{
  "best_metric": 0.7407407407407407,
  "best_model_checkpoint": "vvt-gs-rot-flip-wtoken-f198-4.4-h768-t8.16.16\\checkpoint-5500",
  "epoch": 1.0001818181818183,
  "eval_steps": 500,
  "global_step": 5501,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0018181818181818182,
      "grad_norm": 5.903317928314209,
      "learning_rate": 9.09090909090909e-07,
      "loss": 1.0691,
      "step": 10
    },
    {
      "epoch": 0.0036363636363636364,
      "grad_norm": 11.508553504943848,
      "learning_rate": 1.818181818181818e-06,
      "loss": 1.1499,
      "step": 20
    },
    {
      "epoch": 0.005454545454545455,
      "grad_norm": 9.509793281555176,
      "learning_rate": 2.7272727272727272e-06,
      "loss": 1.0829,
      "step": 30
    },
    {
      "epoch": 0.007272727272727273,
      "grad_norm": 18.586177825927734,
      "learning_rate": 3.636363636363636e-06,
      "loss": 1.1101,
      "step": 40
    },
    {
      "epoch": 0.00909090909090909,
      "grad_norm": 10.552851676940918,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 1.065,
      "step": 50
    },
    {
      "epoch": 0.01090909090909091,
      "grad_norm": 10.60718822479248,
      "learning_rate": 5.4545454545454545e-06,
      "loss": 1.127,
      "step": 60
    },
    {
      "epoch": 0.012727272727272728,
      "grad_norm": 5.415429592132568,
      "learning_rate": 6.363636363636363e-06,
      "loss": 1.0046,
      "step": 70
    },
    {
      "epoch": 0.014545454545454545,
      "grad_norm": 7.580716133117676,
      "learning_rate": 7.272727272727272e-06,
      "loss": 1.1022,
      "step": 80
    },
    {
      "epoch": 0.016363636363636365,
      "grad_norm": 34.73521041870117,
      "learning_rate": 8.181818181818183e-06,
      "loss": 1.0765,
      "step": 90
    },
    {
      "epoch": 0.01818181818181818,
      "grad_norm": 18.438007354736328,
      "learning_rate": 9.090909090909091e-06,
      "loss": 1.0189,
      "step": 100
    },
    {
      "epoch": 0.02,
      "grad_norm": 54.06311798095703,
      "learning_rate": 1e-05,
      "loss": 1.1794,
      "step": 110
    },
    {
      "epoch": 0.02181818181818182,
      "grad_norm": 7.959690570831299,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 1.1737,
      "step": 120
    },
    {
      "epoch": 0.023636363636363636,
      "grad_norm": 33.0672492980957,
      "learning_rate": 1.1818181818181819e-05,
      "loss": 1.117,
      "step": 130
    },
    {
      "epoch": 0.025454545454545455,
      "grad_norm": 20.22812271118164,
      "learning_rate": 1.2727272727272727e-05,
      "loss": 1.1634,
      "step": 140
    },
    {
      "epoch": 0.02727272727272727,
      "grad_norm": 27.637243270874023,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 1.0183,
      "step": 150
    },
    {
      "epoch": 0.02909090909090909,
      "grad_norm": 16.066158294677734,
      "learning_rate": 1.4545454545454545e-05,
      "loss": 1.2349,
      "step": 160
    },
    {
      "epoch": 0.03090909090909091,
      "grad_norm": 28.06402015686035,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 1.0342,
      "step": 170
    },
    {
      "epoch": 0.03272727272727273,
      "grad_norm": 11.528419494628906,
      "learning_rate": 1.6363636363636366e-05,
      "loss": 1.2609,
      "step": 180
    },
    {
      "epoch": 0.034545454545454546,
      "grad_norm": 14.250886917114258,
      "learning_rate": 1.7272727272727274e-05,
      "loss": 1.1208,
      "step": 190
    },
    {
      "epoch": 0.03636363636363636,
      "grad_norm": 9.647354125976562,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 1.1326,
      "step": 200
    },
    {
      "epoch": 0.038181818181818185,
      "grad_norm": 9.057283401489258,
      "learning_rate": 1.9090909090909094e-05,
      "loss": 1.1233,
      "step": 210
    },
    {
      "epoch": 0.04,
      "grad_norm": 6.013422966003418,
      "learning_rate": 2e-05,
      "loss": 1.0266,
      "step": 220
    },
    {
      "epoch": 0.04018181818181818,
      "eval_accuracy": 0.42857142857142855,
      "eval_loss": 1.0405620336532593,
      "eval_runtime": 193.3467,
      "eval_samples_per_second": 0.978,
      "eval_steps_per_second": 0.248,
      "step": 221
    },
    {
      "epoch": 1.0016363636363637,
      "grad_norm": 35.22421646118164,
      "learning_rate": 2.090909090909091e-05,
      "loss": 1.1156,
      "step": 230
    },
    {
      "epoch": 1.0034545454545454,
      "grad_norm": 11.300332069396973,
      "learning_rate": 2.1818181818181818e-05,
      "loss": 1.1326,
      "step": 240
    },
    {
      "epoch": 1.0052727272727273,
      "grad_norm": 16.684968948364258,
      "learning_rate": 2.272727272727273e-05,
      "loss": 1.1189,
      "step": 250
    },
    {
      "epoch": 1.007090909090909,
      "grad_norm": 16.38736915588379,
      "learning_rate": 2.3636363636363637e-05,
      "loss": 1.1015,
      "step": 260
    },
    {
      "epoch": 1.008909090909091,
      "grad_norm": 22.911460876464844,
      "learning_rate": 2.4545454545454545e-05,
      "loss": 1.0275,
      "step": 270
    },
    {
      "epoch": 1.0107272727272727,
      "grad_norm": 18.65599250793457,
      "learning_rate": 2.5454545454545454e-05,
      "loss": 1.1375,
      "step": 280
    },
    {
      "epoch": 1.0125454545454546,
      "grad_norm": 15.462641716003418,
      "learning_rate": 2.636363636363636e-05,
      "loss": 0.9,
      "step": 290
    },
    {
      "epoch": 1.0143636363636364,
      "grad_norm": 38.87071990966797,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 1.226,
      "step": 300
    },
    {
      "epoch": 1.016181818181818,
      "grad_norm": 27.41946029663086,
      "learning_rate": 2.818181818181818e-05,
      "loss": 1.1471,
      "step": 310
    },
    {
      "epoch": 1.018,
      "grad_norm": 14.629090309143066,
      "learning_rate": 2.909090909090909e-05,
      "loss": 1.1017,
      "step": 320
    },
    {
      "epoch": 1.0198181818181817,
      "grad_norm": 11.047767639160156,
      "learning_rate": 3e-05,
      "loss": 1.155,
      "step": 330
    },
    {
      "epoch": 1.0216363636363637,
      "grad_norm": 48.5820198059082,
      "learning_rate": 3.090909090909091e-05,
      "loss": 1.0182,
      "step": 340
    },
    {
      "epoch": 1.0234545454545454,
      "grad_norm": 29.4666805267334,
      "learning_rate": 3.181818181818182e-05,
      "loss": 1.0224,
      "step": 350
    },
    {
      "epoch": 1.0252727272727273,
      "grad_norm": 23.09134292602539,
      "learning_rate": 3.272727272727273e-05,
      "loss": 0.8628,
      "step": 360
    },
    {
      "epoch": 1.027090909090909,
      "grad_norm": 35.83842849731445,
      "learning_rate": 3.3636363636363636e-05,
      "loss": 1.0581,
      "step": 370
    },
    {
      "epoch": 1.028909090909091,
      "grad_norm": 11.075454711914062,
      "learning_rate": 3.454545454545455e-05,
      "loss": 1.1639,
      "step": 380
    },
    {
      "epoch": 1.0307272727272727,
      "grad_norm": 16.66240119934082,
      "learning_rate": 3.545454545454546e-05,
      "loss": 1.1218,
      "step": 390
    },
    {
      "epoch": 1.0325454545454544,
      "grad_norm": 18.380191802978516,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 0.9699,
      "step": 400
    },
    {
      "epoch": 1.0343636363636364,
      "grad_norm": 105.08155822753906,
      "learning_rate": 3.7272727272727276e-05,
      "loss": 1.1986,
      "step": 410
    },
    {
      "epoch": 1.036181818181818,
      "grad_norm": 15.301504135131836,
      "learning_rate": 3.818181818181819e-05,
      "loss": 1.3065,
      "step": 420
    },
    {
      "epoch": 1.038,
      "grad_norm": 11.040570259094238,
      "learning_rate": 3.909090909090909e-05,
      "loss": 1.1796,
      "step": 430
    },
    {
      "epoch": 1.0398181818181818,
      "grad_norm": 20.411191940307617,
      "learning_rate": 4e-05,
      "loss": 1.0928,
      "step": 440
    },
    {
      "epoch": 1.040181818181818,
      "eval_accuracy": 0.5291005291005291,
      "eval_loss": 0.9597143530845642,
      "eval_runtime": 200.8753,
      "eval_samples_per_second": 0.941,
      "eval_steps_per_second": 0.239,
      "step": 442
    },
    {
      "epoch": 2.0014545454545454,
      "grad_norm": 58.89450454711914,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.9651,
      "step": 450
    },
    {
      "epoch": 2.0032727272727273,
      "grad_norm": 17.683551788330078,
      "learning_rate": 4.181818181818182e-05,
      "loss": 1.1622,
      "step": 460
    },
    {
      "epoch": 2.0050909090909093,
      "grad_norm": 16.6800537109375,
      "learning_rate": 4.2727272727272724e-05,
      "loss": 1.0687,
      "step": 470
    },
    {
      "epoch": 2.0069090909090908,
      "grad_norm": 19.49321174621582,
      "learning_rate": 4.3636363636363636e-05,
      "loss": 1.0215,
      "step": 480
    },
    {
      "epoch": 2.0087272727272727,
      "grad_norm": 11.021658897399902,
      "learning_rate": 4.454545454545455e-05,
      "loss": 1.0232,
      "step": 490
    },
    {
      "epoch": 2.0105454545454546,
      "grad_norm": 41.21800994873047,
      "learning_rate": 4.545454545454546e-05,
      "loss": 1.0014,
      "step": 500
    },
    {
      "epoch": 2.0123636363636366,
      "grad_norm": 7.544427871704102,
      "learning_rate": 4.636363636363636e-05,
      "loss": 0.844,
      "step": 510
    },
    {
      "epoch": 2.014181818181818,
      "grad_norm": 41.244773864746094,
      "learning_rate": 4.7272727272727275e-05,
      "loss": 1.2135,
      "step": 520
    },
    {
      "epoch": 2.016,
      "grad_norm": 7.405851364135742,
      "learning_rate": 4.8181818181818186e-05,
      "loss": 1.1618,
      "step": 530
    },
    {
      "epoch": 2.017818181818182,
      "grad_norm": 11.462644577026367,
      "learning_rate": 4.909090909090909e-05,
      "loss": 1.0918,
      "step": 540
    },
    {
      "epoch": 2.0196363636363635,
      "grad_norm": 13.532742500305176,
      "learning_rate": 5e-05,
      "loss": 1.0293,
      "step": 550
    },
    {
      "epoch": 2.0214545454545454,
      "grad_norm": 21.01549530029297,
      "learning_rate": 4.98989898989899e-05,
      "loss": 1.0992,
      "step": 560
    },
    {
      "epoch": 2.0232727272727273,
      "grad_norm": 8.411489486694336,
      "learning_rate": 4.97979797979798e-05,
      "loss": 1.0189,
      "step": 570
    },
    {
      "epoch": 2.0250909090909093,
      "grad_norm": 11.137345314025879,
      "learning_rate": 4.9696969696969694e-05,
      "loss": 1.0663,
      "step": 580
    },
    {
      "epoch": 2.0269090909090908,
      "grad_norm": 30.53400230407715,
      "learning_rate": 4.9595959595959594e-05,
      "loss": 1.0683,
      "step": 590
    },
    {
      "epoch": 2.0287272727272727,
      "grad_norm": 47.95043182373047,
      "learning_rate": 4.94949494949495e-05,
      "loss": 0.9578,
      "step": 600
    },
    {
      "epoch": 2.0305454545454547,
      "grad_norm": 13.816718101501465,
      "learning_rate": 4.93939393939394e-05,
      "loss": 1.0128,
      "step": 610
    },
    {
      "epoch": 2.032363636363636,
      "grad_norm": 40.12334442138672,
      "learning_rate": 4.92929292929293e-05,
      "loss": 1.084,
      "step": 620
    },
    {
      "epoch": 2.034181818181818,
      "grad_norm": 48.186866760253906,
      "learning_rate": 4.919191919191919e-05,
      "loss": 0.9239,
      "step": 630
    },
    {
      "epoch": 2.036,
      "grad_norm": 35.39604949951172,
      "learning_rate": 4.909090909090909e-05,
      "loss": 1.0707,
      "step": 640
    },
    {
      "epoch": 2.037818181818182,
      "grad_norm": 16.182966232299805,
      "learning_rate": 4.898989898989899e-05,
      "loss": 1.2235,
      "step": 650
    },
    {
      "epoch": 2.0396363636363635,
      "grad_norm": 9.64621639251709,
      "learning_rate": 4.888888888888889e-05,
      "loss": 0.862,
      "step": 660
    },
    {
      "epoch": 2.0401818181818183,
      "eval_accuracy": 0.4708994708994709,
      "eval_loss": 0.9646209478378296,
      "eval_runtime": 201.5162,
      "eval_samples_per_second": 0.938,
      "eval_steps_per_second": 0.238,
      "step": 663
    },
    {
      "epoch": 3.001272727272727,
      "grad_norm": 16.516555786132812,
      "learning_rate": 4.878787878787879e-05,
      "loss": 0.9561,
      "step": 670
    },
    {
      "epoch": 3.003090909090909,
      "grad_norm": 13.819075584411621,
      "learning_rate": 4.868686868686869e-05,
      "loss": 1.0881,
      "step": 680
    },
    {
      "epoch": 3.004909090909091,
      "grad_norm": 10.121333122253418,
      "learning_rate": 4.858585858585859e-05,
      "loss": 1.0883,
      "step": 690
    },
    {
      "epoch": 3.006727272727273,
      "grad_norm": 12.964982032775879,
      "learning_rate": 4.848484848484849e-05,
      "loss": 1.0234,
      "step": 700
    },
    {
      "epoch": 3.0085454545454544,
      "grad_norm": 9.034525871276855,
      "learning_rate": 4.838383838383839e-05,
      "loss": 0.8553,
      "step": 710
    },
    {
      "epoch": 3.0103636363636364,
      "grad_norm": 6.48071813583374,
      "learning_rate": 4.828282828282829e-05,
      "loss": 1.3251,
      "step": 720
    },
    {
      "epoch": 3.0121818181818183,
      "grad_norm": 12.058953285217285,
      "learning_rate": 4.8181818181818186e-05,
      "loss": 1.05,
      "step": 730
    },
    {
      "epoch": 3.014,
      "grad_norm": 16.26006317138672,
      "learning_rate": 4.808080808080808e-05,
      "loss": 1.0792,
      "step": 740
    },
    {
      "epoch": 3.0158181818181817,
      "grad_norm": 63.56451416015625,
      "learning_rate": 4.797979797979798e-05,
      "loss": 0.9569,
      "step": 750
    },
    {
      "epoch": 3.0176363636363637,
      "grad_norm": 27.69994354248047,
      "learning_rate": 4.787878787878788e-05,
      "loss": 1.0634,
      "step": 760
    },
    {
      "epoch": 3.0194545454545456,
      "grad_norm": 20.30846405029297,
      "learning_rate": 4.7777777777777784e-05,
      "loss": 1.1572,
      "step": 770
    },
    {
      "epoch": 3.021272727272727,
      "grad_norm": 16.615659713745117,
      "learning_rate": 4.7676767676767684e-05,
      "loss": 1.0451,
      "step": 780
    },
    {
      "epoch": 3.023090909090909,
      "grad_norm": 4.919287204742432,
      "learning_rate": 4.7575757575757576e-05,
      "loss": 1.1037,
      "step": 790
    },
    {
      "epoch": 3.024909090909091,
      "grad_norm": 8.121920585632324,
      "learning_rate": 4.7474747474747476e-05,
      "loss": 0.8158,
      "step": 800
    },
    {
      "epoch": 3.026727272727273,
      "grad_norm": 22.583158493041992,
      "learning_rate": 4.7373737373737375e-05,
      "loss": 0.8648,
      "step": 810
    },
    {
      "epoch": 3.0285454545454544,
      "grad_norm": 55.47071075439453,
      "learning_rate": 4.7272727272727275e-05,
      "loss": 0.8976,
      "step": 820
    },
    {
      "epoch": 3.0303636363636364,
      "grad_norm": 12.427018165588379,
      "learning_rate": 4.7171717171717174e-05,
      "loss": 1.1084,
      "step": 830
    },
    {
      "epoch": 3.0321818181818183,
      "grad_norm": 7.498060703277588,
      "learning_rate": 4.7070707070707074e-05,
      "loss": 0.9779,
      "step": 840
    },
    {
      "epoch": 3.034,
      "grad_norm": 12.233386039733887,
      "learning_rate": 4.696969696969697e-05,
      "loss": 0.9867,
      "step": 850
    },
    {
      "epoch": 3.0358181818181817,
      "grad_norm": 14.040285110473633,
      "learning_rate": 4.686868686868687e-05,
      "loss": 0.9304,
      "step": 860
    },
    {
      "epoch": 3.0376363636363637,
      "grad_norm": 12.534895896911621,
      "learning_rate": 4.676767676767677e-05,
      "loss": 0.8739,
      "step": 870
    },
    {
      "epoch": 3.0394545454545456,
      "grad_norm": 47.45646286010742,
      "learning_rate": 4.666666666666667e-05,
      "loss": 0.9291,
      "step": 880
    },
    {
      "epoch": 3.0401818181818183,
      "eval_accuracy": 0.48677248677248675,
      "eval_loss": 0.9938625693321228,
      "eval_runtime": 203.337,
      "eval_samples_per_second": 0.929,
      "eval_steps_per_second": 0.236,
      "step": 884
    },
    {
      "epoch": 4.001090909090909,
      "grad_norm": 33.868717193603516,
      "learning_rate": 4.656565656565657e-05,
      "loss": 1.0813,
      "step": 890
    },
    {
      "epoch": 4.002909090909091,
      "grad_norm": 8.45543098449707,
      "learning_rate": 4.6464646464646464e-05,
      "loss": 1.0566,
      "step": 900
    },
    {
      "epoch": 4.004727272727273,
      "grad_norm": 13.185288429260254,
      "learning_rate": 4.636363636363636e-05,
      "loss": 1.1579,
      "step": 910
    },
    {
      "epoch": 4.006545454545455,
      "grad_norm": 11.203399658203125,
      "learning_rate": 4.626262626262626e-05,
      "loss": 0.8934,
      "step": 920
    },
    {
      "epoch": 4.008363636363637,
      "grad_norm": 36.20549774169922,
      "learning_rate": 4.616161616161616e-05,
      "loss": 1.1601,
      "step": 930
    },
    {
      "epoch": 4.0101818181818185,
      "grad_norm": 69.12691497802734,
      "learning_rate": 4.606060606060607e-05,
      "loss": 0.9967,
      "step": 940
    },
    {
      "epoch": 4.012,
      "grad_norm": 16.70299530029297,
      "learning_rate": 4.595959595959596e-05,
      "loss": 1.0471,
      "step": 950
    },
    {
      "epoch": 4.0138181818181815,
      "grad_norm": 87.73255157470703,
      "learning_rate": 4.585858585858586e-05,
      "loss": 1.1211,
      "step": 960
    },
    {
      "epoch": 4.0156363636363634,
      "grad_norm": 201.55203247070312,
      "learning_rate": 4.575757575757576e-05,
      "loss": 1.1687,
      "step": 970
    },
    {
      "epoch": 4.017454545454545,
      "grad_norm": 33.13986587524414,
      "learning_rate": 4.565656565656566e-05,
      "loss": 0.9516,
      "step": 980
    },
    {
      "epoch": 4.019272727272727,
      "grad_norm": 9.230415344238281,
      "learning_rate": 4.555555555555556e-05,
      "loss": 1.0018,
      "step": 990
    },
    {
      "epoch": 4.021090909090909,
      "grad_norm": 77.70652770996094,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.9895,
      "step": 1000
    },
    {
      "epoch": 4.022909090909091,
      "grad_norm": 145.90219116210938,
      "learning_rate": 4.535353535353535e-05,
      "loss": 1.2767,
      "step": 1010
    },
    {
      "epoch": 4.024727272727273,
      "grad_norm": 11.502787590026855,
      "learning_rate": 4.525252525252526e-05,
      "loss": 0.8943,
      "step": 1020
    },
    {
      "epoch": 4.026545454545454,
      "grad_norm": 60.249473571777344,
      "learning_rate": 4.515151515151516e-05,
      "loss": 1.2685,
      "step": 1030
    },
    {
      "epoch": 4.028363636363636,
      "grad_norm": 4.921321868896484,
      "learning_rate": 4.5050505050505056e-05,
      "loss": 1.0308,
      "step": 1040
    },
    {
      "epoch": 4.030181818181818,
      "grad_norm": 38.16105651855469,
      "learning_rate": 4.494949494949495e-05,
      "loss": 0.9882,
      "step": 1050
    },
    {
      "epoch": 4.032,
      "grad_norm": 10.788283348083496,
      "learning_rate": 4.484848484848485e-05,
      "loss": 0.9997,
      "step": 1060
    },
    {
      "epoch": 4.033818181818182,
      "grad_norm": 9.961264610290527,
      "learning_rate": 4.474747474747475e-05,
      "loss": 0.8205,
      "step": 1070
    },
    {
      "epoch": 4.035636363636364,
      "grad_norm": 18.4578914642334,
      "learning_rate": 4.464646464646465e-05,
      "loss": 1.0358,
      "step": 1080
    },
    {
      "epoch": 4.037454545454546,
      "grad_norm": 9.838109970092773,
      "learning_rate": 4.454545454545455e-05,
      "loss": 1.0122,
      "step": 1090
    },
    {
      "epoch": 4.039272727272727,
      "grad_norm": 4.6190996170043945,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.8705,
      "step": 1100
    },
    {
      "epoch": 4.040181818181818,
      "eval_accuracy": 0.5502645502645502,
      "eval_loss": 1.0091381072998047,
      "eval_runtime": 192.457,
      "eval_samples_per_second": 0.982,
      "eval_steps_per_second": 0.249,
      "step": 1105
    },
    {
      "epoch": 5.000909090909091,
      "grad_norm": 2.803194284439087,
      "learning_rate": 4.4343434343434346e-05,
      "loss": 1.1634,
      "step": 1110
    },
    {
      "epoch": 5.002727272727273,
      "grad_norm": 12.49787712097168,
      "learning_rate": 4.4242424242424246e-05,
      "loss": 1.2069,
      "step": 1120
    },
    {
      "epoch": 5.004545454545455,
      "grad_norm": 34.8055305480957,
      "learning_rate": 4.4141414141414145e-05,
      "loss": 1.0508,
      "step": 1130
    },
    {
      "epoch": 5.006363636363637,
      "grad_norm": 105.75518035888672,
      "learning_rate": 4.4040404040404044e-05,
      "loss": 0.9281,
      "step": 1140
    },
    {
      "epoch": 5.008181818181818,
      "grad_norm": 11.095450401306152,
      "learning_rate": 4.3939393939393944e-05,
      "loss": 1.0998,
      "step": 1150
    },
    {
      "epoch": 5.01,
      "grad_norm": 12.41756534576416,
      "learning_rate": 4.383838383838384e-05,
      "loss": 1.0209,
      "step": 1160
    },
    {
      "epoch": 5.011818181818182,
      "grad_norm": 9.809650421142578,
      "learning_rate": 4.3737373737373736e-05,
      "loss": 0.8645,
      "step": 1170
    },
    {
      "epoch": 5.013636363636364,
      "grad_norm": 4.656230926513672,
      "learning_rate": 4.3636363636363636e-05,
      "loss": 0.9903,
      "step": 1180
    },
    {
      "epoch": 5.015454545454546,
      "grad_norm": 10.687009811401367,
      "learning_rate": 4.3535353535353535e-05,
      "loss": 0.9723,
      "step": 1190
    },
    {
      "epoch": 5.0172727272727276,
      "grad_norm": 55.024658203125,
      "learning_rate": 4.343434343434344e-05,
      "loss": 0.8759,
      "step": 1200
    },
    {
      "epoch": 5.0190909090909095,
      "grad_norm": 27.996280670166016,
      "learning_rate": 4.3333333333333334e-05,
      "loss": 0.9792,
      "step": 1210
    },
    {
      "epoch": 5.0209090909090905,
      "grad_norm": 9.637792587280273,
      "learning_rate": 4.3232323232323234e-05,
      "loss": 0.9309,
      "step": 1220
    },
    {
      "epoch": 5.0227272727272725,
      "grad_norm": 7.562610626220703,
      "learning_rate": 4.313131313131313e-05,
      "loss": 0.9194,
      "step": 1230
    },
    {
      "epoch": 5.024545454545454,
      "grad_norm": 22.949871063232422,
      "learning_rate": 4.303030303030303e-05,
      "loss": 1.0565,
      "step": 1240
    },
    {
      "epoch": 5.026363636363636,
      "grad_norm": 18.67213249206543,
      "learning_rate": 4.292929292929293e-05,
      "loss": 0.8991,
      "step": 1250
    },
    {
      "epoch": 5.028181818181818,
      "grad_norm": 9.366035461425781,
      "learning_rate": 4.282828282828283e-05,
      "loss": 0.9728,
      "step": 1260
    },
    {
      "epoch": 5.03,
      "grad_norm": 21.2210693359375,
      "learning_rate": 4.2727272727272724e-05,
      "loss": 1.0556,
      "step": 1270
    },
    {
      "epoch": 5.031818181818182,
      "grad_norm": 19.24120330810547,
      "learning_rate": 4.262626262626263e-05,
      "loss": 0.9918,
      "step": 1280
    },
    {
      "epoch": 5.033636363636363,
      "grad_norm": 45.28664779663086,
      "learning_rate": 4.252525252525253e-05,
      "loss": 1.1574,
      "step": 1290
    },
    {
      "epoch": 5.035454545454545,
      "grad_norm": 47.121604919433594,
      "learning_rate": 4.242424242424243e-05,
      "loss": 0.8701,
      "step": 1300
    },
    {
      "epoch": 5.037272727272727,
      "grad_norm": 22.343584060668945,
      "learning_rate": 4.232323232323233e-05,
      "loss": 0.8461,
      "step": 1310
    },
    {
      "epoch": 5.039090909090909,
      "grad_norm": 11.974569320678711,
      "learning_rate": 4.222222222222222e-05,
      "loss": 0.9667,
      "step": 1320
    },
    {
      "epoch": 5.040181818181818,
      "eval_accuracy": 0.5555555555555556,
      "eval_loss": 0.9415688514709473,
      "eval_runtime": 192.3499,
      "eval_samples_per_second": 0.983,
      "eval_steps_per_second": 0.25,
      "step": 1326
    },
    {
      "epoch": 6.000727272727273,
      "grad_norm": 14.801665306091309,
      "learning_rate": 4.212121212121212e-05,
      "loss": 0.9938,
      "step": 1330
    },
    {
      "epoch": 6.002545454545454,
      "grad_norm": 4.720005989074707,
      "learning_rate": 4.202020202020202e-05,
      "loss": 0.9117,
      "step": 1340
    },
    {
      "epoch": 6.004363636363636,
      "grad_norm": 11.161520957946777,
      "learning_rate": 4.191919191919192e-05,
      "loss": 1.1576,
      "step": 1350
    },
    {
      "epoch": 6.006181818181818,
      "grad_norm": 50.02983474731445,
      "learning_rate": 4.181818181818182e-05,
      "loss": 0.9844,
      "step": 1360
    },
    {
      "epoch": 6.008,
      "grad_norm": 6.945914268493652,
      "learning_rate": 4.171717171717172e-05,
      "loss": 0.9107,
      "step": 1370
    },
    {
      "epoch": 6.009818181818182,
      "grad_norm": 12.48246955871582,
      "learning_rate": 4.161616161616162e-05,
      "loss": 0.9041,
      "step": 1380
    },
    {
      "epoch": 6.011636363636364,
      "grad_norm": 71.0044937133789,
      "learning_rate": 4.151515151515152e-05,
      "loss": 0.8598,
      "step": 1390
    },
    {
      "epoch": 6.013454545454546,
      "grad_norm": 6.990677833557129,
      "learning_rate": 4.141414141414142e-05,
      "loss": 1.0659,
      "step": 1400
    },
    {
      "epoch": 6.015272727272727,
      "grad_norm": 9.843575477600098,
      "learning_rate": 4.131313131313132e-05,
      "loss": 0.9873,
      "step": 1410
    },
    {
      "epoch": 6.017090909090909,
      "grad_norm": 4.533863067626953,
      "learning_rate": 4.1212121212121216e-05,
      "loss": 0.9079,
      "step": 1420
    },
    {
      "epoch": 6.018909090909091,
      "grad_norm": 33.53059005737305,
      "learning_rate": 4.111111111111111e-05,
      "loss": 0.7887,
      "step": 1430
    },
    {
      "epoch": 6.020727272727273,
      "grad_norm": 26.704490661621094,
      "learning_rate": 4.101010101010101e-05,
      "loss": 1.0126,
      "step": 1440
    },
    {
      "epoch": 6.022545454545455,
      "grad_norm": 24.7235107421875,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.8291,
      "step": 1450
    },
    {
      "epoch": 6.024363636363637,
      "grad_norm": 13.622713088989258,
      "learning_rate": 4.0808080808080814e-05,
      "loss": 1.0642,
      "step": 1460
    },
    {
      "epoch": 6.0261818181818185,
      "grad_norm": 3.207956314086914,
      "learning_rate": 4.070707070707071e-05,
      "loss": 0.8796,
      "step": 1470
    },
    {
      "epoch": 6.028,
      "grad_norm": 7.998806953430176,
      "learning_rate": 4.0606060606060606e-05,
      "loss": 0.8937,
      "step": 1480
    },
    {
      "epoch": 6.0298181818181815,
      "grad_norm": 8.217198371887207,
      "learning_rate": 4.0505050505050506e-05,
      "loss": 0.9417,
      "step": 1490
    },
    {
      "epoch": 6.0316363636363635,
      "grad_norm": 11.666306495666504,
      "learning_rate": 4.0404040404040405e-05,
      "loss": 0.945,
      "step": 1500
    },
    {
      "epoch": 6.033454545454545,
      "grad_norm": 7.258649826049805,
      "learning_rate": 4.0303030303030305e-05,
      "loss": 0.898,
      "step": 1510
    },
    {
      "epoch": 6.035272727272727,
      "grad_norm": 33.963783264160156,
      "learning_rate": 4.0202020202020204e-05,
      "loss": 1.1643,
      "step": 1520
    },
    {
      "epoch": 6.037090909090909,
      "grad_norm": 20.44380760192871,
      "learning_rate": 4.01010101010101e-05,
      "loss": 0.9079,
      "step": 1530
    },
    {
      "epoch": 6.038909090909091,
      "grad_norm": 47.127071380615234,
      "learning_rate": 4e-05,
      "loss": 0.9227,
      "step": 1540
    },
    {
      "epoch": 6.040181818181818,
      "eval_accuracy": 0.582010582010582,
      "eval_loss": 0.8646734356880188,
      "eval_runtime": 188.7776,
      "eval_samples_per_second": 1.001,
      "eval_steps_per_second": 0.254,
      "step": 1547
    },
    {
      "epoch": 7.000545454545454,
      "grad_norm": 23.793865203857422,
      "learning_rate": 3.98989898989899e-05,
      "loss": 0.9091,
      "step": 1550
    },
    {
      "epoch": 7.002363636363636,
      "grad_norm": 97.33353424072266,
      "learning_rate": 3.97979797979798e-05,
      "loss": 1.0564,
      "step": 1560
    },
    {
      "epoch": 7.004181818181818,
      "grad_norm": 15.528971672058105,
      "learning_rate": 3.96969696969697e-05,
      "loss": 0.9467,
      "step": 1570
    },
    {
      "epoch": 7.006,
      "grad_norm": 13.52229118347168,
      "learning_rate": 3.9595959595959594e-05,
      "loss": 0.9877,
      "step": 1580
    },
    {
      "epoch": 7.007818181818182,
      "grad_norm": 9.012044906616211,
      "learning_rate": 3.9494949494949494e-05,
      "loss": 0.9771,
      "step": 1590
    },
    {
      "epoch": 7.009636363636363,
      "grad_norm": 29.697484970092773,
      "learning_rate": 3.939393939393939e-05,
      "loss": 0.8836,
      "step": 1600
    },
    {
      "epoch": 7.011454545454545,
      "grad_norm": 13.41483211517334,
      "learning_rate": 3.929292929292929e-05,
      "loss": 1.3449,
      "step": 1610
    },
    {
      "epoch": 7.013272727272727,
      "grad_norm": 8.07792854309082,
      "learning_rate": 3.91919191919192e-05,
      "loss": 0.9811,
      "step": 1620
    },
    {
      "epoch": 7.015090909090909,
      "grad_norm": 10.059922218322754,
      "learning_rate": 3.909090909090909e-05,
      "loss": 0.9815,
      "step": 1630
    },
    {
      "epoch": 7.016909090909091,
      "grad_norm": 18.550546646118164,
      "learning_rate": 3.898989898989899e-05,
      "loss": 0.9835,
      "step": 1640
    },
    {
      "epoch": 7.018727272727273,
      "grad_norm": 29.54568099975586,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.9226,
      "step": 1650
    },
    {
      "epoch": 7.020545454545455,
      "grad_norm": 48.8908805847168,
      "learning_rate": 3.878787878787879e-05,
      "loss": 0.9583,
      "step": 1660
    },
    {
      "epoch": 7.022363636363636,
      "grad_norm": 53.94483947753906,
      "learning_rate": 3.868686868686869e-05,
      "loss": 0.8708,
      "step": 1670
    },
    {
      "epoch": 7.024181818181818,
      "grad_norm": 7.722486972808838,
      "learning_rate": 3.858585858585859e-05,
      "loss": 0.9358,
      "step": 1680
    },
    {
      "epoch": 7.026,
      "grad_norm": 23.644926071166992,
      "learning_rate": 3.848484848484848e-05,
      "loss": 1.0788,
      "step": 1690
    },
    {
      "epoch": 7.027818181818182,
      "grad_norm": 7.95550537109375,
      "learning_rate": 3.838383838383838e-05,
      "loss": 0.9001,
      "step": 1700
    },
    {
      "epoch": 7.029636363636364,
      "grad_norm": 32.47743606567383,
      "learning_rate": 3.828282828282829e-05,
      "loss": 0.9419,
      "step": 1710
    },
    {
      "epoch": 7.031454545454546,
      "grad_norm": 28.465147018432617,
      "learning_rate": 3.818181818181819e-05,
      "loss": 0.8118,
      "step": 1720
    },
    {
      "epoch": 7.033272727272728,
      "grad_norm": 27.852588653564453,
      "learning_rate": 3.8080808080808087e-05,
      "loss": 0.9924,
      "step": 1730
    },
    {
      "epoch": 7.0350909090909095,
      "grad_norm": 9.697334289550781,
      "learning_rate": 3.797979797979798e-05,
      "loss": 0.9926,
      "step": 1740
    },
    {
      "epoch": 7.036909090909091,
      "grad_norm": 26.939453125,
      "learning_rate": 3.787878787878788e-05,
      "loss": 0.8222,
      "step": 1750
    },
    {
      "epoch": 7.0387272727272725,
      "grad_norm": 10.071974754333496,
      "learning_rate": 3.777777777777778e-05,
      "loss": 1.0569,
      "step": 1760
    },
    {
      "epoch": 7.040181818181818,
      "eval_accuracy": 0.4126984126984127,
      "eval_loss": 1.107624888420105,
      "eval_runtime": 198.1727,
      "eval_samples_per_second": 0.954,
      "eval_steps_per_second": 0.242,
      "step": 1768
    },
    {
      "epoch": 8.000363636363636,
      "grad_norm": 26.684202194213867,
      "learning_rate": 3.767676767676768e-05,
      "loss": 0.9988,
      "step": 1770
    },
    {
      "epoch": 8.002181818181818,
      "grad_norm": 18.15027618408203,
      "learning_rate": 3.757575757575758e-05,
      "loss": 1.0715,
      "step": 1780
    },
    {
      "epoch": 8.004,
      "grad_norm": 28.486724853515625,
      "learning_rate": 3.747474747474748e-05,
      "loss": 0.8121,
      "step": 1790
    },
    {
      "epoch": 8.005818181818182,
      "grad_norm": 303.9952392578125,
      "learning_rate": 3.7373737373737376e-05,
      "loss": 0.9878,
      "step": 1800
    },
    {
      "epoch": 8.007636363636363,
      "grad_norm": 14.011321067810059,
      "learning_rate": 3.7272727272727276e-05,
      "loss": 1.1017,
      "step": 1810
    },
    {
      "epoch": 8.009454545454545,
      "grad_norm": 22.236854553222656,
      "learning_rate": 3.7171717171717175e-05,
      "loss": 1.0964,
      "step": 1820
    },
    {
      "epoch": 8.011272727272727,
      "grad_norm": 10.44234561920166,
      "learning_rate": 3.7070707070707075e-05,
      "loss": 0.6585,
      "step": 1830
    },
    {
      "epoch": 8.01309090909091,
      "grad_norm": 8.875130653381348,
      "learning_rate": 3.6969696969696974e-05,
      "loss": 0.8471,
      "step": 1840
    },
    {
      "epoch": 8.014909090909091,
      "grad_norm": 10.0613374710083,
      "learning_rate": 3.686868686868687e-05,
      "loss": 0.9548,
      "step": 1850
    },
    {
      "epoch": 8.016727272727273,
      "grad_norm": 24.469505310058594,
      "learning_rate": 3.6767676767676766e-05,
      "loss": 1.2708,
      "step": 1860
    },
    {
      "epoch": 8.018545454545455,
      "grad_norm": 14.466684341430664,
      "learning_rate": 3.6666666666666666e-05,
      "loss": 1.0558,
      "step": 1870
    },
    {
      "epoch": 8.020363636363637,
      "grad_norm": 8.436592102050781,
      "learning_rate": 3.656565656565657e-05,
      "loss": 1.0655,
      "step": 1880
    },
    {
      "epoch": 8.022181818181819,
      "grad_norm": 11.807909965515137,
      "learning_rate": 3.6464646464646465e-05,
      "loss": 1.0226,
      "step": 1890
    },
    {
      "epoch": 8.024,
      "grad_norm": 7.324114799499512,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 1.0769,
      "step": 1900
    },
    {
      "epoch": 8.025818181818181,
      "grad_norm": 7.894458293914795,
      "learning_rate": 3.6262626262626264e-05,
      "loss": 0.9414,
      "step": 1910
    },
    {
      "epoch": 8.027636363636363,
      "grad_norm": 9.577462196350098,
      "learning_rate": 3.616161616161616e-05,
      "loss": 0.9558,
      "step": 1920
    },
    {
      "epoch": 8.029454545454545,
      "grad_norm": 13.144161224365234,
      "learning_rate": 3.606060606060606e-05,
      "loss": 1.0216,
      "step": 1930
    },
    {
      "epoch": 8.031272727272727,
      "grad_norm": 6.9937357902526855,
      "learning_rate": 3.595959595959596e-05,
      "loss": 0.7935,
      "step": 1940
    },
    {
      "epoch": 8.033090909090909,
      "grad_norm": 27.100025177001953,
      "learning_rate": 3.5858585858585855e-05,
      "loss": 0.8731,
      "step": 1950
    },
    {
      "epoch": 8.03490909090909,
      "grad_norm": 71.13379669189453,
      "learning_rate": 3.575757575757576e-05,
      "loss": 0.9298,
      "step": 1960
    },
    {
      "epoch": 8.036727272727273,
      "grad_norm": 118.5480728149414,
      "learning_rate": 3.565656565656566e-05,
      "loss": 1.3363,
      "step": 1970
    },
    {
      "epoch": 8.038545454545455,
      "grad_norm": 8.516532897949219,
      "learning_rate": 3.555555555555556e-05,
      "loss": 0.9919,
      "step": 1980
    },
    {
      "epoch": 8.040181818181818,
      "eval_accuracy": 0.5608465608465608,
      "eval_loss": 0.9307597875595093,
      "eval_runtime": 199.9286,
      "eval_samples_per_second": 0.945,
      "eval_steps_per_second": 0.24,
      "step": 1989
    },
    {
      "epoch": 9.000181818181819,
      "grad_norm": 18.11604881286621,
      "learning_rate": 3.545454545454546e-05,
      "loss": 0.9344,
      "step": 1990
    },
    {
      "epoch": 9.002,
      "grad_norm": 6.241488456726074,
      "learning_rate": 3.535353535353535e-05,
      "loss": 0.9527,
      "step": 2000
    },
    {
      "epoch": 9.003818181818183,
      "grad_norm": 13.090892791748047,
      "learning_rate": 3.525252525252525e-05,
      "loss": 0.6875,
      "step": 2010
    },
    {
      "epoch": 9.005636363636363,
      "grad_norm": 48.75657653808594,
      "learning_rate": 3.515151515151515e-05,
      "loss": 0.9867,
      "step": 2020
    },
    {
      "epoch": 9.007454545454545,
      "grad_norm": 21.59893226623535,
      "learning_rate": 3.505050505050505e-05,
      "loss": 1.3935,
      "step": 2030
    },
    {
      "epoch": 9.009272727272727,
      "grad_norm": 31.900312423706055,
      "learning_rate": 3.494949494949495e-05,
      "loss": 0.7955,
      "step": 2040
    },
    {
      "epoch": 9.011090909090909,
      "grad_norm": 21.120643615722656,
      "learning_rate": 3.484848484848485e-05,
      "loss": 1.1775,
      "step": 2050
    },
    {
      "epoch": 9.01290909090909,
      "grad_norm": 21.606521606445312,
      "learning_rate": 3.474747474747475e-05,
      "loss": 0.7836,
      "step": 2060
    },
    {
      "epoch": 9.014727272727272,
      "grad_norm": 23.555465698242188,
      "learning_rate": 3.464646464646465e-05,
      "loss": 1.0021,
      "step": 2070
    },
    {
      "epoch": 9.016545454545454,
      "grad_norm": 8.791067123413086,
      "learning_rate": 3.454545454545455e-05,
      "loss": 0.8174,
      "step": 2080
    },
    {
      "epoch": 9.018363636363636,
      "grad_norm": 9.401249885559082,
      "learning_rate": 3.444444444444445e-05,
      "loss": 0.9594,
      "step": 2090
    },
    {
      "epoch": 9.020181818181818,
      "grad_norm": 30.353981018066406,
      "learning_rate": 3.434343434343435e-05,
      "loss": 0.9275,
      "step": 2100
    },
    {
      "epoch": 9.022,
      "grad_norm": 14.411357879638672,
      "learning_rate": 3.424242424242424e-05,
      "loss": 0.808,
      "step": 2110
    },
    {
      "epoch": 9.023818181818182,
      "grad_norm": 27.9307861328125,
      "learning_rate": 3.414141414141414e-05,
      "loss": 0.8911,
      "step": 2120
    },
    {
      "epoch": 9.025636363636364,
      "grad_norm": 48.075042724609375,
      "learning_rate": 3.4040404040404045e-05,
      "loss": 1.0519,
      "step": 2130
    },
    {
      "epoch": 9.027454545454546,
      "grad_norm": 24.737884521484375,
      "learning_rate": 3.3939393939393945e-05,
      "loss": 0.8945,
      "step": 2140
    },
    {
      "epoch": 9.029272727272728,
      "grad_norm": 5.976382255554199,
      "learning_rate": 3.3838383838383844e-05,
      "loss": 0.8602,
      "step": 2150
    },
    {
      "epoch": 9.03109090909091,
      "grad_norm": 54.03303146362305,
      "learning_rate": 3.373737373737374e-05,
      "loss": 1.1649,
      "step": 2160
    },
    {
      "epoch": 9.03290909090909,
      "grad_norm": 12.576547622680664,
      "learning_rate": 3.3636363636363636e-05,
      "loss": 1.1853,
      "step": 2170
    },
    {
      "epoch": 9.034727272727272,
      "grad_norm": 12.94636344909668,
      "learning_rate": 3.3535353535353536e-05,
      "loss": 0.9627,
      "step": 2180
    },
    {
      "epoch": 9.036545454545454,
      "grad_norm": 80.18585968017578,
      "learning_rate": 3.3434343434343435e-05,
      "loss": 0.9442,
      "step": 2190
    },
    {
      "epoch": 9.038363636363636,
      "grad_norm": 34.63486862182617,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 1.1025,
      "step": 2200
    },
    {
      "epoch": 9.040181818181818,
      "grad_norm": 21.734012603759766,
      "learning_rate": 3.3232323232323234e-05,
      "loss": 0.7163,
      "step": 2210
    },
    {
      "epoch": 9.040181818181818,
      "eval_accuracy": 0.5185185185185185,
      "eval_loss": 0.9721111059188843,
      "eval_runtime": 188.9985,
      "eval_samples_per_second": 1.0,
      "eval_steps_per_second": 0.254,
      "step": 2210
    },
    {
      "epoch": 10.001818181818182,
      "grad_norm": 22.439699172973633,
      "learning_rate": 3.3131313131313134e-05,
      "loss": 1.1736,
      "step": 2220
    },
    {
      "epoch": 10.003636363636364,
      "grad_norm": 9.710022926330566,
      "learning_rate": 3.303030303030303e-05,
      "loss": 0.9365,
      "step": 2230
    },
    {
      "epoch": 10.005454545454546,
      "grad_norm": 18.069990158081055,
      "learning_rate": 3.292929292929293e-05,
      "loss": 0.9009,
      "step": 2240
    },
    {
      "epoch": 10.007272727272728,
      "grad_norm": 40.750938415527344,
      "learning_rate": 3.282828282828283e-05,
      "loss": 1.1381,
      "step": 2250
    },
    {
      "epoch": 10.00909090909091,
      "grad_norm": 13.775208473205566,
      "learning_rate": 3.272727272727273e-05,
      "loss": 1.022,
      "step": 2260
    },
    {
      "epoch": 10.010909090909092,
      "grad_norm": 14.381951332092285,
      "learning_rate": 3.2626262626262624e-05,
      "loss": 0.9232,
      "step": 2270
    },
    {
      "epoch": 10.012727272727274,
      "grad_norm": 8.365996360778809,
      "learning_rate": 3.2525252525252524e-05,
      "loss": 0.9186,
      "step": 2280
    },
    {
      "epoch": 10.014545454545454,
      "grad_norm": 20.199583053588867,
      "learning_rate": 3.2424242424242423e-05,
      "loss": 0.9464,
      "step": 2290
    },
    {
      "epoch": 10.016363636363636,
      "grad_norm": 17.12706756591797,
      "learning_rate": 3.232323232323233e-05,
      "loss": 0.818,
      "step": 2300
    },
    {
      "epoch": 10.018181818181818,
      "grad_norm": 15.488892555236816,
      "learning_rate": 3.222222222222223e-05,
      "loss": 0.9192,
      "step": 2310
    },
    {
      "epoch": 10.02,
      "grad_norm": 21.11297607421875,
      "learning_rate": 3.212121212121212e-05,
      "loss": 0.9366,
      "step": 2320
    },
    {
      "epoch": 10.021818181818182,
      "grad_norm": 14.04183292388916,
      "learning_rate": 3.202020202020202e-05,
      "loss": 0.9159,
      "step": 2330
    },
    {
      "epoch": 10.023636363636363,
      "grad_norm": 24.148868560791016,
      "learning_rate": 3.191919191919192e-05,
      "loss": 0.7416,
      "step": 2340
    },
    {
      "epoch": 10.025454545454545,
      "grad_norm": 122.17024993896484,
      "learning_rate": 3.181818181818182e-05,
      "loss": 1.0963,
      "step": 2350
    },
    {
      "epoch": 10.027272727272727,
      "grad_norm": 7.737703800201416,
      "learning_rate": 3.171717171717172e-05,
      "loss": 0.8124,
      "step": 2360
    },
    {
      "epoch": 10.02909090909091,
      "grad_norm": 7.881679534912109,
      "learning_rate": 3.161616161616161e-05,
      "loss": 0.8641,
      "step": 2370
    },
    {
      "epoch": 10.030909090909091,
      "grad_norm": 26.089942932128906,
      "learning_rate": 3.151515151515151e-05,
      "loss": 0.8965,
      "step": 2380
    },
    {
      "epoch": 10.032727272727273,
      "grad_norm": 11.208564758300781,
      "learning_rate": 3.141414141414142e-05,
      "loss": 0.8305,
      "step": 2390
    },
    {
      "epoch": 10.034545454545455,
      "grad_norm": 18.578123092651367,
      "learning_rate": 3.131313131313132e-05,
      "loss": 1.0483,
      "step": 2400
    },
    {
      "epoch": 10.036363636363637,
      "grad_norm": 180.40223693847656,
      "learning_rate": 3.121212121212122e-05,
      "loss": 0.9592,
      "step": 2410
    },
    {
      "epoch": 10.038181818181819,
      "grad_norm": 8.062999725341797,
      "learning_rate": 3.111111111111111e-05,
      "loss": 0.8835,
      "step": 2420
    },
    {
      "epoch": 10.04,
      "grad_norm": 27.839271545410156,
      "learning_rate": 3.101010101010101e-05,
      "loss": 0.7941,
      "step": 2430
    },
    {
      "epoch": 10.040181818181818,
      "eval_accuracy": 0.5396825396825397,
      "eval_loss": 0.9407670497894287,
      "eval_runtime": 198.3627,
      "eval_samples_per_second": 0.953,
      "eval_steps_per_second": 0.242,
      "step": 2431
    },
    {
      "epoch": 11.001636363636363,
      "grad_norm": 38.953304290771484,
      "learning_rate": 3.090909090909091e-05,
      "loss": 0.8673,
      "step": 2440
    },
    {
      "epoch": 11.003454545454545,
      "grad_norm": 81.81233215332031,
      "learning_rate": 3.080808080808081e-05,
      "loss": 1.0405,
      "step": 2450
    },
    {
      "epoch": 11.005272727272727,
      "grad_norm": 35.105220794677734,
      "learning_rate": 3.070707070707071e-05,
      "loss": 0.7739,
      "step": 2460
    },
    {
      "epoch": 11.007090909090909,
      "grad_norm": 109.05673217773438,
      "learning_rate": 3.060606060606061e-05,
      "loss": 1.4651,
      "step": 2470
    },
    {
      "epoch": 11.008909090909091,
      "grad_norm": 8.228311538696289,
      "learning_rate": 3.050505050505051e-05,
      "loss": 0.9487,
      "step": 2480
    },
    {
      "epoch": 11.010727272727273,
      "grad_norm": 14.025703430175781,
      "learning_rate": 3.0404040404040406e-05,
      "loss": 0.9334,
      "step": 2490
    },
    {
      "epoch": 11.012545454545455,
      "grad_norm": 34.061092376708984,
      "learning_rate": 3.0303030303030306e-05,
      "loss": 1.045,
      "step": 2500
    },
    {
      "epoch": 11.014363636363637,
      "grad_norm": 20.992385864257812,
      "learning_rate": 3.0202020202020205e-05,
      "loss": 1.0003,
      "step": 2510
    },
    {
      "epoch": 11.016181818181819,
      "grad_norm": 28.905166625976562,
      "learning_rate": 3.01010101010101e-05,
      "loss": 0.7843,
      "step": 2520
    },
    {
      "epoch": 11.018,
      "grad_norm": 14.449548721313477,
      "learning_rate": 3e-05,
      "loss": 1.1539,
      "step": 2530
    },
    {
      "epoch": 11.019818181818183,
      "grad_norm": 178.1973876953125,
      "learning_rate": 2.98989898989899e-05,
      "loss": 1.1367,
      "step": 2540
    },
    {
      "epoch": 11.021636363636363,
      "grad_norm": 31.946022033691406,
      "learning_rate": 2.9797979797979796e-05,
      "loss": 0.832,
      "step": 2550
    },
    {
      "epoch": 11.023454545454545,
      "grad_norm": 7.3684468269348145,
      "learning_rate": 2.96969696969697e-05,
      "loss": 0.9064,
      "step": 2560
    },
    {
      "epoch": 11.025272727272727,
      "grad_norm": 29.028446197509766,
      "learning_rate": 2.95959595959596e-05,
      "loss": 0.9373,
      "step": 2570
    },
    {
      "epoch": 11.027090909090909,
      "grad_norm": 27.086463928222656,
      "learning_rate": 2.9494949494949498e-05,
      "loss": 0.8377,
      "step": 2580
    },
    {
      "epoch": 11.02890909090909,
      "grad_norm": 14.71929931640625,
      "learning_rate": 2.9393939393939394e-05,
      "loss": 0.9033,
      "step": 2590
    },
    {
      "epoch": 11.030727272727272,
      "grad_norm": 12.47122859954834,
      "learning_rate": 2.9292929292929294e-05,
      "loss": 1.0606,
      "step": 2600
    },
    {
      "epoch": 11.032545454545454,
      "grad_norm": 16.485416412353516,
      "learning_rate": 2.9191919191919193e-05,
      "loss": 1.1198,
      "step": 2610
    },
    {
      "epoch": 11.034363636363636,
      "grad_norm": 6.706732273101807,
      "learning_rate": 2.909090909090909e-05,
      "loss": 0.7835,
      "step": 2620
    },
    {
      "epoch": 11.036181818181818,
      "grad_norm": 44.958946228027344,
      "learning_rate": 2.898989898989899e-05,
      "loss": 0.9495,
      "step": 2630
    },
    {
      "epoch": 11.038,
      "grad_norm": 11.914873123168945,
      "learning_rate": 2.8888888888888888e-05,
      "loss": 0.7911,
      "step": 2640
    },
    {
      "epoch": 11.039818181818182,
      "grad_norm": 21.188133239746094,
      "learning_rate": 2.878787878787879e-05,
      "loss": 0.9386,
      "step": 2650
    },
    {
      "epoch": 11.040181818181818,
      "eval_accuracy": 0.5925925925925926,
      "eval_loss": 0.8729740381240845,
      "eval_runtime": 192.0496,
      "eval_samples_per_second": 0.984,
      "eval_steps_per_second": 0.25,
      "step": 2652
    },
    {
      "epoch": 12.001454545454546,
      "grad_norm": 8.996429443359375,
      "learning_rate": 2.868686868686869e-05,
      "loss": 0.7471,
      "step": 2660
    },
    {
      "epoch": 12.003272727272726,
      "grad_norm": 7.39767599105835,
      "learning_rate": 2.8585858585858587e-05,
      "loss": 0.9742,
      "step": 2670
    },
    {
      "epoch": 12.005090909090908,
      "grad_norm": 23.08453369140625,
      "learning_rate": 2.8484848484848486e-05,
      "loss": 1.022,
      "step": 2680
    },
    {
      "epoch": 12.00690909090909,
      "grad_norm": 22.011754989624023,
      "learning_rate": 2.8383838383838386e-05,
      "loss": 0.9812,
      "step": 2690
    },
    {
      "epoch": 12.008727272727272,
      "grad_norm": 14.558695793151855,
      "learning_rate": 2.8282828282828282e-05,
      "loss": 0.8475,
      "step": 2700
    },
    {
      "epoch": 12.010545454545454,
      "grad_norm": 23.893945693969727,
      "learning_rate": 2.818181818181818e-05,
      "loss": 0.7818,
      "step": 2710
    },
    {
      "epoch": 12.012363636363636,
      "grad_norm": 28.504093170166016,
      "learning_rate": 2.808080808080808e-05,
      "loss": 0.8586,
      "step": 2720
    },
    {
      "epoch": 12.014181818181818,
      "grad_norm": 29.902450561523438,
      "learning_rate": 2.7979797979797984e-05,
      "loss": 0.9901,
      "step": 2730
    },
    {
      "epoch": 12.016,
      "grad_norm": 8.041543960571289,
      "learning_rate": 2.7878787878787883e-05,
      "loss": 0.7766,
      "step": 2740
    },
    {
      "epoch": 12.017818181818182,
      "grad_norm": 19.797578811645508,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.1183,
      "step": 2750
    },
    {
      "epoch": 12.019636363636364,
      "grad_norm": 24.141660690307617,
      "learning_rate": 2.767676767676768e-05,
      "loss": 0.7957,
      "step": 2760
    },
    {
      "epoch": 12.021454545454546,
      "grad_norm": 14.93819522857666,
      "learning_rate": 2.7575757575757578e-05,
      "loss": 0.7143,
      "step": 2770
    },
    {
      "epoch": 12.023272727272728,
      "grad_norm": 34.424285888671875,
      "learning_rate": 2.7474747474747474e-05,
      "loss": 0.8015,
      "step": 2780
    },
    {
      "epoch": 12.02509090909091,
      "grad_norm": 14.072427749633789,
      "learning_rate": 2.7373737373737374e-05,
      "loss": 0.8037,
      "step": 2790
    },
    {
      "epoch": 12.026909090909092,
      "grad_norm": 42.77705764770508,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 1.0436,
      "step": 2800
    },
    {
      "epoch": 12.028727272727274,
      "grad_norm": 62.84910583496094,
      "learning_rate": 2.717171717171717e-05,
      "loss": 0.7561,
      "step": 2810
    },
    {
      "epoch": 12.030545454545454,
      "grad_norm": 29.13556671142578,
      "learning_rate": 2.7070707070707075e-05,
      "loss": 1.0289,
      "step": 2820
    },
    {
      "epoch": 12.032363636363636,
      "grad_norm": 13.636876106262207,
      "learning_rate": 2.696969696969697e-05,
      "loss": 0.8254,
      "step": 2830
    },
    {
      "epoch": 12.034181818181818,
      "grad_norm": 24.48445701599121,
      "learning_rate": 2.686868686868687e-05,
      "loss": 0.7901,
      "step": 2840
    },
    {
      "epoch": 12.036,
      "grad_norm": 25.308664321899414,
      "learning_rate": 2.676767676767677e-05,
      "loss": 0.859,
      "step": 2850
    },
    {
      "epoch": 12.037818181818182,
      "grad_norm": 30.3497371673584,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 1.0236,
      "step": 2860
    },
    {
      "epoch": 12.039636363636363,
      "grad_norm": 6.4086198806762695,
      "learning_rate": 2.6565656565656566e-05,
      "loss": 0.857,
      "step": 2870
    },
    {
      "epoch": 12.040181818181818,
      "eval_accuracy": 0.656084656084656,
      "eval_loss": 0.8833221793174744,
      "eval_runtime": 208.1476,
      "eval_samples_per_second": 0.908,
      "eval_steps_per_second": 0.231,
      "step": 2873
    },
    {
      "epoch": 13.001272727272728,
      "grad_norm": 30.714641571044922,
      "learning_rate": 2.6464646464646466e-05,
      "loss": 0.9611,
      "step": 2880
    },
    {
      "epoch": 13.00309090909091,
      "grad_norm": 6.578027248382568,
      "learning_rate": 2.636363636363636e-05,
      "loss": 0.9732,
      "step": 2890
    },
    {
      "epoch": 13.004909090909091,
      "grad_norm": 31.625444412231445,
      "learning_rate": 2.6262626262626268e-05,
      "loss": 0.8233,
      "step": 2900
    },
    {
      "epoch": 13.006727272727273,
      "grad_norm": 10.502728462219238,
      "learning_rate": 2.6161616161616164e-05,
      "loss": 0.8193,
      "step": 2910
    },
    {
      "epoch": 13.008545454545455,
      "grad_norm": 7.814911842346191,
      "learning_rate": 2.6060606060606063e-05,
      "loss": 1.0234,
      "step": 2920
    },
    {
      "epoch": 13.010363636363637,
      "grad_norm": 7.037054061889648,
      "learning_rate": 2.5959595959595963e-05,
      "loss": 1.0141,
      "step": 2930
    },
    {
      "epoch": 13.012181818181817,
      "grad_norm": 73.73834991455078,
      "learning_rate": 2.585858585858586e-05,
      "loss": 1.2231,
      "step": 2940
    },
    {
      "epoch": 13.014,
      "grad_norm": 47.82062911987305,
      "learning_rate": 2.575757575757576e-05,
      "loss": 0.9928,
      "step": 2950
    },
    {
      "epoch": 13.015818181818181,
      "grad_norm": 10.504758834838867,
      "learning_rate": 2.5656565656565658e-05,
      "loss": 0.8383,
      "step": 2960
    },
    {
      "epoch": 13.017636363636363,
      "grad_norm": 43.19855880737305,
      "learning_rate": 2.5555555555555554e-05,
      "loss": 0.7892,
      "step": 2970
    },
    {
      "epoch": 13.019454545454545,
      "grad_norm": 15.07066822052002,
      "learning_rate": 2.5454545454545454e-05,
      "loss": 0.7949,
      "step": 2980
    },
    {
      "epoch": 13.021272727272727,
      "grad_norm": 13.859105110168457,
      "learning_rate": 2.5353535353535356e-05,
      "loss": 0.896,
      "step": 2990
    },
    {
      "epoch": 13.023090909090909,
      "grad_norm": 6.099465370178223,
      "learning_rate": 2.5252525252525256e-05,
      "loss": 0.813,
      "step": 3000
    },
    {
      "epoch": 13.024909090909091,
      "grad_norm": 25.261594772338867,
      "learning_rate": 2.5151515151515155e-05,
      "loss": 0.7597,
      "step": 3010
    },
    {
      "epoch": 13.026727272727273,
      "grad_norm": 9.865255355834961,
      "learning_rate": 2.505050505050505e-05,
      "loss": 0.885,
      "step": 3020
    },
    {
      "epoch": 13.028545454545455,
      "grad_norm": 22.643348693847656,
      "learning_rate": 2.494949494949495e-05,
      "loss": 0.9914,
      "step": 3030
    },
    {
      "epoch": 13.030363636363637,
      "grad_norm": 23.875263214111328,
      "learning_rate": 2.4848484848484847e-05,
      "loss": 0.6312,
      "step": 3040
    },
    {
      "epoch": 13.032181818181819,
      "grad_norm": 18.725353240966797,
      "learning_rate": 2.474747474747475e-05,
      "loss": 0.787,
      "step": 3050
    },
    {
      "epoch": 13.034,
      "grad_norm": 30.778732299804688,
      "learning_rate": 2.464646464646465e-05,
      "loss": 1.0014,
      "step": 3060
    },
    {
      "epoch": 13.035818181818183,
      "grad_norm": 42.08274459838867,
      "learning_rate": 2.4545454545454545e-05,
      "loss": 0.9,
      "step": 3070
    },
    {
      "epoch": 13.037636363636363,
      "grad_norm": 13.115510940551758,
      "learning_rate": 2.4444444444444445e-05,
      "loss": 0.8936,
      "step": 3080
    },
    {
      "epoch": 13.039454545454545,
      "grad_norm": 40.42136001586914,
      "learning_rate": 2.4343434343434344e-05,
      "loss": 0.7059,
      "step": 3090
    },
    {
      "epoch": 13.040181818181818,
      "eval_accuracy": 0.6402116402116402,
      "eval_loss": 0.8672856092453003,
      "eval_runtime": 196.9971,
      "eval_samples_per_second": 0.959,
      "eval_steps_per_second": 0.244,
      "step": 3094
    },
    {
      "epoch": 14.001090909090909,
      "grad_norm": 13.292407989501953,
      "learning_rate": 2.4242424242424244e-05,
      "loss": 0.9836,
      "step": 3100
    },
    {
      "epoch": 14.00290909090909,
      "grad_norm": 34.00928497314453,
      "learning_rate": 2.4141414141414143e-05,
      "loss": 0.8808,
      "step": 3110
    },
    {
      "epoch": 14.004727272727273,
      "grad_norm": 10.223441123962402,
      "learning_rate": 2.404040404040404e-05,
      "loss": 0.7724,
      "step": 3120
    },
    {
      "epoch": 14.006545454545455,
      "grad_norm": 52.84186553955078,
      "learning_rate": 2.393939393939394e-05,
      "loss": 0.9767,
      "step": 3130
    },
    {
      "epoch": 14.008363636363637,
      "grad_norm": 41.73188781738281,
      "learning_rate": 2.3838383838383842e-05,
      "loss": 1.1687,
      "step": 3140
    },
    {
      "epoch": 14.010181818181819,
      "grad_norm": 11.622649192810059,
      "learning_rate": 2.3737373737373738e-05,
      "loss": 1.0479,
      "step": 3150
    },
    {
      "epoch": 14.012,
      "grad_norm": 18.575862884521484,
      "learning_rate": 2.3636363636363637e-05,
      "loss": 0.9784,
      "step": 3160
    },
    {
      "epoch": 14.013818181818182,
      "grad_norm": 4.266345500946045,
      "learning_rate": 2.3535353535353537e-05,
      "loss": 0.8186,
      "step": 3170
    },
    {
      "epoch": 14.015636363636364,
      "grad_norm": 41.91463088989258,
      "learning_rate": 2.3434343434343436e-05,
      "loss": 0.821,
      "step": 3180
    },
    {
      "epoch": 14.017454545454546,
      "grad_norm": 22.115848541259766,
      "learning_rate": 2.3333333333333336e-05,
      "loss": 0.8563,
      "step": 3190
    },
    {
      "epoch": 14.019272727272726,
      "grad_norm": 15.369574546813965,
      "learning_rate": 2.3232323232323232e-05,
      "loss": 0.9562,
      "step": 3200
    },
    {
      "epoch": 14.021090909090908,
      "grad_norm": 35.73291778564453,
      "learning_rate": 2.313131313131313e-05,
      "loss": 0.7987,
      "step": 3210
    },
    {
      "epoch": 14.02290909090909,
      "grad_norm": 13.289374351501465,
      "learning_rate": 2.3030303030303034e-05,
      "loss": 0.953,
      "step": 3220
    },
    {
      "epoch": 14.024727272727272,
      "grad_norm": 9.4387845993042,
      "learning_rate": 2.292929292929293e-05,
      "loss": 0.7263,
      "step": 3230
    },
    {
      "epoch": 14.026545454545454,
      "grad_norm": 19.509065628051758,
      "learning_rate": 2.282828282828283e-05,
      "loss": 0.6561,
      "step": 3240
    },
    {
      "epoch": 14.028363636363636,
      "grad_norm": 11.643479347229004,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.9432,
      "step": 3250
    },
    {
      "epoch": 14.030181818181818,
      "grad_norm": 30.66929817199707,
      "learning_rate": 2.262626262626263e-05,
      "loss": 0.9817,
      "step": 3260
    },
    {
      "epoch": 14.032,
      "grad_norm": 25.15669059753418,
      "learning_rate": 2.2525252525252528e-05,
      "loss": 0.9076,
      "step": 3270
    },
    {
      "epoch": 14.033818181818182,
      "grad_norm": 15.897071838378906,
      "learning_rate": 2.2424242424242424e-05,
      "loss": 0.9879,
      "step": 3280
    },
    {
      "epoch": 14.035636363636364,
      "grad_norm": 25.583818435668945,
      "learning_rate": 2.2323232323232324e-05,
      "loss": 0.8784,
      "step": 3290
    },
    {
      "epoch": 14.037454545454546,
      "grad_norm": 21.015098571777344,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.7533,
      "step": 3300
    },
    {
      "epoch": 14.039272727272728,
      "grad_norm": 23.844053268432617,
      "learning_rate": 2.2121212121212123e-05,
      "loss": 0.8322,
      "step": 3310
    },
    {
      "epoch": 14.040181818181818,
      "eval_accuracy": 0.6296296296296297,
      "eval_loss": 0.8232609033584595,
      "eval_runtime": 190.437,
      "eval_samples_per_second": 0.992,
      "eval_steps_per_second": 0.252,
      "step": 3315
    },
    {
      "epoch": 15.00090909090909,
      "grad_norm": 9.514819145202637,
      "learning_rate": 2.2020202020202022e-05,
      "loss": 1.0237,
      "step": 3320
    },
    {
      "epoch": 15.002727272727272,
      "grad_norm": 57.46669387817383,
      "learning_rate": 2.191919191919192e-05,
      "loss": 0.8501,
      "step": 3330
    },
    {
      "epoch": 15.004545454545454,
      "grad_norm": 77.95063018798828,
      "learning_rate": 2.1818181818181818e-05,
      "loss": 0.858,
      "step": 3340
    },
    {
      "epoch": 15.006363636363636,
      "grad_norm": 18.01690673828125,
      "learning_rate": 2.171717171717172e-05,
      "loss": 0.8786,
      "step": 3350
    },
    {
      "epoch": 15.008181818181818,
      "grad_norm": 20.38191795349121,
      "learning_rate": 2.1616161616161617e-05,
      "loss": 0.9039,
      "step": 3360
    },
    {
      "epoch": 15.01,
      "grad_norm": 41.25912857055664,
      "learning_rate": 2.1515151515151516e-05,
      "loss": 1.1472,
      "step": 3370
    },
    {
      "epoch": 15.011818181818182,
      "grad_norm": 11.111079216003418,
      "learning_rate": 2.1414141414141416e-05,
      "loss": 0.7277,
      "step": 3380
    },
    {
      "epoch": 15.013636363636364,
      "grad_norm": 17.063514709472656,
      "learning_rate": 2.1313131313131315e-05,
      "loss": 0.7826,
      "step": 3390
    },
    {
      "epoch": 15.015454545454546,
      "grad_norm": 10.764243125915527,
      "learning_rate": 2.1212121212121215e-05,
      "loss": 0.743,
      "step": 3400
    },
    {
      "epoch": 15.017272727272728,
      "grad_norm": 5.325133800506592,
      "learning_rate": 2.111111111111111e-05,
      "loss": 0.6667,
      "step": 3410
    },
    {
      "epoch": 15.01909090909091,
      "grad_norm": 12.80745792388916,
      "learning_rate": 2.101010101010101e-05,
      "loss": 0.8126,
      "step": 3420
    },
    {
      "epoch": 15.020909090909091,
      "grad_norm": 20.577905654907227,
      "learning_rate": 2.090909090909091e-05,
      "loss": 0.8996,
      "step": 3430
    },
    {
      "epoch": 15.022727272727273,
      "grad_norm": 17.238866806030273,
      "learning_rate": 2.080808080808081e-05,
      "loss": 0.7257,
      "step": 3440
    },
    {
      "epoch": 15.024545454545455,
      "grad_norm": 11.546747207641602,
      "learning_rate": 2.070707070707071e-05,
      "loss": 0.8113,
      "step": 3450
    },
    {
      "epoch": 15.026363636363635,
      "grad_norm": 19.22821807861328,
      "learning_rate": 2.0606060606060608e-05,
      "loss": 0.9306,
      "step": 3460
    },
    {
      "epoch": 15.028181818181817,
      "grad_norm": 8.0140380859375,
      "learning_rate": 2.0505050505050504e-05,
      "loss": 0.7893,
      "step": 3470
    },
    {
      "epoch": 15.03,
      "grad_norm": 159.8179931640625,
      "learning_rate": 2.0404040404040407e-05,
      "loss": 0.6769,
      "step": 3480
    },
    {
      "epoch": 15.031818181818181,
      "grad_norm": 57.692176818847656,
      "learning_rate": 2.0303030303030303e-05,
      "loss": 0.8511,
      "step": 3490
    },
    {
      "epoch": 15.033636363636363,
      "grad_norm": 17.908693313598633,
      "learning_rate": 2.0202020202020203e-05,
      "loss": 1.0586,
      "step": 3500
    },
    {
      "epoch": 15.035454545454545,
      "grad_norm": 7.324565410614014,
      "learning_rate": 2.0101010101010102e-05,
      "loss": 0.7331,
      "step": 3510
    },
    {
      "epoch": 15.037272727272727,
      "grad_norm": 18.01776123046875,
      "learning_rate": 2e-05,
      "loss": 0.8703,
      "step": 3520
    },
    {
      "epoch": 15.039090909090909,
      "grad_norm": 14.08636474609375,
      "learning_rate": 1.98989898989899e-05,
      "loss": 0.8574,
      "step": 3530
    },
    {
      "epoch": 15.040181818181818,
      "eval_accuracy": 0.6825396825396826,
      "eval_loss": 0.7642742395401001,
      "eval_runtime": 195.2966,
      "eval_samples_per_second": 0.968,
      "eval_steps_per_second": 0.246,
      "step": 3536
    },
    {
      "epoch": 16.00072727272727,
      "grad_norm": 8.01131820678711,
      "learning_rate": 1.9797979797979797e-05,
      "loss": 0.6872,
      "step": 3540
    },
    {
      "epoch": 16.002545454545455,
      "grad_norm": 83.17420196533203,
      "learning_rate": 1.9696969696969697e-05,
      "loss": 0.9136,
      "step": 3550
    },
    {
      "epoch": 16.004363636363635,
      "grad_norm": 16.785919189453125,
      "learning_rate": 1.95959595959596e-05,
      "loss": 1.0454,
      "step": 3560
    },
    {
      "epoch": 16.00618181818182,
      "grad_norm": 14.56745719909668,
      "learning_rate": 1.9494949494949496e-05,
      "loss": 0.8476,
      "step": 3570
    },
    {
      "epoch": 16.008,
      "grad_norm": 15.242620468139648,
      "learning_rate": 1.9393939393939395e-05,
      "loss": 0.8815,
      "step": 3580
    },
    {
      "epoch": 16.009818181818183,
      "grad_norm": 13.342121124267578,
      "learning_rate": 1.9292929292929295e-05,
      "loss": 1.018,
      "step": 3590
    },
    {
      "epoch": 16.011636363636363,
      "grad_norm": 16.53753662109375,
      "learning_rate": 1.919191919191919e-05,
      "loss": 0.8287,
      "step": 3600
    },
    {
      "epoch": 16.013454545454547,
      "grad_norm": 13.803275108337402,
      "learning_rate": 1.9090909090909094e-05,
      "loss": 1.0207,
      "step": 3610
    },
    {
      "epoch": 16.015272727272727,
      "grad_norm": 17.9733943939209,
      "learning_rate": 1.898989898989899e-05,
      "loss": 0.7543,
      "step": 3620
    },
    {
      "epoch": 16.01709090909091,
      "grad_norm": 24.71270179748535,
      "learning_rate": 1.888888888888889e-05,
      "loss": 0.7426,
      "step": 3630
    },
    {
      "epoch": 16.01890909090909,
      "grad_norm": 74.74566650390625,
      "learning_rate": 1.878787878787879e-05,
      "loss": 0.6937,
      "step": 3640
    },
    {
      "epoch": 16.020727272727274,
      "grad_norm": 16.248302459716797,
      "learning_rate": 1.8686868686868688e-05,
      "loss": 0.8434,
      "step": 3650
    },
    {
      "epoch": 16.022545454545455,
      "grad_norm": 21.447010040283203,
      "learning_rate": 1.8585858585858588e-05,
      "loss": 0.6861,
      "step": 3660
    },
    {
      "epoch": 16.024363636363635,
      "grad_norm": 18.33003807067871,
      "learning_rate": 1.8484848484848487e-05,
      "loss": 0.8178,
      "step": 3670
    },
    {
      "epoch": 16.02618181818182,
      "grad_norm": 11.963517189025879,
      "learning_rate": 1.8383838383838383e-05,
      "loss": 0.7212,
      "step": 3680
    },
    {
      "epoch": 16.028,
      "grad_norm": 48.92180633544922,
      "learning_rate": 1.8282828282828286e-05,
      "loss": 0.8028,
      "step": 3690
    },
    {
      "epoch": 16.029818181818182,
      "grad_norm": 82.1492919921875,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.827,
      "step": 3700
    },
    {
      "epoch": 16.031636363636363,
      "grad_norm": 22.397035598754883,
      "learning_rate": 1.808080808080808e-05,
      "loss": 0.9913,
      "step": 3710
    },
    {
      "epoch": 16.033454545454546,
      "grad_norm": 92.46865844726562,
      "learning_rate": 1.797979797979798e-05,
      "loss": 0.7714,
      "step": 3720
    },
    {
      "epoch": 16.035272727272726,
      "grad_norm": 44.989501953125,
      "learning_rate": 1.787878787878788e-05,
      "loss": 0.7027,
      "step": 3730
    },
    {
      "epoch": 16.03709090909091,
      "grad_norm": 61.908939361572266,
      "learning_rate": 1.777777777777778e-05,
      "loss": 0.7927,
      "step": 3740
    },
    {
      "epoch": 16.03890909090909,
      "grad_norm": 124.97834014892578,
      "learning_rate": 1.7676767676767676e-05,
      "loss": 0.7092,
      "step": 3750
    },
    {
      "epoch": 16.040181818181818,
      "eval_accuracy": 0.671957671957672,
      "eval_loss": 0.7971976399421692,
      "eval_runtime": 193.9592,
      "eval_samples_per_second": 0.974,
      "eval_steps_per_second": 0.247,
      "step": 3757
    },
    {
      "epoch": 17.000545454545456,
      "grad_norm": 40.8574104309082,
      "learning_rate": 1.7575757575757576e-05,
      "loss": 0.7822,
      "step": 3760
    },
    {
      "epoch": 17.002363636363636,
      "grad_norm": 41.2142448425293,
      "learning_rate": 1.7474747474747475e-05,
      "loss": 0.7753,
      "step": 3770
    },
    {
      "epoch": 17.004181818181817,
      "grad_norm": 24.647171020507812,
      "learning_rate": 1.7373737373737375e-05,
      "loss": 0.5532,
      "step": 3780
    },
    {
      "epoch": 17.006,
      "grad_norm": 95.03191375732422,
      "learning_rate": 1.7272727272727274e-05,
      "loss": 0.9703,
      "step": 3790
    },
    {
      "epoch": 17.00781818181818,
      "grad_norm": 23.023130416870117,
      "learning_rate": 1.7171717171717173e-05,
      "loss": 0.959,
      "step": 3800
    },
    {
      "epoch": 17.009636363636364,
      "grad_norm": 11.591606140136719,
      "learning_rate": 1.707070707070707e-05,
      "loss": 0.7913,
      "step": 3810
    },
    {
      "epoch": 17.011454545454544,
      "grad_norm": 9.28178882598877,
      "learning_rate": 1.6969696969696972e-05,
      "loss": 0.8629,
      "step": 3820
    },
    {
      "epoch": 17.013272727272728,
      "grad_norm": 8.856142044067383,
      "learning_rate": 1.686868686868687e-05,
      "loss": 0.7767,
      "step": 3830
    },
    {
      "epoch": 17.015090909090908,
      "grad_norm": 32.30220413208008,
      "learning_rate": 1.6767676767676768e-05,
      "loss": 0.7585,
      "step": 3840
    },
    {
      "epoch": 17.016909090909092,
      "grad_norm": 30.16693878173828,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.7905,
      "step": 3850
    },
    {
      "epoch": 17.018727272727272,
      "grad_norm": 10.049819946289062,
      "learning_rate": 1.6565656565656567e-05,
      "loss": 0.8439,
      "step": 3860
    },
    {
      "epoch": 17.020545454545456,
      "grad_norm": 27.414104461669922,
      "learning_rate": 1.6464646464646466e-05,
      "loss": 0.8226,
      "step": 3870
    },
    {
      "epoch": 17.022363636363636,
      "grad_norm": 37.73158645629883,
      "learning_rate": 1.6363636363636366e-05,
      "loss": 0.8542,
      "step": 3880
    },
    {
      "epoch": 17.02418181818182,
      "grad_norm": 24.757152557373047,
      "learning_rate": 1.6262626262626262e-05,
      "loss": 0.7458,
      "step": 3890
    },
    {
      "epoch": 17.026,
      "grad_norm": 33.96831512451172,
      "learning_rate": 1.6161616161616165e-05,
      "loss": 0.751,
      "step": 3900
    },
    {
      "epoch": 17.027818181818184,
      "grad_norm": 39.2791633605957,
      "learning_rate": 1.606060606060606e-05,
      "loss": 0.7804,
      "step": 3910
    },
    {
      "epoch": 17.029636363636364,
      "grad_norm": 13.831995010375977,
      "learning_rate": 1.595959595959596e-05,
      "loss": 0.7447,
      "step": 3920
    },
    {
      "epoch": 17.031454545454544,
      "grad_norm": 29.68831443786621,
      "learning_rate": 1.585858585858586e-05,
      "loss": 0.8449,
      "step": 3930
    },
    {
      "epoch": 17.033272727272728,
      "grad_norm": 89.53331756591797,
      "learning_rate": 1.5757575757575756e-05,
      "loss": 0.6897,
      "step": 3940
    },
    {
      "epoch": 17.035090909090908,
      "grad_norm": 31.637697219848633,
      "learning_rate": 1.565656565656566e-05,
      "loss": 0.8963,
      "step": 3950
    },
    {
      "epoch": 17.03690909090909,
      "grad_norm": 8.320121765136719,
      "learning_rate": 1.5555555555555555e-05,
      "loss": 0.4999,
      "step": 3960
    },
    {
      "epoch": 17.03872727272727,
      "grad_norm": 20.025047302246094,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 0.6816,
      "step": 3970
    },
    {
      "epoch": 17.040181818181818,
      "eval_accuracy": 0.708994708994709,
      "eval_loss": 0.712218165397644,
      "eval_runtime": 190.8911,
      "eval_samples_per_second": 0.99,
      "eval_steps_per_second": 0.251,
      "step": 3978
    },
    {
      "epoch": 18.000363636363637,
      "grad_norm": 11.93933391571045,
      "learning_rate": 1.5353535353535354e-05,
      "loss": 0.788,
      "step": 3980
    },
    {
      "epoch": 18.002181818181818,
      "grad_norm": 57.62703323364258,
      "learning_rate": 1.5252525252525255e-05,
      "loss": 0.9024,
      "step": 3990
    },
    {
      "epoch": 18.004,
      "grad_norm": 16.53204917907715,
      "learning_rate": 1.5151515151515153e-05,
      "loss": 0.7909,
      "step": 4000
    },
    {
      "epoch": 18.00581818181818,
      "grad_norm": 4.855262279510498,
      "learning_rate": 1.505050505050505e-05,
      "loss": 0.7213,
      "step": 4010
    },
    {
      "epoch": 18.007636363636365,
      "grad_norm": 7.463796138763428,
      "learning_rate": 1.494949494949495e-05,
      "loss": 0.9467,
      "step": 4020
    },
    {
      "epoch": 18.009454545454545,
      "grad_norm": 13.145658493041992,
      "learning_rate": 1.484848484848485e-05,
      "loss": 0.7383,
      "step": 4030
    },
    {
      "epoch": 18.011272727272726,
      "grad_norm": 29.503515243530273,
      "learning_rate": 1.4747474747474749e-05,
      "loss": 0.9726,
      "step": 4040
    },
    {
      "epoch": 18.01309090909091,
      "grad_norm": 52.793174743652344,
      "learning_rate": 1.4646464646464647e-05,
      "loss": 0.7689,
      "step": 4050
    },
    {
      "epoch": 18.01490909090909,
      "grad_norm": 43.86151123046875,
      "learning_rate": 1.4545454545454545e-05,
      "loss": 0.8847,
      "step": 4060
    },
    {
      "epoch": 18.016727272727273,
      "grad_norm": 5.855788707733154,
      "learning_rate": 1.4444444444444444e-05,
      "loss": 0.653,
      "step": 4070
    },
    {
      "epoch": 18.018545454545453,
      "grad_norm": 50.7757453918457,
      "learning_rate": 1.4343434343434345e-05,
      "loss": 0.74,
      "step": 4080
    },
    {
      "epoch": 18.020363636363637,
      "grad_norm": 119.08275604248047,
      "learning_rate": 1.4242424242424243e-05,
      "loss": 0.9263,
      "step": 4090
    },
    {
      "epoch": 18.022181818181817,
      "grad_norm": 18.412038803100586,
      "learning_rate": 1.4141414141414141e-05,
      "loss": 0.638,
      "step": 4100
    },
    {
      "epoch": 18.024,
      "grad_norm": 5.919767379760742,
      "learning_rate": 1.404040404040404e-05,
      "loss": 0.7176,
      "step": 4110
    },
    {
      "epoch": 18.02581818181818,
      "grad_norm": 22.25209617614746,
      "learning_rate": 1.3939393939393942e-05,
      "loss": 0.7239,
      "step": 4120
    },
    {
      "epoch": 18.027636363636365,
      "grad_norm": 18.18853187561035,
      "learning_rate": 1.383838383838384e-05,
      "loss": 0.7939,
      "step": 4130
    },
    {
      "epoch": 18.029454545454545,
      "grad_norm": 25.958633422851562,
      "learning_rate": 1.3737373737373737e-05,
      "loss": 0.7458,
      "step": 4140
    },
    {
      "epoch": 18.03127272727273,
      "grad_norm": 49.500728607177734,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.8482,
      "step": 4150
    },
    {
      "epoch": 18.03309090909091,
      "grad_norm": 50.96277618408203,
      "learning_rate": 1.3535353535353538e-05,
      "loss": 0.7981,
      "step": 4160
    },
    {
      "epoch": 18.034909090909093,
      "grad_norm": 48.261512756347656,
      "learning_rate": 1.3434343434343436e-05,
      "loss": 0.9629,
      "step": 4170
    },
    {
      "epoch": 18.036727272727273,
      "grad_norm": 70.20667266845703,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.9913,
      "step": 4180
    },
    {
      "epoch": 18.038545454545453,
      "grad_norm": 18.857158660888672,
      "learning_rate": 1.3232323232323233e-05,
      "loss": 0.839,
      "step": 4190
    },
    {
      "epoch": 18.040181818181818,
      "eval_accuracy": 0.7142857142857143,
      "eval_loss": 0.7404036521911621,
      "eval_runtime": 194.8175,
      "eval_samples_per_second": 0.97,
      "eval_steps_per_second": 0.246,
      "step": 4199
    },
    {
      "epoch": 19.00018181818182,
      "grad_norm": 13.861733436584473,
      "learning_rate": 1.3131313131313134e-05,
      "loss": 0.9389,
      "step": 4200
    },
    {
      "epoch": 19.002,
      "grad_norm": 85.8123550415039,
      "learning_rate": 1.3030303030303032e-05,
      "loss": 0.7941,
      "step": 4210
    },
    {
      "epoch": 19.003818181818183,
      "grad_norm": 28.504823684692383,
      "learning_rate": 1.292929292929293e-05,
      "loss": 0.7037,
      "step": 4220
    },
    {
      "epoch": 19.005636363636363,
      "grad_norm": 45.82936096191406,
      "learning_rate": 1.2828282828282829e-05,
      "loss": 0.7746,
      "step": 4230
    },
    {
      "epoch": 19.007454545454546,
      "grad_norm": 31.655263900756836,
      "learning_rate": 1.2727272727272727e-05,
      "loss": 0.7592,
      "step": 4240
    },
    {
      "epoch": 19.009272727272727,
      "grad_norm": 15.960184097290039,
      "learning_rate": 1.2626262626262628e-05,
      "loss": 0.6311,
      "step": 4250
    },
    {
      "epoch": 19.01109090909091,
      "grad_norm": 6.695078372955322,
      "learning_rate": 1.2525252525252526e-05,
      "loss": 0.7749,
      "step": 4260
    },
    {
      "epoch": 19.01290909090909,
      "grad_norm": 24.639005661010742,
      "learning_rate": 1.2424242424242424e-05,
      "loss": 0.827,
      "step": 4270
    },
    {
      "epoch": 19.014727272727274,
      "grad_norm": 16.702285766601562,
      "learning_rate": 1.2323232323232325e-05,
      "loss": 0.9035,
      "step": 4280
    },
    {
      "epoch": 19.016545454545454,
      "grad_norm": 27.27808952331543,
      "learning_rate": 1.2222222222222222e-05,
      "loss": 0.9385,
      "step": 4290
    },
    {
      "epoch": 19.018363636363638,
      "grad_norm": 13.049288749694824,
      "learning_rate": 1.2121212121212122e-05,
      "loss": 0.9155,
      "step": 4300
    },
    {
      "epoch": 19.02018181818182,
      "grad_norm": 12.153098106384277,
      "learning_rate": 1.202020202020202e-05,
      "loss": 0.7098,
      "step": 4310
    },
    {
      "epoch": 19.022,
      "grad_norm": 56.02737808227539,
      "learning_rate": 1.1919191919191921e-05,
      "loss": 0.7314,
      "step": 4320
    },
    {
      "epoch": 19.023818181818182,
      "grad_norm": 25.432979583740234,
      "learning_rate": 1.1818181818181819e-05,
      "loss": 0.8002,
      "step": 4330
    },
    {
      "epoch": 19.025636363636362,
      "grad_norm": 10.923587799072266,
      "learning_rate": 1.1717171717171718e-05,
      "loss": 0.889,
      "step": 4340
    },
    {
      "epoch": 19.027454545454546,
      "grad_norm": 30.052419662475586,
      "learning_rate": 1.1616161616161616e-05,
      "loss": 0.7825,
      "step": 4350
    },
    {
      "epoch": 19.029272727272726,
      "grad_norm": 13.3062744140625,
      "learning_rate": 1.1515151515151517e-05,
      "loss": 0.7734,
      "step": 4360
    },
    {
      "epoch": 19.03109090909091,
      "grad_norm": 81.07504272460938,
      "learning_rate": 1.1414141414141415e-05,
      "loss": 0.7578,
      "step": 4370
    },
    {
      "epoch": 19.03290909090909,
      "grad_norm": 97.549072265625,
      "learning_rate": 1.1313131313131314e-05,
      "loss": 0.9045,
      "step": 4380
    },
    {
      "epoch": 19.034727272727274,
      "grad_norm": 57.642093658447266,
      "learning_rate": 1.1212121212121212e-05,
      "loss": 0.5685,
      "step": 4390
    },
    {
      "epoch": 19.036545454545454,
      "grad_norm": 66.42289733886719,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 0.5724,
      "step": 4400
    },
    {
      "epoch": 19.038363636363638,
      "grad_norm": 10.334883689880371,
      "learning_rate": 1.1010101010101011e-05,
      "loss": 0.6384,
      "step": 4410
    },
    {
      "epoch": 19.040181818181818,
      "grad_norm": 4.034684181213379,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 0.4672,
      "step": 4420
    },
    {
      "epoch": 19.040181818181818,
      "eval_accuracy": 0.7195767195767195,
      "eval_loss": 0.7093486189842224,
      "eval_runtime": 195.9324,
      "eval_samples_per_second": 0.965,
      "eval_steps_per_second": 0.245,
      "step": 4420
    },
    {
      "epoch": 20.00181818181818,
      "grad_norm": 38.80495834350586,
      "learning_rate": 1.0808080808080808e-05,
      "loss": 0.7484,
      "step": 4430
    },
    {
      "epoch": 20.003636363636364,
      "grad_norm": 96.9532241821289,
      "learning_rate": 1.0707070707070708e-05,
      "loss": 0.8687,
      "step": 4440
    },
    {
      "epoch": 20.005454545454544,
      "grad_norm": 31.927587509155273,
      "learning_rate": 1.0606060606060607e-05,
      "loss": 0.6293,
      "step": 4450
    },
    {
      "epoch": 20.007272727272728,
      "grad_norm": 31.61720848083496,
      "learning_rate": 1.0505050505050505e-05,
      "loss": 0.8344,
      "step": 4460
    },
    {
      "epoch": 20.009090909090908,
      "grad_norm": 17.091989517211914,
      "learning_rate": 1.0404040404040405e-05,
      "loss": 0.8526,
      "step": 4470
    },
    {
      "epoch": 20.01090909090909,
      "grad_norm": 11.5504150390625,
      "learning_rate": 1.0303030303030304e-05,
      "loss": 1.0841,
      "step": 4480
    },
    {
      "epoch": 20.012727272727272,
      "grad_norm": 17.285606384277344,
      "learning_rate": 1.0202020202020204e-05,
      "loss": 0.6781,
      "step": 4490
    },
    {
      "epoch": 20.014545454545456,
      "grad_norm": 7.431727409362793,
      "learning_rate": 1.0101010101010101e-05,
      "loss": 0.6636,
      "step": 4500
    },
    {
      "epoch": 20.016363636363636,
      "grad_norm": 64.10966491699219,
      "learning_rate": 1e-05,
      "loss": 0.7486,
      "step": 4510
    },
    {
      "epoch": 20.01818181818182,
      "grad_norm": 37.64710235595703,
      "learning_rate": 9.898989898989899e-06,
      "loss": 0.7398,
      "step": 4520
    },
    {
      "epoch": 20.02,
      "grad_norm": 69.10208892822266,
      "learning_rate": 9.7979797979798e-06,
      "loss": 1.0162,
      "step": 4530
    },
    {
      "epoch": 20.021818181818183,
      "grad_norm": 27.21761703491211,
      "learning_rate": 9.696969696969698e-06,
      "loss": 0.7102,
      "step": 4540
    },
    {
      "epoch": 20.023636363636363,
      "grad_norm": 28.60508155822754,
      "learning_rate": 9.595959595959595e-06,
      "loss": 0.5108,
      "step": 4550
    },
    {
      "epoch": 20.025454545454547,
      "grad_norm": 124.87145233154297,
      "learning_rate": 9.494949494949495e-06,
      "loss": 0.7203,
      "step": 4560
    },
    {
      "epoch": 20.027272727272727,
      "grad_norm": 40.59431838989258,
      "learning_rate": 9.393939393939394e-06,
      "loss": 0.9909,
      "step": 4570
    },
    {
      "epoch": 20.029090909090908,
      "grad_norm": 18.73350715637207,
      "learning_rate": 9.292929292929294e-06,
      "loss": 0.6461,
      "step": 4580
    },
    {
      "epoch": 20.03090909090909,
      "grad_norm": 20.62631607055664,
      "learning_rate": 9.191919191919192e-06,
      "loss": 0.774,
      "step": 4590
    },
    {
      "epoch": 20.03272727272727,
      "grad_norm": 86.15074157714844,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.8109,
      "step": 4600
    },
    {
      "epoch": 20.034545454545455,
      "grad_norm": 43.705196380615234,
      "learning_rate": 8.98989898989899e-06,
      "loss": 0.6988,
      "step": 4610
    },
    {
      "epoch": 20.036363636363635,
      "grad_norm": 9.670613288879395,
      "learning_rate": 8.88888888888889e-06,
      "loss": 0.8493,
      "step": 4620
    },
    {
      "epoch": 20.03818181818182,
      "grad_norm": 36.182037353515625,
      "learning_rate": 8.787878787878788e-06,
      "loss": 0.6715,
      "step": 4630
    },
    {
      "epoch": 20.04,
      "grad_norm": 22.325069427490234,
      "learning_rate": 8.686868686868687e-06,
      "loss": 0.5119,
      "step": 4640
    },
    {
      "epoch": 20.040181818181818,
      "eval_accuracy": 0.6825396825396826,
      "eval_loss": 0.7170515656471252,
      "eval_runtime": 194.357,
      "eval_samples_per_second": 0.972,
      "eval_steps_per_second": 0.247,
      "step": 4641
    },
    {
      "epoch": 21.001636363636365,
      "grad_norm": 30.531394958496094,
      "learning_rate": 8.585858585858587e-06,
      "loss": 0.5961,
      "step": 4650
    },
    {
      "epoch": 21.003454545454545,
      "grad_norm": 24.909801483154297,
      "learning_rate": 8.484848484848486e-06,
      "loss": 0.8621,
      "step": 4660
    },
    {
      "epoch": 21.00527272727273,
      "grad_norm": 84.33313751220703,
      "learning_rate": 8.383838383838384e-06,
      "loss": 0.7983,
      "step": 4670
    },
    {
      "epoch": 21.00709090909091,
      "grad_norm": 33.05415344238281,
      "learning_rate": 8.282828282828283e-06,
      "loss": 0.8508,
      "step": 4680
    },
    {
      "epoch": 21.00890909090909,
      "grad_norm": 55.02461242675781,
      "learning_rate": 8.181818181818183e-06,
      "loss": 0.7546,
      "step": 4690
    },
    {
      "epoch": 21.010727272727273,
      "grad_norm": 19.233810424804688,
      "learning_rate": 8.080808080808082e-06,
      "loss": 0.58,
      "step": 4700
    },
    {
      "epoch": 21.012545454545453,
      "grad_norm": 8.127425193786621,
      "learning_rate": 7.97979797979798e-06,
      "loss": 0.7661,
      "step": 4710
    },
    {
      "epoch": 21.014363636363637,
      "grad_norm": 28.833446502685547,
      "learning_rate": 7.878787878787878e-06,
      "loss": 0.653,
      "step": 4720
    },
    {
      "epoch": 21.016181818181817,
      "grad_norm": 24.907699584960938,
      "learning_rate": 7.777777777777777e-06,
      "loss": 0.6454,
      "step": 4730
    },
    {
      "epoch": 21.018,
      "grad_norm": 20.222206115722656,
      "learning_rate": 7.676767676767677e-06,
      "loss": 0.847,
      "step": 4740
    },
    {
      "epoch": 21.01981818181818,
      "grad_norm": 7.958954811096191,
      "learning_rate": 7.5757575757575764e-06,
      "loss": 0.8589,
      "step": 4750
    },
    {
      "epoch": 21.021636363636365,
      "grad_norm": 75.67617797851562,
      "learning_rate": 7.474747474747475e-06,
      "loss": 0.6666,
      "step": 4760
    },
    {
      "epoch": 21.023454545454545,
      "grad_norm": 2.3386125564575195,
      "learning_rate": 7.3737373737373745e-06,
      "loss": 0.8145,
      "step": 4770
    },
    {
      "epoch": 21.02527272727273,
      "grad_norm": 69.61566162109375,
      "learning_rate": 7.272727272727272e-06,
      "loss": 0.4374,
      "step": 4780
    },
    {
      "epoch": 21.02709090909091,
      "grad_norm": 26.695165634155273,
      "learning_rate": 7.171717171717173e-06,
      "loss": 0.6145,
      "step": 4790
    },
    {
      "epoch": 21.028909090909092,
      "grad_norm": 19.140888214111328,
      "learning_rate": 7.0707070707070704e-06,
      "loss": 0.6424,
      "step": 4800
    },
    {
      "epoch": 21.030727272727272,
      "grad_norm": 8.267212867736816,
      "learning_rate": 6.969696969696971e-06,
      "loss": 0.7443,
      "step": 4810
    },
    {
      "epoch": 21.032545454545456,
      "grad_norm": 23.349206924438477,
      "learning_rate": 6.8686868686868685e-06,
      "loss": 0.5858,
      "step": 4820
    },
    {
      "epoch": 21.034363636363636,
      "grad_norm": 24.537132263183594,
      "learning_rate": 6.767676767676769e-06,
      "loss": 0.7272,
      "step": 4830
    },
    {
      "epoch": 21.036181818181817,
      "grad_norm": 27.91595458984375,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.9845,
      "step": 4840
    },
    {
      "epoch": 21.038,
      "grad_norm": 20.382572174072266,
      "learning_rate": 6.565656565656567e-06,
      "loss": 0.629,
      "step": 4850
    },
    {
      "epoch": 21.03981818181818,
      "grad_norm": 49.914493560791016,
      "learning_rate": 6.464646464646465e-06,
      "loss": 0.7769,
      "step": 4860
    },
    {
      "epoch": 21.040181818181818,
      "eval_accuracy": 0.6931216931216931,
      "eval_loss": 0.7071664929389954,
      "eval_runtime": 196.3708,
      "eval_samples_per_second": 0.962,
      "eval_steps_per_second": 0.244,
      "step": 4862
    },
    {
      "epoch": 22.001454545454546,
      "grad_norm": 47.80362319946289,
      "learning_rate": 6.363636363636363e-06,
      "loss": 0.9012,
      "step": 4870
    },
    {
      "epoch": 22.003272727272726,
      "grad_norm": 41.31081771850586,
      "learning_rate": 6.262626262626263e-06,
      "loss": 0.7611,
      "step": 4880
    },
    {
      "epoch": 22.00509090909091,
      "grad_norm": 24.996232986450195,
      "learning_rate": 6.161616161616162e-06,
      "loss": 0.479,
      "step": 4890
    },
    {
      "epoch": 22.00690909090909,
      "grad_norm": 18.948902130126953,
      "learning_rate": 6.060606060606061e-06,
      "loss": 0.6003,
      "step": 4900
    },
    {
      "epoch": 22.008727272727274,
      "grad_norm": 59.650230407714844,
      "learning_rate": 5.9595959595959605e-06,
      "loss": 0.5891,
      "step": 4910
    },
    {
      "epoch": 22.010545454545454,
      "grad_norm": 96.68114471435547,
      "learning_rate": 5.858585858585859e-06,
      "loss": 0.8473,
      "step": 4920
    },
    {
      "epoch": 22.012363636363638,
      "grad_norm": 29.867095947265625,
      "learning_rate": 5.7575757575757586e-06,
      "loss": 0.6998,
      "step": 4930
    },
    {
      "epoch": 22.014181818181818,
      "grad_norm": 25.995758056640625,
      "learning_rate": 5.656565656565657e-06,
      "loss": 0.6509,
      "step": 4940
    },
    {
      "epoch": 22.016,
      "grad_norm": 54.34788131713867,
      "learning_rate": 5.555555555555556e-06,
      "loss": 0.5927,
      "step": 4950
    },
    {
      "epoch": 22.017818181818182,
      "grad_norm": 42.82137680053711,
      "learning_rate": 5.4545454545454545e-06,
      "loss": 0.7703,
      "step": 4960
    },
    {
      "epoch": 22.019636363636362,
      "grad_norm": 34.68387222290039,
      "learning_rate": 5.353535353535354e-06,
      "loss": 0.7322,
      "step": 4970
    },
    {
      "epoch": 22.021454545454546,
      "grad_norm": 27.881155014038086,
      "learning_rate": 5.2525252525252526e-06,
      "loss": 0.5349,
      "step": 4980
    },
    {
      "epoch": 22.023272727272726,
      "grad_norm": 22.914897918701172,
      "learning_rate": 5.151515151515152e-06,
      "loss": 0.8067,
      "step": 4990
    },
    {
      "epoch": 22.02509090909091,
      "grad_norm": 18.1747989654541,
      "learning_rate": 5.050505050505051e-06,
      "loss": 0.7394,
      "step": 5000
    },
    {
      "epoch": 22.02690909090909,
      "grad_norm": 18.75357437133789,
      "learning_rate": 4.949494949494949e-06,
      "loss": 0.7725,
      "step": 5010
    },
    {
      "epoch": 22.028727272727274,
      "grad_norm": 88.3263931274414,
      "learning_rate": 4.848484848484849e-06,
      "loss": 0.6074,
      "step": 5020
    },
    {
      "epoch": 22.030545454545454,
      "grad_norm": 34.32087326049805,
      "learning_rate": 4.747474747474747e-06,
      "loss": 0.7811,
      "step": 5030
    },
    {
      "epoch": 22.032363636363637,
      "grad_norm": 32.58205032348633,
      "learning_rate": 4.646464646464647e-06,
      "loss": 0.8519,
      "step": 5040
    },
    {
      "epoch": 22.034181818181818,
      "grad_norm": 28.88677215576172,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.6139,
      "step": 5050
    },
    {
      "epoch": 22.036,
      "grad_norm": 14.495834350585938,
      "learning_rate": 4.444444444444445e-06,
      "loss": 0.5717,
      "step": 5060
    },
    {
      "epoch": 22.03781818181818,
      "grad_norm": 42.042579650878906,
      "learning_rate": 4.343434343434344e-06,
      "loss": 0.6286,
      "step": 5070
    },
    {
      "epoch": 22.039636363636365,
      "grad_norm": 102.42108917236328,
      "learning_rate": 4.242424242424243e-06,
      "loss": 0.6635,
      "step": 5080
    },
    {
      "epoch": 22.040181818181818,
      "eval_accuracy": 0.7195767195767195,
      "eval_loss": 0.7095865607261658,
      "eval_runtime": 197.7202,
      "eval_samples_per_second": 0.956,
      "eval_steps_per_second": 0.243,
      "step": 5083
    },
    {
      "epoch": 23.001272727272728,
      "grad_norm": 71.68008422851562,
      "learning_rate": 4.141414141414142e-06,
      "loss": 0.6267,
      "step": 5090
    },
    {
      "epoch": 23.003090909090908,
      "grad_norm": 50.31315612792969,
      "learning_rate": 4.040404040404041e-06,
      "loss": 0.7405,
      "step": 5100
    },
    {
      "epoch": 23.00490909090909,
      "grad_norm": 40.68128204345703,
      "learning_rate": 3.939393939393939e-06,
      "loss": 0.7911,
      "step": 5110
    },
    {
      "epoch": 23.00672727272727,
      "grad_norm": 35.65066146850586,
      "learning_rate": 3.8383838383838385e-06,
      "loss": 0.8023,
      "step": 5120
    },
    {
      "epoch": 23.008545454545455,
      "grad_norm": 24.422685623168945,
      "learning_rate": 3.7373737373737375e-06,
      "loss": 0.5818,
      "step": 5130
    },
    {
      "epoch": 23.010363636363635,
      "grad_norm": 19.126468658447266,
      "learning_rate": 3.636363636363636e-06,
      "loss": 0.6642,
      "step": 5140
    },
    {
      "epoch": 23.01218181818182,
      "grad_norm": 19.357858657836914,
      "learning_rate": 3.5353535353535352e-06,
      "loss": 0.6839,
      "step": 5150
    },
    {
      "epoch": 23.014,
      "grad_norm": 38.63173294067383,
      "learning_rate": 3.4343434343434343e-06,
      "loss": 0.5163,
      "step": 5160
    },
    {
      "epoch": 23.015818181818183,
      "grad_norm": 17.257484436035156,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.7778,
      "step": 5170
    },
    {
      "epoch": 23.017636363636363,
      "grad_norm": 38.906246185302734,
      "learning_rate": 3.2323232323232324e-06,
      "loss": 0.6887,
      "step": 5180
    },
    {
      "epoch": 23.019454545454547,
      "grad_norm": 103.85768127441406,
      "learning_rate": 3.1313131313131314e-06,
      "loss": 0.7222,
      "step": 5190
    },
    {
      "epoch": 23.021272727272727,
      "grad_norm": 28.443326950073242,
      "learning_rate": 3.0303030303030305e-06,
      "loss": 0.7608,
      "step": 5200
    },
    {
      "epoch": 23.02309090909091,
      "grad_norm": 38.45210647583008,
      "learning_rate": 2.9292929292929295e-06,
      "loss": 0.782,
      "step": 5210
    },
    {
      "epoch": 23.02490909090909,
      "grad_norm": 28.65218734741211,
      "learning_rate": 2.8282828282828286e-06,
      "loss": 0.4031,
      "step": 5220
    },
    {
      "epoch": 23.02672727272727,
      "grad_norm": 59.21070098876953,
      "learning_rate": 2.7272727272727272e-06,
      "loss": 0.6224,
      "step": 5230
    },
    {
      "epoch": 23.028545454545455,
      "grad_norm": 136.05392456054688,
      "learning_rate": 2.6262626262626263e-06,
      "loss": 0.7196,
      "step": 5240
    },
    {
      "epoch": 23.030363636363635,
      "grad_norm": 183.45257568359375,
      "learning_rate": 2.5252525252525253e-06,
      "loss": 1.0722,
      "step": 5250
    },
    {
      "epoch": 23.03218181818182,
      "grad_norm": 17.261632919311523,
      "learning_rate": 2.4242424242424244e-06,
      "loss": 0.5863,
      "step": 5260
    },
    {
      "epoch": 23.034,
      "grad_norm": 100.2103500366211,
      "learning_rate": 2.3232323232323234e-06,
      "loss": 0.4809,
      "step": 5270
    },
    {
      "epoch": 23.035818181818183,
      "grad_norm": 32.065818786621094,
      "learning_rate": 2.2222222222222225e-06,
      "loss": 0.5746,
      "step": 5280
    },
    {
      "epoch": 23.037636363636363,
      "grad_norm": 9.926220893859863,
      "learning_rate": 2.1212121212121216e-06,
      "loss": 0.451,
      "step": 5290
    },
    {
      "epoch": 23.039454545454547,
      "grad_norm": 27.379253387451172,
      "learning_rate": 2.0202020202020206e-06,
      "loss": 0.8027,
      "step": 5300
    },
    {
      "epoch": 23.040181818181818,
      "eval_accuracy": 0.7037037037037037,
      "eval_loss": 0.6887486577033997,
      "eval_runtime": 194.3489,
      "eval_samples_per_second": 0.972,
      "eval_steps_per_second": 0.247,
      "step": 5304
    },
    {
      "epoch": 24.00109090909091,
      "grad_norm": 71.3526840209961,
      "learning_rate": 1.9191919191919192e-06,
      "loss": 0.5815,
      "step": 5310
    },
    {
      "epoch": 24.002909090909093,
      "grad_norm": 4.591304779052734,
      "learning_rate": 1.818181818181818e-06,
      "loss": 0.8352,
      "step": 5320
    },
    {
      "epoch": 24.004727272727273,
      "grad_norm": 7.72051477432251,
      "learning_rate": 1.7171717171717171e-06,
      "loss": 0.8424,
      "step": 5330
    },
    {
      "epoch": 24.006545454545453,
      "grad_norm": 71.4774169921875,
      "learning_rate": 1.6161616161616162e-06,
      "loss": 0.9185,
      "step": 5340
    },
    {
      "epoch": 24.008363636363637,
      "grad_norm": 25.500089645385742,
      "learning_rate": 1.5151515151515152e-06,
      "loss": 0.5875,
      "step": 5350
    },
    {
      "epoch": 24.010181818181817,
      "grad_norm": 80.47840118408203,
      "learning_rate": 1.4141414141414143e-06,
      "loss": 0.9027,
      "step": 5360
    },
    {
      "epoch": 24.012,
      "grad_norm": 15.391636848449707,
      "learning_rate": 1.3131313131313131e-06,
      "loss": 0.4817,
      "step": 5370
    },
    {
      "epoch": 24.01381818181818,
      "grad_norm": 29.841554641723633,
      "learning_rate": 1.2121212121212122e-06,
      "loss": 0.7169,
      "step": 5380
    },
    {
      "epoch": 24.015636363636364,
      "grad_norm": 50.79997634887695,
      "learning_rate": 1.1111111111111112e-06,
      "loss": 0.7159,
      "step": 5390
    },
    {
      "epoch": 24.017454545454545,
      "grad_norm": 108.86124420166016,
      "learning_rate": 1.0101010101010103e-06,
      "loss": 1.0093,
      "step": 5400
    },
    {
      "epoch": 24.019272727272728,
      "grad_norm": 29.645301818847656,
      "learning_rate": 9.09090909090909e-07,
      "loss": 0.6033,
      "step": 5410
    },
    {
      "epoch": 24.02109090909091,
      "grad_norm": 28.632356643676758,
      "learning_rate": 8.080808080808081e-07,
      "loss": 0.5634,
      "step": 5420
    },
    {
      "epoch": 24.022909090909092,
      "grad_norm": 24.16314697265625,
      "learning_rate": 7.070707070707071e-07,
      "loss": 0.5789,
      "step": 5430
    },
    {
      "epoch": 24.024727272727272,
      "grad_norm": 69.10330200195312,
      "learning_rate": 6.060606060606061e-07,
      "loss": 0.523,
      "step": 5440
    },
    {
      "epoch": 24.026545454545456,
      "grad_norm": 153.1522216796875,
      "learning_rate": 5.050505050505052e-07,
      "loss": 0.7189,
      "step": 5450
    },
    {
      "epoch": 24.028363636363636,
      "grad_norm": 29.681365966796875,
      "learning_rate": 4.0404040404040405e-07,
      "loss": 0.4212,
      "step": 5460
    },
    {
      "epoch": 24.03018181818182,
      "grad_norm": 51.57491683959961,
      "learning_rate": 3.0303030303030305e-07,
      "loss": 0.724,
      "step": 5470
    },
    {
      "epoch": 24.032,
      "grad_norm": 19.3199520111084,
      "learning_rate": 2.0202020202020202e-07,
      "loss": 0.5203,
      "step": 5480
    },
    {
      "epoch": 24.03381818181818,
      "grad_norm": 138.87806701660156,
      "learning_rate": 1.0101010101010101e-07,
      "loss": 0.9596,
      "step": 5490
    },
    {
      "epoch": 24.035636363636364,
      "grad_norm": 38.64838790893555,
      "learning_rate": 0.0,
      "loss": 0.4765,
      "step": 5500
    },
    {
      "epoch": 24.035636363636364,
      "eval_accuracy": 0.7407407407407407,
      "eval_loss": 0.6834621429443359,
      "eval_runtime": 207.2262,
      "eval_samples_per_second": 0.912,
      "eval_steps_per_second": 0.232,
      "step": 5500
    },
    {
      "epoch": 1.0001818181818183,
      "eval_accuracy": 0.671957671957672,
      "eval_loss": 0.7627605199813843,
      "eval_runtime": 237.6048,
      "eval_samples_per_second": 0.795,
      "eval_steps_per_second": 0.202,
      "step": 5501
    }
  ],
  "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": 4.389253736670403e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}