
      "step": 4400
    },
    {
      "epoch": 39.01472727272727,
      "grad_norm": 11.222505569458008,
      "learning_rate": 1.1010101010101011e-05,
      "loss": 0.806,
      "step": 4410
    },
    {
      "epoch": 39.01654545454546,
      "grad_norm": 13.890769958496094,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 0.8326,
      "step": 4420
    },
    {
      "epoch": 39.01836363636364,
      "grad_norm": 14.440450668334961,
      "learning_rate": 1.0808080808080808e-05,
      "loss": 0.7795,
      "step": 4430
    },
    {
      "epoch": 39.02018181818182,
      "grad_norm": 7.134138584136963,
      "learning_rate": 1.0707070707070708e-05,
      "loss": 0.8065,
      "step": 4440
    },
    {
      "epoch": 39.02018181818182,
      "eval_accuracy": 0.5925925925925926,
      "eval_loss": 0.9015074372291565,
      "eval_runtime": 145.6584,
      "eval_samples_per_second": 5.19,
      "eval_steps_per_second": 0.652,
      "step": 4440
    },
    {
      "epoch": 40.00181818181818,
      "grad_norm": 6.643571376800537,
      "learning_rate": 1.0606060606060607e-05,
      "loss": 0.7311,
      "step": 4450
    },
    {
      "epoch": 40.00363636363636,
      "grad_norm": 8.529888153076172,
      "learning_rate": 1.0505050505050505e-05,
      "loss": 0.7338,
      "step": 4460
    },
    {
      "epoch": 40.00545454545455,
      "grad_norm": 6.874202728271484,
      "learning_rate": 1.0404040404040405e-05,
      "loss": 0.7362,
      "step": 4470
    },
    {
      "epoch": 40.00727272727273,
      "grad_norm": 12.398289680480957,
      "learning_rate": 1.0303030303030304e-05,
      "loss": 0.9457,
      "step": 4480
    },
    {
      "epoch": 40.00909090909091,
      "grad_norm": 9.524188041687012,
      "learning_rate": 1.0202020202020204e-05,
      "loss": 0.6851,
      "step": 4490
    },
    {
      "epoch": 40.01090909090909,
      "grad_norm": 8.724706649780273,
      "learning_rate": 1.0101010101010101e-05,
      "loss": 0.8305,
      "step": 4500
    },
    {
      "epoch": 40.012727272727275,
      "grad_norm": 7.507053852081299,
      "learning_rate": 1e-05,
      "loss": 0.8247,
      "step": 4510
    },
    {
      "epoch": 40.014545454545456,
      "grad_norm": 17.32568359375,
      "learning_rate": 9.898989898989899e-06,
      "loss": 0.7536,
      "step": 4520
    },
    {
      "epoch": 40.016363636363636,
      "grad_norm": 9.178523063659668,
      "learning_rate": 9.7979797979798e-06,
      "loss": 0.8062,
      "step": 4530
    },
    {
      "epoch": 40.018181818181816,
      "grad_norm": 9.23918342590332,
      "learning_rate": 9.696969696969698e-06,
      "loss": 0.7517,
      "step": 4540
    },
    {
      "epoch": 40.02,
      "grad_norm": 5.558004379272461,
      "learning_rate": 9.595959595959595e-06,
      "loss": 0.7989,
      "step": 4550
    },
    {
      "epoch": 40.02018181818182,
      "eval_accuracy": 0.5886243386243386,
      "eval_loss": 0.9139331579208374,
      "eval_runtime": 149.2134,
      "eval_samples_per_second": 5.067,
      "eval_steps_per_second": 0.637,
      "step": 4551
    },
    {
      "epoch": 41.001636363636365,
      "grad_norm": 10.740436553955078,
      "learning_rate": 9.494949494949495e-06,
      "loss": 0.8113,
      "step": 4560
    },
    {
      "epoch": 41.003454545454545,
      "grad_norm": 12.195236206054688,
      "learning_rate": 9.393939393939394e-06,
      "loss": 0.8884,
      "step": 4570
    },
    {
      "epoch": 41.005272727272725,
      "grad_norm": 9.886882781982422,
      "learning_rate": 9.292929292929294e-06,
      "loss": 0.6887,
      "step": 4580
    },
    {
      "epoch": 41.00709090909091,
      "grad_norm": 14.517189025878906,
      "learning_rate": 9.191919191919192e-06,
      "loss": 0.7739,
      "step": 4590
    },
    {
      "epoch": 41.00890909090909,
      "grad_norm": 11.440573692321777,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.6939,
      "step": 4600
    },
    {
      "epoch": 41.01072727272727,
      "grad_norm": 8.60914421081543,
      "learning_rate": 8.98989898989899e-06,
      "loss": 0.6769,
      "step": 4610
    },
    {
      "epoch": 41.01254545454545,
      "grad_norm": 5.601386070251465,
      "learning_rate": 8.88888888888889e-06,
      "loss": 0.7519,
      "step": 4620
    },
    {
      "epoch": 41.01436363636363,
      "grad_norm": 7.747749328613281,
      "learning_rate": 8.787878787878788e-06,
      "loss": 0.7842,
      "step": 4630
    },
    {
      "epoch": 41.01618181818182,
      "grad_norm": 11.570269584655762,
      "learning_rate": 8.686868686868687e-06,
      "loss": 0.7867,
      "step": 4640
    },
    {
      "epoch": 41.018,
      "grad_norm": 15.08242416381836,
      "learning_rate": 8.585858585858587e-06,
      "loss": 0.7528,
      "step": 4650
    },
    {
      "epoch": 41.01981818181818,
      "grad_norm": 7.267955780029297,
      "learning_rate": 8.484848484848486e-06,
      "loss": 0.6323,
      "step": 4660
    },
    {
      "epoch": 41.02018181818182,
      "eval_accuracy": 0.5992063492063492,
      "eval_loss": 0.9004061818122864,
      "eval_runtime": 148.3224,
      "eval_samples_per_second": 5.097,
      "eval_steps_per_second": 0.64,
      "step": 4662
    },
    {
      "epoch": 42.00145454545454,
      "grad_norm": 9.609107971191406,
      "learning_rate": 8.383838383838384e-06,
      "loss": 0.7849,
      "step": 4670
    },
    {
      "epoch": 42.00327272727273,
      "grad_norm": 12.100410461425781,
      "learning_rate": 8.282828282828283e-06,
      "loss": 0.806,
      "step": 4680
    },
    {
      "epoch": 42.00509090909091,
      "grad_norm": 10.153985977172852,
      "learning_rate": 8.181818181818183e-06,
      "loss": 0.669,
      "step": 4690
    },
    {
      "epoch": 42.00690909090909,
      "grad_norm": 14.402464866638184,
      "learning_rate": 8.080808080808082e-06,
      "loss": 0.6636,
      "step": 4700
    },
    {
      "epoch": 42.00872727272727,
      "grad_norm": 14.053479194641113,
      "learning_rate": 7.97979797979798e-06,
      "loss": 0.769,
      "step": 4710
    },
    {
      "epoch": 42.01054545454546,
      "grad_norm": 7.945107460021973,
      "learning_rate": 7.878787878787878e-06,
      "loss": 0.6945,
      "step": 4720
    },
    {
      "epoch": 42.01236363636364,
      "grad_norm": 8.050100326538086,
      "learning_rate": 7.777777777777777e-06,
      "loss": 0.6112,
      "step": 4730
    },
    {
      "epoch": 42.01418181818182,
      "grad_norm": 13.17475414276123,
      "learning_rate": 7.676767676767677e-06,
      "loss": 0.8035,
      "step": 4740
    },
    {
      "epoch": 42.016,
      "grad_norm": 16.584096908569336,
      "learning_rate": 7.5757575757575764e-06,
      "loss": 0.7637,
      "step": 4750
    },
    {
      "epoch": 42.01781818181818,
      "grad_norm": 11.471580505371094,
      "learning_rate": 7.474747474747475e-06,
      "loss": 0.7247,
      "step": 4760
    },
    {
      "epoch": 42.019636363636366,
      "grad_norm": 11.29985237121582,
      "learning_rate": 7.3737373737373745e-06,
      "loss": 0.6847,
      "step": 4770
    },
    {
      "epoch": 42.02018181818182,
      "eval_accuracy": 0.6124338624338624,
      "eval_loss": 0.9082908034324646,
      "eval_runtime": 147.6625,
      "eval_samples_per_second": 5.12,
      "eval_steps_per_second": 0.643,
      "step": 4773
    },
    {
      "epoch": 43.00127272727273,
      "grad_norm": 10.324508666992188,
      "learning_rate": 7.272727272727272e-06,
      "loss": 0.5586,
      "step": 4780
    },
    {
      "epoch": 43.00309090909091,
      "grad_norm": 10.838313102722168,
      "learning_rate": 7.171717171717173e-06,
      "loss": 0.7385,
      "step": 4790
    },
    {
      "epoch": 43.00490909090909,
      "grad_norm": 6.829562187194824,
      "learning_rate": 7.0707070707070704e-06,
      "loss": 0.6887,
      "step": 4800
    },
    {
      "epoch": 43.006727272727275,
      "grad_norm": 19.654333114624023,
      "learning_rate": 6.969696969696971e-06,
      "loss": 0.8211,
      "step": 4810
    },
    {
      "epoch": 43.008545454545455,
      "grad_norm": 11.248620986938477,
      "learning_rate": 6.8686868686868685e-06,
      "loss": 0.7529,
      "step": 4820
    },
    {
      "epoch": 43.010363636363635,
      "grad_norm": 14.996559143066406,
      "learning_rate": 6.767676767676769e-06,
      "loss": 0.764,
      "step": 4830
    },
    {
      "epoch": 43.012181818181816,
      "grad_norm": 13.435885429382324,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.7792,
      "step": 4840
    },
    {
      "epoch": 43.014,
      "grad_norm": 16.040420532226562,
      "learning_rate": 6.565656565656567e-06,
      "loss": 0.9152,
      "step": 4850
    },
    {
      "epoch": 43.01581818181818,
      "grad_norm": 9.068537712097168,
      "learning_rate": 6.464646464646465e-06,
      "loss": 0.7527,
      "step": 4860
    },
    {
      "epoch": 43.01763636363636,
      "grad_norm": 9.895467758178711,
      "learning_rate": 6.363636363636363e-06,
      "loss": 0.7526,
      "step": 4870
    },
    {
      "epoch": 43.01945454545454,
      "grad_norm": 12.693032264709473,
      "learning_rate": 6.262626262626263e-06,
      "loss": 0.7711,
      "step": 4880
    },
    {
      "epoch": 43.02018181818182,
      "eval_accuracy": 0.5978835978835979,
      "eval_loss": 0.9023100137710571,
      "eval_runtime": 146.9949,
      "eval_samples_per_second": 5.143,
      "eval_steps_per_second": 0.646,
      "step": 4884
    },
    {
      "epoch": 44.00109090909091,
      "grad_norm": 9.122892379760742,
      "learning_rate": 6.161616161616162e-06,
      "loss": 0.8924,
      "step": 4890
    },
    {
      "epoch": 44.00290909090909,
      "grad_norm": 16.091691970825195,
      "learning_rate": 6.060606060606061e-06,
      "loss": 0.8003,
      "step": 4900
    },
    {
      "epoch": 44.00472727272727,
      "grad_norm": 11.019819259643555,
      "learning_rate": 5.9595959595959605e-06,
      "loss": 0.7017,
      "step": 4910
    },
    {
      "epoch": 44.00654545454545,
      "grad_norm": 5.1934332847595215,
      "learning_rate": 5.858585858585859e-06,
      "loss": 0.5691,
      "step": 4920
    },
    {
      "epoch": 44.00836363636363,
      "grad_norm": 15.653286933898926,
      "learning_rate": 5.7575757575757586e-06,
      "loss": 0.7065,
      "step": 4930
    },
    {
      "epoch": 44.01018181818182,
      "grad_norm": 11.122913360595703,
      "learning_rate": 5.656565656565657e-06,
      "loss": 0.706,
      "step": 4940
    },
    {
      "epoch": 44.012,
      "grad_norm": 24.985450744628906,
      "learning_rate": 5.555555555555556e-06,
      "loss": 0.8322,
      "step": 4950
    },
    {
      "epoch": 44.01381818181818,
      "grad_norm": 19.167444229125977,
      "learning_rate": 5.4545454545454545e-06,
      "loss": 0.6324,
      "step": 4960
    },
    {
      "epoch": 44.01563636363636,
      "grad_norm": 10.249974250793457,
      "learning_rate": 5.353535353535354e-06,
      "loss": 0.7215,
      "step": 4970
    },
    {
      "epoch": 44.01745454545455,
      "grad_norm": 5.51936149597168,
      "learning_rate": 5.2525252525252526e-06,
      "loss": 0.6117,
      "step": 4980
    },
    {
      "epoch": 44.01927272727273,
      "grad_norm": 6.55401611328125,
      "learning_rate": 5.151515151515152e-06,
      "loss": 0.5815,
      "step": 4990
    },
    {
      "epoch": 44.02018181818182,
      "eval_accuracy": 0.6058201058201058,
      "eval_loss": 0.9247404336929321,
      "eval_runtime": 148.7396,
      "eval_samples_per_second": 5.083,
      "eval_steps_per_second": 0.639,
      "step": 4995
    },
    {
      "epoch": 45.00090909090909,
      "grad_norm": 10.085493087768555,
      "learning_rate": 5.050505050505051e-06,
      "loss": 0.7191,
      "step": 5000
    },
    {
      "epoch": 45.00272727272727,
      "grad_norm": 9.831320762634277,
      "learning_rate": 4.949494949494949e-06,
      "loss": 0.7192,
      "step": 5010
    },
    {
      "epoch": 45.00454545454546,
      "grad_norm": 12.230847358703613,
      "learning_rate": 4.848484848484849e-06,
      "loss": 0.5632,
      "step": 5020
    },
    {
      "epoch": 45.00636363636364,
      "grad_norm": 12.282224655151367,
      "learning_rate": 4.747474747474747e-06,
      "loss": 0.9129,
      "step": 5030
    },
    {
      "epoch": 45.00818181818182,
      "grad_norm": 13.632772445678711,
      "learning_rate": 4.646464646464647e-06,
      "loss": 0.6591,
      "step": 5040
    },
    {
      "epoch": 45.01,
      "grad_norm": 9.214385986328125,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.7707,
      "step": 5050
    },
    {
      "epoch": 45.011818181818185,
      "grad_norm": 10.159058570861816,
      "learning_rate": 4.444444444444445e-06,
      "loss": 0.7463,
      "step": 5060
    },
    {
      "epoch": 45.013636363636365,
      "grad_norm": 12.011441230773926,
      "learning_rate": 4.343434343434344e-06,
      "loss": 0.6294,
      "step": 5070
    },
    {
      "epoch": 45.015454545454546,
      "grad_norm": 5.8238115310668945,
      "learning_rate": 4.242424242424243e-06,
      "loss": 0.7006,
      "step": 5080
    },
    {
      "epoch": 45.017272727272726,
      "grad_norm": 12.484506607055664,
      "learning_rate": 4.141414141414142e-06,
      "loss": 0.7661,
      "step": 5090
    },
    {
      "epoch": 45.019090909090906,
      "grad_norm": 13.438883781433105,
      "learning_rate": 4.040404040404041e-06,
      "loss": 0.8821,
      "step": 5100
    },
    {
      "epoch": 45.02018181818182,
      "eval_accuracy": 0.6058201058201058,
      "eval_loss": 0.9070749282836914,
      "eval_runtime": 148.301,
      "eval_samples_per_second": 5.098,
      "eval_steps_per_second": 0.641,
      "step": 5106
    },
    {
      "epoch": 46.000727272727275,
      "grad_norm": 8.703226089477539,
      "learning_rate": 3.939393939393939e-06,
      "loss": 0.7494,
      "step": 5110
    },
    {
      "epoch": 46.002545454545455,
      "grad_norm": 14.767084121704102,
      "learning_rate": 3.8383838383838385e-06,
      "loss": 0.7142,
      "step": 5120
    },
    {
      "epoch": 46.004363636363635,
      "grad_norm": 12.91554069519043,
      "learning_rate": 3.7373737373737375e-06,
      "loss": 0.6795,
      "step": 5130
    },
    {
      "epoch": 46.006181818181815,
      "grad_norm": 10.742305755615234,
      "learning_rate": 3.636363636363636e-06,
      "loss": 0.6425,
      "step": 5140
    },
    {
      "epoch": 46.008,
      "grad_norm": 14.223603248596191,
      "learning_rate": 3.5353535353535352e-06,
      "loss": 0.8141,
      "step": 5150
    },
    {
      "epoch": 46.00981818181818,
      "grad_norm": 8.773728370666504,
      "learning_rate": 3.4343434343434343e-06,
      "loss": 0.6197,
      "step": 5160
    },
    {
      "epoch": 46.01163636363636,
      "grad_norm": 5.544559478759766,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.6588,
      "step": 5170
    },
    {
      "epoch": 46.01345454545454,
      "grad_norm": 7.724081039428711,
      "learning_rate": 3.2323232323232324e-06,
      "loss": 0.7757,
      "step": 5180
    },
    {
      "epoch": 46.01527272727273,
      "grad_norm": 15.705133438110352,
      "learning_rate": 3.1313131313131314e-06,
      "loss": 0.8568,
      "step": 5190
    },
    {
      "epoch": 46.01709090909091,
      "grad_norm": 10.857434272766113,
      "learning_rate": 3.0303030303030305e-06,
      "loss": 0.6365,
      "step": 5200
    },
    {
      "epoch": 46.01890909090909,
      "grad_norm": 11.509218215942383,
      "learning_rate": 2.9292929292929295e-06,
      "loss": 0.7436,
      "step": 5210
    },
    {
      "epoch": 46.02018181818182,
      "eval_accuracy": 0.6084656084656085,
      "eval_loss": 0.8924344182014465,
      "eval_runtime": 144.2666,
      "eval_samples_per_second": 5.24,
      "eval_steps_per_second": 0.659,
      "step": 5217
    },
    {
      "epoch": 47.00054545454545,
      "grad_norm": 11.28231143951416,
      "learning_rate": 2.8282828282828286e-06,
      "loss": 0.7133,
      "step": 5220
    },
    {
      "epoch": 47.00236363636364,
      "grad_norm": 25.72589683532715,
      "learning_rate": 2.7272727272727272e-06,
      "loss": 0.7184,
      "step": 5230
    },
    {
      "epoch": 47.00418181818182,
      "grad_norm": 13.874494552612305,
      "learning_rate": 2.6262626262626263e-06,
      "loss": 0.8662,
      "step": 5240
    },
    {
      "epoch": 47.006,
      "grad_norm": 10.938508987426758,
      "learning_rate": 2.5252525252525253e-06,
      "loss": 0.5704,
      "step": 5250
    },
    {
      "epoch": 47.00781818181818,
      "grad_norm": 7.734257698059082,
      "learning_rate": 2.4242424242424244e-06,
      "loss": 0.6648,
      "step": 5260
    },
    {
      "epoch": 47.00963636363636,
      "grad_norm": 8.946194648742676,
      "learning_rate": 2.3232323232323234e-06,
      "loss": 0.8833,
      "step": 5270
    },
    {
      "epoch": 47.01145454545455,
      "grad_norm": 20.528419494628906,
      "learning_rate": 2.2222222222222225e-06,
      "loss": 0.7525,
      "step": 5280
    },
    {
      "epoch": 47.01327272727273,
      "grad_norm": 15.024425506591797,
      "learning_rate": 2.1212121212121216e-06,
      "loss": 0.6796,
      "step": 5290
    },
    {
      "epoch": 47.01509090909091,
      "grad_norm": 7.91912317276001,
      "learning_rate": 2.0202020202020206e-06,
      "loss": 0.7555,
      "step": 5300
    },
    {
      "epoch": 47.01690909090909,
      "grad_norm": 16.979326248168945,
      "learning_rate": 1.9191919191919192e-06,
      "loss": 0.5724,
      "step": 5310
    },
    {
      "epoch": 47.018727272727276,
      "grad_norm": 7.550643444061279,
      "learning_rate": 1.818181818181818e-06,
      "loss": 0.6863,
      "step": 5320
    },
    {
      "epoch": 47.02018181818182,
      "eval_accuracy": 0.6111111111111112,
      "eval_loss": 0.8965290784835815,
      "eval_runtime": 146.2721,
      "eval_samples_per_second": 5.168,
      "eval_steps_per_second": 0.649,
      "step": 5328
    },
    {
      "epoch": 48.00036363636364,
      "grad_norm": 13.027327537536621,
      "learning_rate": 1.7171717171717171e-06,
      "loss": 0.5784,
      "step": 5330
    },
    {
      "epoch": 48.00218181818182,
      "grad_norm": 9.1978759765625,
      "learning_rate": 1.6161616161616162e-06,
      "loss": 0.6893,
      "step": 5340
    },
    {
      "epoch": 48.004,
      "grad_norm": 8.523101806640625,
      "learning_rate": 1.5151515151515152e-06,
      "loss": 0.6099,
      "step": 5350
    },
    {
      "epoch": 48.005818181818185,
      "grad_norm": 15.44604778289795,
      "learning_rate": 1.4141414141414143e-06,
      "loss": 0.7395,
      "step": 5360
    },
    {
      "epoch": 48.007636363636365,
      "grad_norm": 12.837162971496582,
      "learning_rate": 1.3131313131313131e-06,
      "loss": 0.7887,
      "step": 5370
    },
    {
      "epoch": 48.009454545454545,
      "grad_norm": 16.95429039001465,
      "learning_rate": 1.2121212121212122e-06,
      "loss": 0.8216,
      "step": 5380
    },
    {
      "epoch": 48.011272727272726,
      "grad_norm": 11.841358184814453,
      "learning_rate": 1.1111111111111112e-06,
      "loss": 0.7102,
      "step": 5390
    },
    {
      "epoch": 48.013090909090906,
      "grad_norm": 12.675179481506348,
      "learning_rate": 1.0101010101010103e-06,
      "loss": 0.8177,
      "step": 5400
    },
    {
      "epoch": 48.01490909090909,
      "grad_norm": 13.496634483337402,
      "learning_rate": 9.09090909090909e-07,
      "loss": 0.7247,
      "step": 5410
    },
    {
      "epoch": 48.01672727272727,
      "grad_norm": 11.762472152709961,
      "learning_rate": 8.080808080808081e-07,
      "loss": 0.6307,
      "step": 5420
    },
    {
      "epoch": 48.01854545454545,
      "grad_norm": 25.6706485748291,
      "learning_rate": 7.070707070707071e-07,
      "loss": 0.7035,
      "step": 5430
    },
    {
      "epoch": 48.02018181818182,
      "eval_accuracy": 0.6044973544973545,
      "eval_loss": 0.8940863013267517,
      "eval_runtime": 149.5287,
      "eval_samples_per_second": 5.056,
      "eval_steps_per_second": 0.635,
      "step": 5439
    },
    {
      "epoch": 49.000181818181815,
      "grad_norm": 12.681800842285156,
      "learning_rate": 6.060606060606061e-07,
      "loss": 0.5424,
      "step": 5440
    },
    {
      "epoch": 49.002,
      "grad_norm": 10.071161270141602,
      "learning_rate": 5.050505050505052e-07,
      "loss": 0.6097,
      "step": 5450
    },
    {
      "epoch": 49.00381818181818,
      "grad_norm": 11.494267463684082,
      "learning_rate": 4.0404040404040405e-07,
      "loss": 0.5644,
      "step": 5460
    },
    {
      "epoch": 49.00563636363636,
      "grad_norm": 15.089944839477539,
      "learning_rate": 3.0303030303030305e-07,
      "loss": 0.714,
      "step": 5470
    },
    {
      "epoch": 49.00745454545454,
      "grad_norm": 8.669146537780762,
      "learning_rate": 2.0202020202020202e-07,
      "loss": 0.7086,
      "step": 5480
    },
    {
      "epoch": 49.00927272727273,
      "grad_norm": 10.859286308288574,
      "learning_rate": 1.0101010101010101e-07,
      "loss": 0.6569,
      "step": 5490
    },
    {
      "epoch": 49.01109090909091,
      "grad_norm": 10.49977970123291,
      "learning_rate": 0.0,
      "loss": 0.6348,
      "step": 5500
    },
    {
      "epoch": 49.01109090909091,
      "eval_accuracy": 0.6124338624338624,
      "eval_loss": 0.8949828743934631,
      "eval_runtime": 148.7108,
      "eval_samples_per_second": 5.084,
      "eval_steps_per_second": 0.639,
      "step": 5500
    },
    {
      "epoch": 49.01109090909091,
      "step": 5500,
      "total_flos": 2.3965589411203056e+19,
      "train_loss": 0.8993404138738459,
      "train_runtime": 16738.8602,
      "train_samples_per_second": 2.629,
      "train_steps_per_second": 0.329
    },
    {
      "epoch": 49.01109090909091,
      "eval_accuracy": 0.5595238095238095,
      "eval_loss": 0.9923801422119141,
      "eval_runtime": 148.7567,
      "eval_samples_per_second": 5.082,
      "eval_steps_per_second": 0.639,
      "step": 5500
    }
  ],
  "logging_steps": 10,
  "max_steps": 5500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.3965589411203056e+19,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}