{
  "best_metric": 0.5704381465911865,
  "best_model_checkpoint": "ckpt/llama2_13b_fuze30_no_sys/news_commentary_es/checkpoint-3000",
  "epoch": 5.0,
  "eval_steps": 200,
  "global_step": 3985,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 0.5029562711715698,
      "learning_rate": 5e-06,
      "loss": 1.2458,
      "step": 10
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.6982043981552124,
      "learning_rate": 1e-05,
      "loss": 1.266,
      "step": 20
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.7867264151573181,
      "learning_rate": 9.99984305369385e-06,
      "loss": 1.2072,
      "step": 30
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.7258906364440918,
      "learning_rate": 9.999372224628255e-06,
      "loss": 1.0755,
      "step": 40
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.6007847189903259,
      "learning_rate": 9.998587542361168e-06,
      "loss": 0.9002,
      "step": 50
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.3854537606239319,
      "learning_rate": 9.997489056153784e-06,
      "loss": 0.7568,
      "step": 60
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.3243318796157837,
      "learning_rate": 9.996076834967442e-06,
      "loss": 0.7157,
      "step": 70
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.24546250700950623,
      "learning_rate": 9.994350967459303e-06,
      "loss": 0.6984,
      "step": 80
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.34287428855895996,
      "learning_rate": 9.99231156197678e-06,
      "loss": 0.6996,
      "step": 90
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.3073829412460327,
      "learning_rate": 9.989958746550732e-06,
      "loss": 0.7129,
      "step": 100
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.2948177456855774,
      "learning_rate": 9.987292668887439e-06,
      "loss": 0.6977,
      "step": 110
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.3140040934085846,
      "learning_rate": 9.984313496359317e-06,
      "loss": 0.6272,
      "step": 120
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.2903355658054352,
      "learning_rate": 9.981021415994413e-06,
      "loss": 0.6285,
      "step": 130
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.3696918189525604,
      "learning_rate": 9.97741663446467e-06,
      "loss": 0.651,
      "step": 140
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.24330583214759827,
      "learning_rate": 9.973499378072947e-06,
      "loss": 0.6376,
      "step": 150
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.26600131392478943,
      "learning_rate": 9.96926989273881e-06,
      "loss": 0.6402,
      "step": 160
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.34049201011657715,
      "learning_rate": 9.9647284439831e-06,
      "loss": 0.6662,
      "step": 170
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.3909376859664917,
      "learning_rate": 9.959875316911261e-06,
      "loss": 0.6134,
      "step": 180
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.3953985869884491,
      "learning_rate": 9.95471081619544e-06,
      "loss": 0.6054,
      "step": 190
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.42383915185928345,
      "learning_rate": 9.949235266055356e-06,
      "loss": 0.6296,
      "step": 200
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.6343660950660706,
      "eval_runtime": 71.7569,
      "eval_samples_per_second": 62.712,
      "eval_steps_per_second": 1.965,
      "step": 200
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.37584221363067627,
      "learning_rate": 9.943449010237965e-06,
      "loss": 0.6724,
      "step": 210
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.368352472782135,
      "learning_rate": 9.937352411995851e-06,
      "loss": 0.6065,
      "step": 220
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.30686938762664795,
      "learning_rate": 9.930945854064447e-06,
      "loss": 0.5998,
      "step": 230
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.2913750410079956,
      "learning_rate": 9.924229738637992e-06,
      "loss": 0.6274,
      "step": 240
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.30744606256484985,
      "learning_rate": 9.91720448734429e-06,
      "loss": 0.6442,
      "step": 250
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.315236359834671,
      "learning_rate": 9.909870541218238e-06,
      "loss": 0.6026,
      "step": 260
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.28183722496032715,
      "learning_rate": 9.902228360674137e-06,
      "loss": 0.6511,
      "step": 270
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.33093082904815674,
      "learning_rate": 9.89427842547679e-06,
      "loss": 0.6575,
      "step": 280
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.3711361587047577,
      "learning_rate": 9.886021234711382e-06,
      "loss": 0.6209,
      "step": 290
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.35551759600639343,
      "learning_rate": 9.877457306752149e-06,
      "loss": 0.6265,
      "step": 300
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.3064817786216736,
      "learning_rate": 9.868587179229834e-06,
      "loss": 0.6214,
      "step": 310
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.4653981924057007,
      "learning_rate": 9.859411408997939e-06,
      "loss": 0.6234,
      "step": 320
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.3303908407688141,
      "learning_rate": 9.84993057209776e-06,
      "loss": 0.6502,
      "step": 330
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.3864089548587799,
      "learning_rate": 9.84014526372223e-06,
      "loss": 0.6275,
      "step": 340
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.3393557071685791,
      "learning_rate": 9.83005609817855e-06,
      "loss": 0.5913,
      "step": 350
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.3689006567001343,
      "learning_rate": 9.819663708849624e-06,
      "loss": 0.5962,
      "step": 360
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.32422763109207153,
      "learning_rate": 9.808968748154302e-06,
      "loss": 0.5896,
      "step": 370
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.3583005964756012,
      "learning_rate": 9.797971887506413e-06,
      "loss": 0.617,
      "step": 380
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.331403911113739,
      "learning_rate": 9.786673817272621e-06,
      "loss": 0.5994,
      "step": 390
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.3698234558105469,
      "learning_rate": 9.77507524672908e-06,
      "loss": 0.6272,
      "step": 400
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.612372875213623,
      "eval_runtime": 68.9986,
      "eval_samples_per_second": 65.219,
      "eval_steps_per_second": 2.044,
      "step": 400
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.3716101348400116,
      "learning_rate": 9.763176904016914e-06,
      "loss": 0.6246,
      "step": 410
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.4471350610256195,
      "learning_rate": 9.750979536096495e-06,
      "loss": 0.645,
      "step": 420
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.31366682052612305,
      "learning_rate": 9.738483908700563e-06,
      "loss": 0.6397,
      "step": 430
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.3895745575428009,
      "learning_rate": 9.72569080628614e-06,
      "loss": 0.6118,
      "step": 440
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.42779144644737244,
      "learning_rate": 9.71260103198529e-06,
      "loss": 0.6359,
      "step": 450
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.37752124667167664,
      "learning_rate": 9.699215407554712e-06,
      "loss": 0.5863,
      "step": 460
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.43578141927719116,
      "learning_rate": 9.685534773324124e-06,
      "loss": 0.6244,
      "step": 470
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.5181850790977478,
      "learning_rate": 9.671559988143529e-06,
      "loss": 0.5962,
      "step": 480
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.43225836753845215,
      "learning_rate": 9.657291929329293e-06,
      "loss": 0.631,
      "step": 490
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.36941251158714294,
      "learning_rate": 9.64273149260907e-06,
      "loss": 0.6398,
      "step": 500
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.3803509473800659,
      "learning_rate": 9.627879592065558e-06,
      "loss": 0.6393,
      "step": 510
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.3586438298225403,
      "learning_rate": 9.612737160079134e-06,
      "loss": 0.5944,
      "step": 520
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.41518551111221313,
      "learning_rate": 9.597305147269303e-06,
      "loss": 0.6224,
      "step": 530
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.3281481862068176,
      "learning_rate": 9.581584522435025e-06,
      "loss": 0.5853,
      "step": 540
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.3723869323730469,
      "learning_rate": 9.565576272493902e-06,
      "loss": 0.591,
      "step": 550
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.37543514370918274,
      "learning_rate": 9.549281402420212e-06,
      "loss": 0.5883,
      "step": 560
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.39954090118408203,
      "learning_rate": 9.532700935181822e-06,
      "loss": 0.5942,
      "step": 570
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.4060945212841034,
      "learning_rate": 9.515835911675968e-06,
      "loss": 0.6082,
      "step": 580
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.32310950756073,
      "learning_rate": 9.498687390663905e-06,
      "loss": 0.5836,
      "step": 590
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.37066149711608887,
      "learning_rate": 9.481256448704445e-06,
      "loss": 0.6167,
      "step": 600
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.6020291447639465,
      "eval_runtime": 71.3713,
      "eval_samples_per_second": 63.051,
      "eval_steps_per_second": 1.976,
      "step": 600
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.38338422775268555,
      "learning_rate": 9.463544180086368e-06,
      "loss": 0.6211,
      "step": 610
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.3963073790073395,
      "learning_rate": 9.445551696759732e-06,
      "loss": 0.5827,
      "step": 620
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.4199773967266083,
      "learning_rate": 9.427280128266049e-06,
      "loss": 0.5782,
      "step": 630
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.39564579725265503,
      "learning_rate": 9.4087306216674e-06,
      "loss": 0.6028,
      "step": 640
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.43293875455856323,
      "learning_rate": 9.389904341474394e-06,
      "loss": 0.5791,
      "step": 650
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.3330543041229248,
      "learning_rate": 9.370802469573088e-06,
      "loss": 0.5875,
      "step": 660
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.4487074613571167,
      "learning_rate": 9.351426205150778e-06,
      "loss": 0.6148,
      "step": 670
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.4278540015220642,
      "learning_rate": 9.331776764620711e-06,
      "loss": 0.6271,
      "step": 680
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.4620065987110138,
      "learning_rate": 9.311855381545735e-06,
      "loss": 0.6439,
      "step": 690
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.5026012063026428,
      "learning_rate": 9.291663306560843e-06,
      "loss": 0.614,
      "step": 700
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.3847663700580597,
      "learning_rate": 9.271201807294667e-06,
      "loss": 0.6058,
      "step": 710
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.38854289054870605,
      "learning_rate": 9.250472168289901e-06,
      "loss": 0.5859,
      "step": 720
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.4080798625946045,
      "learning_rate": 9.22947569092265e-06,
      "loss": 0.5874,
      "step": 730
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.587581992149353,
      "learning_rate": 9.20821369332074e-06,
      "loss": 0.6077,
      "step": 740
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.5180934071540833,
      "learning_rate": 9.186687510280968e-06,
      "loss": 0.6005,
      "step": 750
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.4063900411128998,
      "learning_rate": 9.164898493185295e-06,
      "loss": 0.6007,
      "step": 760
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.3806501626968384,
      "learning_rate": 9.142848009916025e-06,
      "loss": 0.6,
      "step": 770
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.4454844892024994,
      "learning_rate": 9.120537444769912e-06,
      "loss": 0.5704,
      "step": 780
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.4750034511089325,
      "learning_rate": 9.097968198371277e-06,
      "loss": 0.626,
      "step": 790
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.44586676359176636,
      "learning_rate": 9.075141687584056e-06,
      "loss": 0.5882,
      "step": 800
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.5939783453941345,
      "eval_runtime": 72.333,
      "eval_samples_per_second": 62.212,
      "eval_steps_per_second": 1.949,
      "step": 800
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.4351860284805298,
      "learning_rate": 9.052059345422874e-06,
      "loss": 0.5908,
      "step": 810
    },
    {
      "epoch": 1.03,
      "grad_norm": 0.5268151760101318,
      "learning_rate": 9.028722620963066e-06,
      "loss": 0.6155,
      "step": 820
    },
    {
      "epoch": 1.04,
      "grad_norm": 0.5248316526412964,
      "learning_rate": 9.00513297924971e-06,
      "loss": 0.5736,
      "step": 830
    },
    {
      "epoch": 1.05,
      "grad_norm": 0.43092867732048035,
      "learning_rate": 8.98129190120566e-06,
      "loss": 0.5602,
      "step": 840
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.4282706379890442,
      "learning_rate": 8.957200883538568e-06,
      "loss": 0.5737,
      "step": 850
    },
    {
      "epoch": 1.08,
      "grad_norm": 0.6257330775260925,
      "learning_rate": 8.932861438646929e-06,
      "loss": 0.5982,
      "step": 860
    },
    {
      "epoch": 1.09,
      "grad_norm": 0.44131189584732056,
      "learning_rate": 8.908275094525132e-06,
      "loss": 0.6144,
      "step": 870
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.49996450543403625,
      "learning_rate": 8.88344339466753e-06,
      "loss": 0.6138,
      "step": 880
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.41754284501075745,
      "learning_rate": 8.858367897971555e-06,
      "loss": 0.5641,
      "step": 890
    },
    {
      "epoch": 1.13,
      "grad_norm": 0.5091477036476135,
      "learning_rate": 8.833050178639835e-06,
      "loss": 0.5455,
      "step": 900
    },
    {
      "epoch": 1.14,
      "grad_norm": 0.5106736421585083,
      "learning_rate": 8.807491826081384e-06,
      "loss": 0.5913,
      "step": 910
    },
    {
      "epoch": 1.15,
      "grad_norm": 0.48033618927001953,
      "learning_rate": 8.781694444811813e-06,
      "loss": 0.5773,
      "step": 920
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.5638559460639954,
      "learning_rate": 8.755659654352599e-06,
      "loss": 0.5732,
      "step": 930
    },
    {
      "epoch": 1.18,
      "grad_norm": 0.45605936646461487,
      "learning_rate": 8.729389089129423e-06,
      "loss": 0.5911,
      "step": 940
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.4803718328475952,
      "learning_rate": 8.702884398369552e-06,
      "loss": 0.5909,
      "step": 950
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.5215190649032593,
      "learning_rate": 8.676147245998307e-06,
      "loss": 0.614,
      "step": 960
    },
    {
      "epoch": 1.22,
      "grad_norm": 0.490953654050827,
      "learning_rate": 8.649179310534612e-06,
      "loss": 0.5687,
      "step": 970
    },
    {
      "epoch": 1.23,
      "grad_norm": 0.5034527778625488,
      "learning_rate": 8.621982284985611e-06,
      "loss": 0.5948,
      "step": 980
    },
    {
      "epoch": 1.24,
      "grad_norm": 0.5091165900230408,
      "learning_rate": 8.59455787674038e-06,
      "loss": 0.5695,
      "step": 990
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.49061551690101624,
      "learning_rate": 8.56690780746275e-06,
      "loss": 0.5522,
      "step": 1000
    },
    {
      "epoch": 1.25,
      "eval_loss": 0.5888673067092896,
      "eval_runtime": 73.1303,
      "eval_samples_per_second": 61.534,
      "eval_steps_per_second": 1.928,
      "step": 1000
    },
    {
      "epoch": 1.27,
      "grad_norm": 0.4834423065185547,
      "learning_rate": 8.539033812983215e-06,
      "loss": 0.5413,
      "step": 1010
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.6021168828010559,
      "learning_rate": 8.510937643189962e-06,
      "loss": 0.5882,
      "step": 1020
    },
    {
      "epoch": 1.29,
      "grad_norm": 0.5302062630653381,
      "learning_rate": 8.482621061919021e-06,
      "loss": 0.6059,
      "step": 1030
    },
    {
      "epoch": 1.3,
      "grad_norm": 0.4995095729827881,
      "learning_rate": 8.454085846843523e-06,
      "loss": 0.5419,
      "step": 1040
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.5463216304779053,
      "learning_rate": 8.425333789362108e-06,
      "loss": 0.5941,
      "step": 1050
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.536431610584259,
      "learning_rate": 8.396366694486466e-06,
      "loss": 0.6284,
      "step": 1060
    },
    {
      "epoch": 1.34,
      "grad_norm": 0.5217477083206177,
      "learning_rate": 8.36718638072801e-06,
      "loss": 0.6178,
      "step": 1070
    },
    {
      "epoch": 1.36,
      "grad_norm": 0.4792711138725281,
      "learning_rate": 8.337794679983724e-06,
      "loss": 0.5858,
      "step": 1080
    },
    {
      "epoch": 1.37,
      "grad_norm": 0.5548479557037354,
      "learning_rate": 8.30819343742115e-06,
      "loss": 0.5861,
      "step": 1090
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.4717773199081421,
      "learning_rate": 8.278384511362562e-06,
      "loss": 0.6047,
      "step": 1100
    },
    {
      "epoch": 1.39,
      "grad_norm": 0.5312036275863647,
      "learning_rate": 8.248369773168297e-06,
      "loss": 0.5994,
      "step": 1110
    },
    {
      "epoch": 1.41,
      "grad_norm": 0.4545268714427948,
      "learning_rate": 8.218151107119266e-06,
      "loss": 0.5618,
      "step": 1120
    },
    {
      "epoch": 1.42,
      "grad_norm": 0.6215177774429321,
      "learning_rate": 8.187730410298675e-06,
      "loss": 0.6228,
      "step": 1130
    },
    {
      "epoch": 1.43,
      "grad_norm": 0.5464479923248291,
      "learning_rate": 8.157109592472924e-06,
      "loss": 0.5774,
      "step": 1140
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.5225650668144226,
      "learning_rate": 8.126290575971714e-06,
      "loss": 0.5804,
      "step": 1150
    },
    {
      "epoch": 1.46,
      "grad_norm": 0.6188873648643494,
      "learning_rate": 8.095275295567361e-06,
      "loss": 0.5498,
      "step": 1160
    },
    {
      "epoch": 1.47,
      "grad_norm": 0.5218258500099182,
      "learning_rate": 8.064065698353344e-06,
      "loss": 0.5829,
      "step": 1170
    },
    {
      "epoch": 1.48,
      "grad_norm": 0.5987175703048706,
      "learning_rate": 8.032663743622065e-06,
      "loss": 0.5914,
      "step": 1180
    },
    {
      "epoch": 1.49,
      "grad_norm": 0.527808666229248,
      "learning_rate": 8.001071402741843e-06,
      "loss": 0.5682,
      "step": 1190
    },
    {
      "epoch": 1.51,
      "grad_norm": 0.5273358821868896,
      "learning_rate": 7.969290659033156e-06,
      "loss": 0.5919,
      "step": 1200
    },
    {
      "epoch": 1.51,
      "eval_loss": 0.5836370587348938,
      "eval_runtime": 71.4972,
      "eval_samples_per_second": 62.939,
      "eval_steps_per_second": 1.972,
      "step": 1200
    },
    {
      "epoch": 1.52,
      "grad_norm": 0.6059889197349548,
      "learning_rate": 7.93732350764414e-06,
      "loss": 0.5763,
      "step": 1210
    },
    {
      "epoch": 1.53,
      "grad_norm": 0.5306861996650696,
      "learning_rate": 7.905171955425328e-06,
      "loss": 0.5567,
      "step": 1220
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.5967408418655396,
      "learning_rate": 7.872838020803662e-06,
      "loss": 0.5609,
      "step": 1230
    },
    {
      "epoch": 1.56,
      "grad_norm": 0.6726911067962646,
      "learning_rate": 7.84032373365578e-06,
      "loss": 0.6471,
      "step": 1240
    },
    {
      "epoch": 1.57,
      "grad_norm": 0.5502099990844727,
      "learning_rate": 7.807631135180591e-06,
      "loss": 0.5886,
      "step": 1250
    },
    {
      "epoch": 1.58,
      "grad_norm": 0.615433394908905,
      "learning_rate": 7.774762277771123e-06,
      "loss": 0.5559,
      "step": 1260
    },
    {
      "epoch": 1.59,
      "grad_norm": 0.5481686592102051,
      "learning_rate": 7.741719224885675e-06,
      "loss": 0.5928,
      "step": 1270
    },
    {
      "epoch": 1.61,
      "grad_norm": 0.620995044708252,
      "learning_rate": 7.708504050918288e-06,
      "loss": 0.5683,
      "step": 1280
    },
    {
      "epoch": 1.62,
      "grad_norm": 0.5698446035385132,
      "learning_rate": 7.675118841068507e-06,
      "loss": 0.5909,
      "step": 1290
    },
    {
      "epoch": 1.63,
      "grad_norm": 0.5968176126480103,
      "learning_rate": 7.64156569121048e-06,
      "loss": 0.573,
      "step": 1300
    },
    {
      "epoch": 1.64,
      "grad_norm": 0.7233667373657227,
      "learning_rate": 7.6078467077613746e-06,
      "loss": 0.5888,
      "step": 1310
    },
    {
      "epoch": 1.66,
      "grad_norm": 0.5293125510215759,
      "learning_rate": 7.5739640075491546e-06,
      "loss": 0.6316,
      "step": 1320
    },
    {
      "epoch": 1.67,
      "grad_norm": 0.5775229334831238,
      "learning_rate": 7.5399197176796735e-06,
      "loss": 0.591,
      "step": 1330
    },
    {
      "epoch": 1.68,
      "grad_norm": 0.7166782021522522,
      "learning_rate": 7.50571597540315e-06,
      "loss": 0.5505,
      "step": 1340
    },
    {
      "epoch": 1.69,
      "grad_norm": 0.5287649631500244,
      "learning_rate": 7.471354927979984e-06,
      "loss": 0.6036,
      "step": 1350
    },
    {
      "epoch": 1.71,
      "grad_norm": 0.5856715440750122,
      "learning_rate": 7.436838732545964e-06,
      "loss": 0.5682,
      "step": 1360
    },
    {
      "epoch": 1.72,
      "grad_norm": 0.5226353406906128,
      "learning_rate": 7.402169555976841e-06,
      "loss": 0.5593,
      "step": 1370
    },
    {
      "epoch": 1.73,
      "grad_norm": 0.6625097393989563,
      "learning_rate": 7.367349574752296e-06,
      "loss": 0.5853,
      "step": 1380
    },
    {
      "epoch": 1.74,
      "grad_norm": 0.5979090929031372,
      "learning_rate": 7.332380974819301e-06,
      "loss": 0.5493,
      "step": 1390
    },
    {
      "epoch": 1.76,
      "grad_norm": 0.6176837682723999,
      "learning_rate": 7.297265951454891e-06,
      "loss": 0.6038,
      "step": 1400
    },
    {
      "epoch": 1.76,
      "eval_loss": 0.5800889730453491,
      "eval_runtime": 69.6671,
      "eval_samples_per_second": 64.593,
      "eval_steps_per_second": 2.024,
      "step": 1400
    },
    {
      "epoch": 1.77,
      "grad_norm": 0.6878402233123779,
      "learning_rate": 7.262006709128351e-06,
      "loss": 0.5453,
      "step": 1410
    },
    {
      "epoch": 1.78,
      "grad_norm": 0.6541485786437988,
      "learning_rate": 7.226605461362817e-06,
      "loss": 0.586,
      "step": 1420
    },
    {
      "epoch": 1.79,
      "grad_norm": 0.6143850088119507,
      "learning_rate": 7.191064430596315e-06,
      "loss": 0.5422,
      "step": 1430
    },
    {
      "epoch": 1.81,
      "grad_norm": 0.6676700115203857,
      "learning_rate": 7.155385848042245e-06,
      "loss": 0.5806,
      "step": 1440
    },
    {
      "epoch": 1.82,
      "grad_norm": 0.6332250833511353,
      "learning_rate": 7.119571953549305e-06,
      "loss": 0.5585,
      "step": 1450
    },
    {
      "epoch": 1.83,
      "grad_norm": 0.7023981809616089,
      "learning_rate": 7.08362499546087e-06,
      "loss": 0.5833,
      "step": 1460
    },
    {
      "epoch": 1.84,
      "grad_norm": 0.665985107421875,
      "learning_rate": 7.047547230473861e-06,
      "loss": 0.569,
      "step": 1470
    },
    {
      "epoch": 1.86,
      "grad_norm": 0.5501468777656555,
      "learning_rate": 7.011340923497055e-06,
      "loss": 0.5824,
      "step": 1480
    },
    {
      "epoch": 1.87,
      "grad_norm": 0.5913044810295105,
      "learning_rate": 6.975008347508906e-06,
      "loss": 0.5581,
      "step": 1490
    },
    {
      "epoch": 1.88,
      "grad_norm": 0.6394723653793335,
      "learning_rate": 6.938551783414856e-06,
      "loss": 0.5813,
      "step": 1500
    },
    {
      "epoch": 1.89,
      "grad_norm": 0.6261583566665649,
      "learning_rate": 6.901973519904128e-06,
      "loss": 0.5804,
      "step": 1510
    },
    {
      "epoch": 1.91,
      "grad_norm": 0.5661801695823669,
      "learning_rate": 6.865275853306062e-06,
      "loss": 0.5587,
      "step": 1520
    },
    {
      "epoch": 1.92,
      "grad_norm": 0.5941905975341797,
      "learning_rate": 6.828461087445946e-06,
      "loss": 0.5735,
      "step": 1530
    },
    {
      "epoch": 1.93,
      "grad_norm": 0.5802645087242126,
      "learning_rate": 6.791531533500384e-06,
      "loss": 0.5854,
      "step": 1540
    },
    {
      "epoch": 1.94,
      "grad_norm": 0.5715972781181335,
      "learning_rate": 6.754489509852208e-06,
      "loss": 0.5627,
      "step": 1550
    },
    {
      "epoch": 1.96,
      "grad_norm": 0.6669459342956543,
      "learning_rate": 6.7173373419449295e-06,
      "loss": 0.5647,
      "step": 1560
    },
    {
      "epoch": 1.97,
      "grad_norm": 0.6907923221588135,
      "learning_rate": 6.68007736213676e-06,
      "loss": 0.5364,
      "step": 1570
    },
    {
      "epoch": 1.98,
      "grad_norm": 0.5516065359115601,
      "learning_rate": 6.6427119095541745e-06,
      "loss": 0.5643,
      "step": 1580
    },
    {
      "epoch": 1.99,
      "grad_norm": 0.6715050935745239,
      "learning_rate": 6.60524332994508e-06,
      "loss": 0.6026,
      "step": 1590
    },
    {
      "epoch": 2.01,
      "grad_norm": 0.5935114026069641,
      "learning_rate": 6.567673975531542e-06,
      "loss": 0.5882,
      "step": 1600
    },
    {
      "epoch": 2.01,
      "eval_loss": 0.5773123502731323,
      "eval_runtime": 71.3304,
      "eval_samples_per_second": 63.087,
      "eval_steps_per_second": 1.977,
      "step": 1600
    },
    {
      "epoch": 2.02,
      "grad_norm": 0.7686386108398438,
      "learning_rate": 6.530006204862121e-06,
      "loss": 0.5391,
      "step": 1610
    },
    {
      "epoch": 2.03,
      "grad_norm": 0.6634293794631958,
      "learning_rate": 6.492242382663804e-06,
      "loss": 0.5495,
      "step": 1620
    },
    {
      "epoch": 2.05,
      "grad_norm": 0.8547865748405457,
      "learning_rate": 6.454384879693549e-06,
      "loss": 0.5456,
      "step": 1630
    },
    {
      "epoch": 2.06,
      "grad_norm": 0.8082049489021301,
      "learning_rate": 6.41643607258946e-06,
      "loss": 0.5903,
      "step": 1640
    },
    {
      "epoch": 2.07,
      "grad_norm": 0.6522872447967529,
      "learning_rate": 6.378398343721575e-06,
      "loss": 0.5476,
      "step": 1650
    },
    {
      "epoch": 2.08,
      "grad_norm": 0.663383960723877,
      "learning_rate": 6.340274081042307e-06,
      "loss": 0.5271,
      "step": 1660
    },
    {
      "epoch": 2.1,
      "grad_norm": 0.8022051453590393,
      "learning_rate": 6.3020656779365416e-06,
      "loss": 0.5656,
      "step": 1670
    },
    {
      "epoch": 2.11,
      "grad_norm": 0.7163932919502258,
      "learning_rate": 6.263775533071368e-06,
      "loss": 0.5581,
      "step": 1680
    },
    {
      "epoch": 2.12,
      "grad_norm": 0.6052596569061279,
      "learning_rate": 6.225406050245507e-06,
      "loss": 0.554,
      "step": 1690
    },
    {
      "epoch": 2.13,
      "grad_norm": 0.6838263273239136,
      "learning_rate": 6.186959638238397e-06,
      "loss": 0.5812,
      "step": 1700
    },
    {
      "epoch": 2.15,
      "grad_norm": 0.6921390891075134,
      "learning_rate": 6.148438710658979e-06,
      "loss": 0.5237,
      "step": 1710
    },
    {
      "epoch": 2.16,
      "grad_norm": 0.883646547794342,
      "learning_rate": 6.109845685794169e-06,
      "loss": 0.5267,
      "step": 1720
    },
    {
      "epoch": 2.17,
      "grad_norm": 0.6446242928504944,
      "learning_rate": 6.071182986457046e-06,
      "loss": 0.5466,
      "step": 1730
    },
    {
      "epoch": 2.18,
      "grad_norm": 0.7913056015968323,
      "learning_rate": 6.032453039834748e-06,
      "loss": 0.573,
      "step": 1740
    },
    {
      "epoch": 2.2,
      "grad_norm": 0.7966864705085754,
      "learning_rate": 5.993658277336101e-06,
      "loss": 0.5498,
      "step": 1750
    },
    {
      "epoch": 2.21,
      "grad_norm": 0.7100770473480225,
      "learning_rate": 5.954801134438971e-06,
      "loss": 0.5624,
      "step": 1760
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.5544886589050293,
      "learning_rate": 5.915884050537378e-06,
      "loss": 0.5349,
      "step": 1770
    },
    {
      "epoch": 2.23,
      "grad_norm": 0.7371806502342224,
      "learning_rate": 5.8769094687883484e-06,
      "loss": 0.5789,
      "step": 1780
    },
    {
      "epoch": 2.25,
      "grad_norm": 0.7266082763671875,
      "learning_rate": 5.837879835958535e-06,
      "loss": 0.5573,
      "step": 1790
    },
    {
      "epoch": 2.26,
      "grad_norm": 0.6665763854980469,
      "learning_rate": 5.798797602270623e-06,
      "loss": 0.5302,
      "step": 1800
    },
    {
      "epoch": 2.26,
      "eval_loss": 0.5754047632217407,
      "eval_runtime": 72.9642,
      "eval_samples_per_second": 61.674,
      "eval_steps_per_second": 1.932,
      "step": 1800
    },
    {
      "epoch": 2.27,
      "grad_norm": 0.7089454531669617,
      "learning_rate": 5.7596652212494944e-06,
      "loss": 0.5652,
      "step": 1810
    },
    {
      "epoch": 2.28,
      "grad_norm": 0.790558397769928,
      "learning_rate": 5.720485149568211e-06,
      "loss": 0.5547,
      "step": 1820
    },
    {
      "epoch": 2.3,
      "grad_norm": 0.7344884872436523,
      "learning_rate": 5.681259846893783e-06,
      "loss": 0.5703,
      "step": 1830
    },
    {
      "epoch": 2.31,
      "grad_norm": 0.7799204587936401,
      "learning_rate": 5.641991775732756e-06,
      "loss": 0.5369,
      "step": 1840
    },
    {
      "epoch": 2.32,
      "grad_norm": 0.7437064051628113,
      "learning_rate": 5.6026834012766155e-06,
      "loss": 0.5814,
      "step": 1850
    },
    {
      "epoch": 2.33,
      "grad_norm": 0.7207692861557007,
      "learning_rate": 5.5633371912470334e-06,
      "loss": 0.5535,
      "step": 1860
    },
    {
      "epoch": 2.35,
      "grad_norm": 0.7650673985481262,
      "learning_rate": 5.5239556157409366e-06,
      "loss": 0.5507,
      "step": 1870
    },
    {
      "epoch": 2.36,
      "grad_norm": 0.8299598097801208,
      "learning_rate": 5.484541147075449e-06,
      "loss": 0.5678,
      "step": 1880
    },
    {
      "epoch": 2.37,
      "grad_norm": 0.6018086671829224,
      "learning_rate": 5.4450962596326776e-06,
      "loss": 0.5749,
      "step": 1890
    },
    {
      "epoch": 2.38,
      "grad_norm": 0.7759122848510742,
      "learning_rate": 5.405623429704373e-06,
      "loss": 0.5561,
      "step": 1900
    },
    {
      "epoch": 2.4,
      "grad_norm": 0.7874650359153748,
      "learning_rate": 5.366125135336477e-06,
      "loss": 0.568,
      "step": 1910
    },
    {
      "epoch": 2.41,
      "grad_norm": 0.642218828201294,
      "learning_rate": 5.326603856173548e-06,
      "loss": 0.5468,
      "step": 1920
    },
    {
      "epoch": 2.42,
      "grad_norm": 0.7558509707450867,
      "learning_rate": 5.2870620733030994e-06,
      "loss": 0.5315,
      "step": 1930
    },
    {
      "epoch": 2.43,
      "grad_norm": 0.846817672252655,
      "learning_rate": 5.247502269099834e-06,
      "loss": 0.5511,
      "step": 1940
    },
    {
      "epoch": 2.45,
      "grad_norm": 0.725319504737854,
      "learning_rate": 5.207926927069809e-06,
      "loss": 0.5098,
      "step": 1950
    },
    {
      "epoch": 2.46,
      "grad_norm": 0.7795179486274719,
      "learning_rate": 5.168338531694523e-06,
      "loss": 0.5417,
      "step": 1960
    },
    {
      "epoch": 2.47,
      "grad_norm": 0.764240562915802,
      "learning_rate": 5.1287395682749444e-06,
      "loss": 0.5502,
      "step": 1970
    },
    {
      "epoch": 2.48,
      "grad_norm": 0.7871197462081909,
      "learning_rate": 5.089132522775486e-06,
      "loss": 0.575,
      "step": 1980
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.8167646527290344,
      "learning_rate": 5.049519881667945e-06,
      "loss": 0.5637,
      "step": 1990
    },
    {
      "epoch": 2.51,
      "grad_norm": 0.7696388959884644,
      "learning_rate": 5.009904131775402e-06,
      "loss": 0.5331,
      "step": 2000
    },
    {
      "epoch": 2.51,
      "eval_loss": 0.5732148289680481,
      "eval_runtime": 73.9926,
      "eval_samples_per_second": 60.817,
      "eval_steps_per_second": 1.906,
      "step": 2000
    },
    {
      "epoch": 2.52,
      "grad_norm": 0.8611487746238708,
      "learning_rate": 4.970287760116098e-06,
      "loss": 0.544,
      "step": 2010
    },
    {
      "epoch": 2.53,
      "grad_norm": 0.7960777878761292,
      "learning_rate": 4.930673253747314e-06,
      "loss": 0.5702,
      "step": 2020
    },
    {
      "epoch": 2.55,
      "grad_norm": 0.8783027529716492,
      "learning_rate": 4.8910630996092265e-06,
      "loss": 0.5309,
      "step": 2030
    },
    {
      "epoch": 2.56,
      "grad_norm": 0.7091485261917114,
      "learning_rate": 4.851459784368787e-06,
      "loss": 0.5565,
      "step": 2040
    },
    {
      "epoch": 2.57,
      "grad_norm": 0.7412302494049072,
      "learning_rate": 4.811865794263611e-06,
      "loss": 0.5546,
      "step": 2050
    },
    {
      "epoch": 2.58,
      "grad_norm": 0.7188848257064819,
      "learning_rate": 4.772283614945895e-06,
      "loss": 0.5927,
      "step": 2060
    },
    {
      "epoch": 2.6,
      "grad_norm": 0.9045892953872681,
      "learning_rate": 4.732715731326372e-06,
      "loss": 0.5515,
      "step": 2070
    },
    {
      "epoch": 2.61,
      "grad_norm": 0.7781686186790466,
      "learning_rate": 4.693164627418315e-06,
      "loss": 0.5289,
      "step": 2080
    },
    {
      "epoch": 2.62,
      "grad_norm": 0.8868610858917236,
      "learning_rate": 4.653632786181587e-06,
      "loss": 0.5934,
      "step": 2090
    },
    {
      "epoch": 2.63,
      "grad_norm": 0.7511059641838074,
      "learning_rate": 4.614122689366769e-06,
      "loss": 0.5464,
      "step": 2100
    },
    {
      "epoch": 2.65,
      "grad_norm": 0.7453913688659668,
      "learning_rate": 4.574636817359365e-06,
      "loss": 0.5318,
      "step": 2110
    },
    {
      "epoch": 2.66,
      "grad_norm": 0.6865512132644653,
      "learning_rate": 4.535177649024077e-06,
      "loss": 0.5527,
      "step": 2120
    },
    {
      "epoch": 2.67,
      "grad_norm": 0.8288848400115967,
      "learning_rate": 4.495747661549187e-06,
      "loss": 0.5368,
      "step": 2130
    },
    {
      "epoch": 2.69,
      "grad_norm": 0.8635351657867432,
      "learning_rate": 4.456349330291055e-06,
      "loss": 0.5549,
      "step": 2140
    },
    {
      "epoch": 2.7,
      "grad_norm": 0.7543601393699646,
      "learning_rate": 4.4169851286187005e-06,
      "loss": 0.5435,
      "step": 2150
    },
    {
      "epoch": 2.71,
      "grad_norm": 0.8399900794029236,
      "learning_rate": 4.3776575277585445e-06,
      "loss": 0.5853,
      "step": 2160
    },
    {
      "epoch": 2.72,
      "grad_norm": 0.9056334495544434,
      "learning_rate": 4.3383689966392604e-06,
      "loss": 0.5737,
      "step": 2170
    },
    {
      "epoch": 2.74,
      "grad_norm": 0.7865429520606995,
      "learning_rate": 4.2991220017367795e-06,
      "loss": 0.5908,
      "step": 2180
    },
    {
      "epoch": 2.75,
      "grad_norm": 0.9085346460342407,
      "learning_rate": 4.259919006919456e-06,
      "loss": 0.5468,
      "step": 2190
    },
    {
      "epoch": 2.76,
      "grad_norm": 0.844716489315033,
      "learning_rate": 4.220762473293379e-06,
      "loss": 0.5669,
      "step": 2200
    },
    {
      "epoch": 2.76,
      "eval_loss": 0.5714393854141235,
      "eval_runtime": 74.9812,
      "eval_samples_per_second": 60.015,
      "eval_steps_per_second": 1.88,
      "step": 2200
    },
    {
      "epoch": 2.77,
      "grad_norm": 0.8216796517372131,
      "learning_rate": 4.181654859047873e-06,
      "loss": 0.5516,
      "step": 2210
    },
    {
      "epoch": 2.79,
      "grad_norm": 0.8482044339179993,
      "learning_rate": 4.1425986193011805e-06,
      "loss": 0.5463,
      "step": 2220
    },
    {
      "epoch": 2.8,
      "grad_norm": 0.8037722706794739,
      "learning_rate": 4.103596205946323e-06,
      "loss": 0.5387,
      "step": 2230
    },
    {
      "epoch": 2.81,
      "grad_norm": 0.8523209691047668,
      "learning_rate": 4.0646500674971826e-06,
      "loss": 0.5282,
      "step": 2240
    },
    {
      "epoch": 2.82,
      "grad_norm": 0.8726175427436829,
      "learning_rate": 4.0257626489347895e-06,
      "loss": 0.5476,
      "step": 2250
    },
    {
      "epoch": 2.84,
      "grad_norm": 0.9041292071342468,
      "learning_rate": 3.986936391553822e-06,
      "loss": 0.5677,
      "step": 2260
    },
    {
      "epoch": 2.85,
      "grad_norm": 0.8939766883850098,
      "learning_rate": 3.948173732809351e-06,
      "loss": 0.5711,
      "step": 2270
    },
    {
      "epoch": 2.86,
      "grad_norm": 0.7777726650238037,
      "learning_rate": 3.909477106163818e-06,
      "loss": 0.5289,
      "step": 2280
    },
    {
      "epoch": 2.87,
      "grad_norm": 0.7699301242828369,
      "learning_rate": 3.87084894093427e-06,
      "loss": 0.55,
      "step": 2290
    },
    {
      "epoch": 2.89,
      "grad_norm": 0.9905028939247131,
      "learning_rate": 3.832291662139845e-06,
      "loss": 0.541,
      "step": 2300
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.8651571869850159,
      "learning_rate": 3.7938076903495363e-06,
      "loss": 0.5533,
      "step": 2310
    },
    {
      "epoch": 2.91,
      "grad_norm": 0.7639217376708984,
      "learning_rate": 3.7553994415302274e-06,
      "loss": 0.5511,
      "step": 2320
    },
    {
      "epoch": 2.92,
      "grad_norm": 0.8610684871673584,
      "learning_rate": 3.7170693268950354e-06,
      "loss": 0.5119,
      "step": 2330
    },
    {
      "epoch": 2.94,
      "grad_norm": 0.7909624576568604,
      "learning_rate": 3.67881975275192e-06,
      "loss": 0.5881,
      "step": 2340
    },
    {
      "epoch": 2.95,
      "grad_norm": 0.9253331422805786,
      "learning_rate": 3.6406531203526274e-06,
      "loss": 0.5612,
      "step": 2350
    },
    {
      "epoch": 2.96,
      "grad_norm": 0.8581608533859253,
      "learning_rate": 3.6025718257419532e-06,
      "loss": 0.5322,
      "step": 2360
    },
    {
      "epoch": 2.97,
      "grad_norm": 0.9016821384429932,
      "learning_rate": 3.564578259607303e-06,
      "loss": 0.5487,
      "step": 2370
    },
    {
      "epoch": 2.99,
      "grad_norm": 0.8990395665168762,
      "learning_rate": 3.5266748071286205e-06,
      "loss": 0.5471,
      "step": 2380
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9548435211181641,
      "learning_rate": 3.488863847828652e-06,
      "loss": 0.547,
      "step": 2390
    },
    {
      "epoch": 3.01,
      "grad_norm": 0.771557092666626,
      "learning_rate": 3.4511477554235516e-06,
      "loss": 0.5551,
      "step": 2400
    },
    {
      "epoch": 3.01,
      "eval_loss": 0.5705960392951965,
      "eval_runtime": 76.5829,
      "eval_samples_per_second": 58.76,
      "eval_steps_per_second": 1.841,
      "step": 2400
    },
    {
      "epoch": 3.02,
      "grad_norm": 0.8474746942520142,
      "learning_rate": 3.4135288976738762e-06,
      "loss": 0.5552,
      "step": 2410
    },
    {
      "epoch": 3.04,
      "grad_norm": 1.0370606184005737,
      "learning_rate": 3.3760096362359306e-06,
      "loss": 0.5456,
      "step": 2420
    },
    {
      "epoch": 3.05,
      "grad_norm": 0.7640136480331421,
      "learning_rate": 3.33859232651351e-06,
      "loss": 0.5228,
      "step": 2430
    },
    {
      "epoch": 3.06,
      "grad_norm": 1.0315712690353394,
      "learning_rate": 3.301279317510038e-06,
      "loss": 0.5389,
      "step": 2440
    },
    {
      "epoch": 3.07,
      "grad_norm": 1.0031884908676147,
      "learning_rate": 3.2640729516810844e-06,
      "loss": 0.5153,
      "step": 2450
    },
    {
      "epoch": 3.09,
      "grad_norm": 0.8636874556541443,
      "learning_rate": 3.226975564787322e-06,
      "loss": 0.527,
      "step": 2460
    },
    {
      "epoch": 3.1,
      "grad_norm": 0.9102957248687744,
      "learning_rate": 3.1899894857478897e-06,
      "loss": 0.5397,
      "step": 2470
    },
    {
      "epoch": 3.11,
      "grad_norm": 0.8529887795448303,
      "learning_rate": 3.15311703649418e-06,
      "loss": 0.5184,
      "step": 2480
    },
    {
      "epoch": 3.12,
      "grad_norm": 0.9462995529174805,
      "learning_rate": 3.116360531824074e-06,
      "loss": 0.5559,
      "step": 2490
    },
    {
      "epoch": 3.14,
      "grad_norm": 0.8420120477676392,
      "learning_rate": 3.0797222792566306e-06,
      "loss": 0.5245,
      "step": 2500
    },
    {
      "epoch": 3.15,
      "grad_norm": 0.7965397238731384,
      "learning_rate": 3.043204578887209e-06,
      "loss": 0.5585,
      "step": 2510
    },
    {
      "epoch": 3.16,
      "grad_norm": 0.9633888602256775,
      "learning_rate": 3.00680972324308e-06,
      "loss": 0.4956,
      "step": 2520
    },
    {
      "epoch": 3.17,
      "grad_norm": 1.1136802434921265,
      "learning_rate": 2.970539997139509e-06,
      "loss": 0.5444,
      "step": 2530
    },
    {
      "epoch": 3.19,
      "grad_norm": 0.8797339797019958,
      "learning_rate": 2.9343976775363097e-06,
      "loss": 0.5253,
      "step": 2540
    },
    {
      "epoch": 3.2,
      "grad_norm": 1.0063321590423584,
      "learning_rate": 2.8983850333949065e-06,
      "loss": 0.5427,
      "step": 2550
    },
    {
      "epoch": 3.21,
      "grad_norm": 0.8839291930198669,
      "learning_rate": 2.862504325535886e-06,
      "loss": 0.5346,
      "step": 2560
    },
    {
      "epoch": 3.22,
      "grad_norm": 0.9732418060302734,
      "learning_rate": 2.8267578064970735e-06,
      "loss": 0.5889,
      "step": 2570
    },
    {
      "epoch": 3.24,
      "grad_norm": 1.0936757326126099,
      "learning_rate": 2.791147720392118e-06,
      "loss": 0.5136,
      "step": 2580
    },
    {
      "epoch": 3.25,
      "grad_norm": 0.9694124460220337,
      "learning_rate": 2.755676302769611e-06,
      "loss": 0.5276,
      "step": 2590
    },
    {
      "epoch": 3.26,
      "grad_norm": 0.9088903069496155,
      "learning_rate": 2.7203457804727364e-06,
      "loss": 0.5499,
      "step": 2600
    },
    {
      "epoch": 3.26,
      "eval_loss": 0.5715650320053101,
      "eval_runtime": 75.4356,
      "eval_samples_per_second": 59.653,
      "eval_steps_per_second": 1.869,
      "step": 2600
    },
    {
      "epoch": 3.27,
      "grad_norm": 0.8832049369812012,
      "learning_rate": 2.6851583714994843e-06,
      "loss": 0.5392,
      "step": 2610
    },
    {
      "epoch": 3.29,
      "grad_norm": 1.0563007593154907,
      "learning_rate": 2.6501162848634023e-06,
      "loss": 0.5952,
      "step": 2620
    },
    {
      "epoch": 3.3,
      "grad_norm": 0.9844443798065186,
      "learning_rate": 2.6152217204549045e-06,
      "loss": 0.5304,
      "step": 2630
    },
    {
      "epoch": 3.31,
      "grad_norm": 1.0739080905914307,
      "learning_rate": 2.580476868903197e-06,
      "loss": 0.4941,
      "step": 2640
    },
    {
      "epoch": 3.32,
      "grad_norm": 0.8901020884513855,
      "learning_rate": 2.5458839114387146e-06,
      "loss": 0.5418,
      "step": 2650
    },
    {
      "epoch": 3.34,
      "grad_norm": 0.9322318434715271,
      "learning_rate": 2.5114450197562214e-06,
      "loss": 0.4952,
      "step": 2660
    },
    {
      "epoch": 3.35,
      "grad_norm": 0.9946108460426331,
      "learning_rate": 2.4771623558784474e-06,
      "loss": 0.5155,
      "step": 2670
    },
    {
      "epoch": 3.36,
      "grad_norm": 0.9886655211448669,
      "learning_rate": 2.4430380720203785e-06,
      "loss": 0.5374,
      "step": 2680
    },
    {
      "epoch": 3.38,
      "grad_norm": 0.9973739385604858,
      "learning_rate": 2.409074310454136e-06,
      "loss": 0.5055,
      "step": 2690
    },
    {
      "epoch": 3.39,
      "grad_norm": 0.9865736961364746,
      "learning_rate": 2.375273203374489e-06,
      "loss": 0.5287,
      "step": 2700
    },
    {
      "epoch": 3.4,
      "grad_norm": 0.9992399215698242,
      "learning_rate": 2.341636872764994e-06,
      "loss": 0.5753,
      "step": 2710
    },
    {
      "epoch": 3.41,
      "grad_norm": 1.0160070657730103,
      "learning_rate": 2.3081674302647905e-06,
      "loss": 0.5102,
      "step": 2720
    },
    {
      "epoch": 3.43,
      "grad_norm": 0.7992890477180481,
      "learning_rate": 2.2748669770360273e-06,
      "loss": 0.5551,
      "step": 2730
    },
    {
      "epoch": 3.44,
      "grad_norm": 1.0403869152069092,
      "learning_rate": 2.2417376036319495e-06,
      "loss": 0.5231,
      "step": 2740
    },
    {
      "epoch": 3.45,
      "grad_norm": 0.9026238918304443,
      "learning_rate": 2.2087813898656775e-06,
      "loss": 0.5148,
      "step": 2750
    },
    {
      "epoch": 3.46,
      "grad_norm": 1.078641414642334,
      "learning_rate": 2.176000404679612e-06,
      "loss": 0.5481,
      "step": 2760
    },
    {
      "epoch": 3.48,
      "grad_norm": 0.8496494889259338,
      "learning_rate": 2.1433967060155687e-06,
      "loss": 0.4961,
      "step": 2770
    },
    {
      "epoch": 3.49,
      "grad_norm": 0.9248107671737671,
      "learning_rate": 2.110972340685577e-06,
      "loss": 0.5402,
      "step": 2780
    },
    {
      "epoch": 3.5,
      "grad_norm": 1.0704095363616943,
      "learning_rate": 2.078729344243384e-06,
      "loss": 0.5228,
      "step": 2790
    },
    {
      "epoch": 3.51,
      "grad_norm": 0.8805019855499268,
      "learning_rate": 2.0466697408566664e-06,
      "loss": 0.545,
      "step": 2800
    },
    {
      "epoch": 3.51,
      "eval_loss": 0.5704147815704346,
      "eval_runtime": 77.346,
      "eval_samples_per_second": 58.18,
      "eval_steps_per_second": 1.823,
      "step": 2800
    },
    {
      "epoch": 3.53,
      "grad_norm": 1.0108119249343872,
      "learning_rate": 2.014795543179957e-06,
      "loss": 0.5361,
      "step": 2810
    },
    {
      "epoch": 3.54,
      "grad_norm": 1.01576828956604,
      "learning_rate": 1.9831087522282865e-06,
      "loss": 0.5552,
      "step": 2820
    },
    {
      "epoch": 3.55,
      "grad_norm": 0.9220107197761536,
      "learning_rate": 1.951611357251575e-06,
      "loss": 0.5753,
      "step": 2830
    },
    {
      "epoch": 3.56,
      "grad_norm": 0.9482864141464233,
      "learning_rate": 1.92030533560974e-06,
      "loss": 0.5339,
      "step": 2840
    },
    {
      "epoch": 3.58,
      "grad_norm": 0.9235343933105469,
      "learning_rate": 1.8891926526485643e-06,
      "loss": 0.538,
      "step": 2850
    },
    {
      "epoch": 3.59,
      "grad_norm": 0.6921684145927429,
      "learning_rate": 1.8582752615763161e-06,
      "loss": 0.5331,
      "step": 2860
    },
    {
      "epoch": 3.6,
      "grad_norm": 0.930234968662262,
      "learning_rate": 1.8275551033411209e-06,
      "loss": 0.5226,
      "step": 2870
    },
    {
      "epoch": 3.61,
      "grad_norm": 0.9894780516624451,
      "learning_rate": 1.7970341065091246e-06,
      "loss": 0.5294,
      "step": 2880
    },
    {
      "epoch": 3.63,
      "grad_norm": 1.0227577686309814,
      "learning_rate": 1.766714187143413e-06,
      "loss": 0.5186,
      "step": 2890
    },
    {
      "epoch": 3.64,
      "grad_norm": 0.9115012884140015,
      "learning_rate": 1.736597248683724e-06,
      "loss": 0.5156,
      "step": 2900
    },
    {
      "epoch": 3.65,
      "grad_norm": 1.050728440284729,
      "learning_rate": 1.7066851818269558e-06,
      "loss": 0.521,
      "step": 2910
    },
    {
      "epoch": 3.66,
      "grad_norm": 1.0954750776290894,
      "learning_rate": 1.676979864408471e-06,
      "loss": 0.5451,
      "step": 2920
    },
    {
      "epoch": 3.68,
      "grad_norm": 0.7569468021392822,
      "learning_rate": 1.647483161284203e-06,
      "loss": 0.5765,
      "step": 2930
    },
    {
      "epoch": 3.69,
      "grad_norm": 0.8553792238235474,
      "learning_rate": 1.6181969242135926e-06,
      "loss": 0.5335,
      "step": 2940
    },
    {
      "epoch": 3.7,
      "grad_norm": 0.918138861656189,
      "learning_rate": 1.5891229917433316e-06,
      "loss": 0.5343,
      "step": 2950
    },
    {
      "epoch": 3.71,
      "grad_norm": 0.988727867603302,
      "learning_rate": 1.5602631890919428e-06,
      "loss": 0.5288,
      "step": 2960
    },
    {
      "epoch": 3.73,
      "grad_norm": 1.130867600440979,
      "learning_rate": 1.531619328035197e-06,
      "loss": 0.4952,
      "step": 2970
    },
    {
      "epoch": 3.74,
      "grad_norm": 1.1033731698989868,
      "learning_rate": 1.503193206792366e-06,
      "loss": 0.5053,
      "step": 2980
    },
    {
      "epoch": 3.75,
      "grad_norm": 1.0361392498016357,
      "learning_rate": 1.4749866099133409e-06,
      "loss": 0.555,
      "step": 2990
    },
    {
      "epoch": 3.76,
      "grad_norm": 1.1811940670013428,
      "learning_rate": 1.447001308166599e-06,
      "loss": 0.5253,
      "step": 3000
    },
    {
      "epoch": 3.76,
      "eval_loss": 0.5704381465911865,
      "eval_runtime": 72.5701,
      "eval_samples_per_second": 62.009,
      "eval_steps_per_second": 1.943,
      "step": 3000
    },
    {
      "epoch": 3.78,
      "grad_norm": 1.0649921894073486,
      "learning_rate": 1.4192390584280347e-06,
      "loss": 0.5397,
      "step": 3010
    },
    {
      "epoch": 3.79,
      "grad_norm": 1.006032943725586,
      "learning_rate": 1.3917016035706632e-06,
      "loss": 0.5321,
      "step": 3020
    },
    {
      "epoch": 3.8,
      "grad_norm": 1.1129940748214722,
      "learning_rate": 1.3643906723552186e-06,
      "loss": 0.5246,
      "step": 3030
    },
    {
      "epoch": 3.81,
      "grad_norm": 1.1166002750396729,
      "learning_rate": 1.3373079793216037e-06,
      "loss": 0.5405,
      "step": 3040
    },
    {
      "epoch": 3.83,
      "grad_norm": 0.9260038137435913,
      "learning_rate": 1.3104552246812768e-06,
      "loss": 0.543,
      "step": 3050
    },
    {
      "epoch": 3.84,
      "grad_norm": 1.0943797826766968,
      "learning_rate": 1.2838340942104944e-06,
      "loss": 0.5147,
      "step": 3060
    },
    {
      "epoch": 3.85,
      "grad_norm": 1.0058125257492065,
      "learning_rate": 1.257446259144494e-06,
      "loss": 0.4995,
      "step": 3070
    },
    {
      "epoch": 3.86,
      "grad_norm": 1.022620677947998,
      "learning_rate": 1.2312933760725753e-06,
      "loss": 0.5485,
      "step": 3080
    },
    {
      "epoch": 3.88,
      "grad_norm": 1.077143907546997,
      "learning_rate": 1.2053770868340914e-06,
      "loss": 0.5483,
      "step": 3090
    },
    {
      "epoch": 3.89,
      "grad_norm": 1.3038913011550903,
      "learning_rate": 1.17969901841539e-06,
      "loss": 0.5388,
      "step": 3100
    },
    {
      "epoch": 3.9,
      "grad_norm": 0.8922839760780334,
      "learning_rate": 1.1542607828476665e-06,
      "loss": 0.5404,
      "step": 3110
    },
    {
      "epoch": 3.91,
      "grad_norm": 0.9709892272949219,
      "learning_rate": 1.1290639771057648e-06,
      "loss": 0.5204,
      "step": 3120
    },
    {
      "epoch": 3.93,
      "grad_norm": 1.1010524034500122,
      "learning_rate": 1.1041101830079159e-06,
      "loss": 0.5429,
      "step": 3130
    },
    {
      "epoch": 3.94,
      "grad_norm": 1.1041542291641235,
      "learning_rate": 1.0794009671164484e-06,
      "loss": 0.5256,
      "step": 3140
    },
    {
      "epoch": 3.95,
      "grad_norm": 1.1114507913589478,
      "learning_rate": 1.054937880639424e-06,
      "loss": 0.5185,
      "step": 3150
    },
    {
      "epoch": 3.96,
      "grad_norm": 0.9926356077194214,
      "learning_rate": 1.030722459333267e-06,
      "loss": 0.5161,
      "step": 3160
    },
    {
      "epoch": 3.98,
      "grad_norm": 1.00425386428833,
      "learning_rate": 1.0067562234063477e-06,
      "loss": 0.5222,
      "step": 3170
    },
    {
      "epoch": 3.99,
      "grad_norm": 1.0224225521087646,
      "learning_rate": 9.830406774235474e-07,
      "loss": 0.521,
      "step": 3180
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9822995662689209,
      "learning_rate": 9.595773102118022e-07,
      "loss": 0.528,
      "step": 3190
    },
    {
      "epoch": 4.02,
      "grad_norm": 1.1153786182403564,
      "learning_rate": 9.363675947666362e-07,
      "loss": 0.5418,
      "step": 3200
    },
    {
      "epoch": 4.02,
      "eval_loss": 0.5702469944953918,
      "eval_runtime": 71.9125,
      "eval_samples_per_second": 62.576,
      "eval_steps_per_second": 1.961,
      "step": 3200
    },
    {
      "epoch": 4.03,
      "grad_norm": 1.056578278541565,
      "learning_rate": 9.134129881596926e-07,
      "loss": 0.5326,
      "step": 3210
    },
    {
      "epoch": 4.04,
      "grad_norm": 0.9958868026733398,
      "learning_rate": 8.907149314472579e-07,
      "loss": 0.558,
      "step": 3220
    },
    {
      "epoch": 4.05,
      "grad_norm": 1.1032071113586426,
      "learning_rate": 8.682748495797965e-07,
      "loss": 0.5423,
      "step": 3230
    },
    {
      "epoch": 4.07,
      "grad_norm": 1.0517561435699463,
      "learning_rate": 8.460941513124876e-07,
      "loss": 0.5189,
      "step": 3240
    },
    {
      "epoch": 4.08,
      "grad_norm": 1.1289842128753662,
      "learning_rate": 8.241742291168015e-07,
      "loss": 0.5002,
      "step": 3250
    },
    {
      "epoch": 4.09,
      "grad_norm": 0.9873390793800354,
      "learning_rate": 8.025164590930623e-07,
      "loss": 0.5589,
      "step": 3260
    },
    {
      "epoch": 4.1,
      "grad_norm": 0.885833740234375,
      "learning_rate": 7.811222008840719e-07,
      "loss": 0.5031,
      "step": 3270
    },
    {
      "epoch": 4.12,
      "grad_norm": 0.7554776072502136,
      "learning_rate": 7.599927975897514e-07,
      "loss": 0.5128,
      "step": 3280
    },
    {
      "epoch": 4.13,
      "grad_norm": 1.0165417194366455,
      "learning_rate": 7.391295756828187e-07,
      "loss": 0.5369,
      "step": 3290
    },
    {
      "epoch": 4.14,
      "grad_norm": 0.9958201050758362,
      "learning_rate": 7.185338449255197e-07,
      "loss": 0.4948,
      "step": 3300
    },
    {
      "epoch": 4.15,
      "grad_norm": 1.1157667636871338,
      "learning_rate": 6.982068982874019e-07,
      "loss": 0.5181,
      "step": 3310
    },
    {
      "epoch": 4.17,
      "grad_norm": 1.0274876356124878,
      "learning_rate": 6.781500118641377e-07,
      "loss": 0.5172,
      "step": 3320
    },
    {
      "epoch": 4.18,
      "grad_norm": 0.9985108375549316,
      "learning_rate": 6.583644447974241e-07,
      "loss": 0.5304,
      "step": 3330
    },
    {
      "epoch": 4.19,
      "grad_norm": 1.0142953395843506,
      "learning_rate": 6.388514391959283e-07,
      "loss": 0.5278,
      "step": 3340
    },
    {
      "epoch": 4.2,
      "grad_norm": 1.0171020030975342,
      "learning_rate": 6.196122200573073e-07,
      "loss": 0.5269,
      "step": 3350
    },
    {
      "epoch": 4.22,
      "grad_norm": 1.2027987241744995,
      "learning_rate": 6.006479951913168e-07,
      "loss": 0.5379,
      "step": 3360
    },
    {
      "epoch": 4.23,
      "grad_norm": 1.064353108406067,
      "learning_rate": 5.819599551439703e-07,
      "loss": 0.4879,
      "step": 3370
    },
    {
      "epoch": 4.24,
      "grad_norm": 1.0336158275604248,
      "learning_rate": 5.635492731228098e-07,
      "loss": 0.5363,
      "step": 3380
    },
    {
      "epoch": 4.25,
      "grad_norm": 1.143526554107666,
      "learning_rate": 5.45417104923251e-07,
      "loss": 0.5371,
      "step": 3390
    },
    {
      "epoch": 4.27,
      "grad_norm": 0.9199160933494568,
      "learning_rate": 5.275645888560233e-07,
      "loss": 0.5289,
      "step": 3400
    },
    {
      "epoch": 4.27,
      "eval_loss": 0.570408046245575,
      "eval_runtime": 72.3442,
      "eval_samples_per_second": 62.203,
      "eval_steps_per_second": 1.949,
      "step": 3400
    },
    {
      "epoch": 4.28,
      "grad_norm": 1.0219011306762695,
      "learning_rate": 5.099928456757036e-07,
      "loss": 0.5115,
      "step": 3410
    },
    {
      "epoch": 4.29,
      "grad_norm": 1.0187703371047974,
      "learning_rate": 4.927029785103715e-07,
      "loss": 0.5171,
      "step": 3420
    },
    {
      "epoch": 4.3,
      "grad_norm": 1.006913423538208,
      "learning_rate": 4.756960727923371e-07,
      "loss": 0.5113,
      "step": 3430
    },
    {
      "epoch": 4.32,
      "grad_norm": 1.0249481201171875,
      "learning_rate": 4.5897319619001404e-07,
      "loss": 0.552,
      "step": 3440
    },
    {
      "epoch": 4.33,
      "grad_norm": 1.016176462173462,
      "learning_rate": 4.42535398540887e-07,
      "loss": 0.5246,
      "step": 3450
    },
    {
      "epoch": 4.34,
      "grad_norm": 0.9975274205207825,
      "learning_rate": 4.263837117856046e-07,
      "loss": 0.5328,
      "step": 3460
    },
    {
      "epoch": 4.35,
      "grad_norm": 0.8603295683860779,
      "learning_rate": 4.10519149903198e-07,
      "loss": 0.5409,
      "step": 3470
    },
    {
      "epoch": 4.37,
      "grad_norm": 1.079770565032959,
      "learning_rate": 3.949427088474189e-07,
      "loss": 0.5506,
      "step": 3480
    },
    {
      "epoch": 4.38,
      "grad_norm": 0.9214140176773071,
      "learning_rate": 3.79655366484224e-07,
      "loss": 0.5175,
      "step": 3490
    },
    {
      "epoch": 4.39,
      "grad_norm": 0.9992761015892029,
      "learning_rate": 3.646580825303786e-07,
      "loss": 0.5197,
      "step": 3500
    },
    {
      "epoch": 4.4,
      "grad_norm": 0.9221498370170593,
      "learning_rate": 3.49951798493211e-07,
      "loss": 0.5083,
      "step": 3510
    },
    {
      "epoch": 4.42,
      "grad_norm": 1.0544548034667969,
      "learning_rate": 3.355374376115017e-07,
      "loss": 0.5187,
      "step": 3520
    },
    {
      "epoch": 4.43,
      "grad_norm": 0.8989390134811401,
      "learning_rate": 3.214159047975324e-07,
      "loss": 0.5137,
      "step": 3530
    },
    {
      "epoch": 4.44,
      "grad_norm": 0.9407683610916138,
      "learning_rate": 3.0758808658026496e-07,
      "loss": 0.5045,
      "step": 3540
    },
    {
      "epoch": 4.45,
      "grad_norm": 1.034885287284851,
      "learning_rate": 2.940548510496977e-07,
      "loss": 0.5369,
      "step": 3550
    },
    {
      "epoch": 4.47,
      "grad_norm": 0.9151519536972046,
      "learning_rate": 2.8081704780236083e-07,
      "loss": 0.5619,
      "step": 3560
    },
    {
      "epoch": 4.48,
      "grad_norm": 1.22360098361969,
      "learning_rate": 2.6787550788798243e-07,
      "loss": 0.5078,
      "step": 3570
    },
    {
      "epoch": 4.49,
      "grad_norm": 1.11081063747406,
      "learning_rate": 2.55231043757318e-07,
      "loss": 0.5391,
      "step": 3580
    },
    {
      "epoch": 4.5,
      "grad_norm": 0.9431936144828796,
      "learning_rate": 2.428844492111415e-07,
      "loss": 0.581,
      "step": 3590
    },
    {
      "epoch": 4.52,
      "grad_norm": 0.993818461894989,
      "learning_rate": 2.3083649935041642e-07,
      "loss": 0.4983,
      "step": 3600
    },
    {
      "epoch": 4.52,
      "eval_loss": 0.570551335811615,
      "eval_runtime": 73.4513,
      "eval_samples_per_second": 61.265,
      "eval_steps_per_second": 1.92,
      "step": 3600
    },
    {
      "epoch": 4.53,
      "grad_norm": 0.9568440318107605,
      "learning_rate": 2.1908795052763388e-07,
      "loss": 0.5094,
      "step": 3610
    },
    {
      "epoch": 4.54,
      "grad_norm": 1.0978604555130005,
      "learning_rate": 2.0763954029933085e-07,
      "loss": 0.5262,
      "step": 3620
    },
    {
      "epoch": 4.55,
      "grad_norm": 0.9831346273422241,
      "learning_rate": 1.9649198737978325e-07,
      "loss": 0.5381,
      "step": 3630
    },
    {
      "epoch": 4.57,
      "grad_norm": 1.1549886465072632,
      "learning_rate": 1.856459915958958e-07,
      "loss": 0.5069,
      "step": 3640
    },
    {
      "epoch": 4.58,
      "grad_norm": 1.122920274734497,
      "learning_rate": 1.7510223384325674e-07,
      "loss": 0.5323,
      "step": 3650
    },
    {
      "epoch": 4.59,
      "grad_norm": 1.0226640701293945,
      "learning_rate": 1.6486137604339813e-07,
      "loss": 0.5206,
      "step": 3660
    },
    {
      "epoch": 4.6,
      "grad_norm": 0.9887501001358032,
      "learning_rate": 1.5492406110224289e-07,
      "loss": 0.549,
      "step": 3670
    },
    {
      "epoch": 4.62,
      "grad_norm": 1.069175124168396,
      "learning_rate": 1.4529091286973994e-07,
      "loss": 0.535,
      "step": 3680
    },
    {
      "epoch": 4.63,
      "grad_norm": 1.0605347156524658,
      "learning_rate": 1.3596253610070165e-07,
      "loss": 0.4874,
      "step": 3690
    },
    {
      "epoch": 4.64,
      "grad_norm": 1.1193597316741943,
      "learning_rate": 1.2693951641683856e-07,
      "loss": 0.5137,
      "step": 3700
    },
    {
      "epoch": 4.65,
      "grad_norm": 1.0010859966278076,
      "learning_rate": 1.1822242026999443e-07,
      "loss": 0.4834,
      "step": 3710
    },
    {
      "epoch": 4.67,
      "grad_norm": 1.1622557640075684,
      "learning_rate": 1.0981179490658522e-07,
      "loss": 0.5739,
      "step": 3720
    },
    {
      "epoch": 4.68,
      "grad_norm": 1.0148837566375732,
      "learning_rate": 1.0170816833324549e-07,
      "loss": 0.5182,
      "step": 3730
    },
    {
      "epoch": 4.69,
      "grad_norm": 1.1132534742355347,
      "learning_rate": 9.391204928367648e-08,
      "loss": 0.517,
      "step": 3740
    },
    {
      "epoch": 4.71,
      "grad_norm": 1.0367366075515747,
      "learning_rate": 8.642392718671455e-08,
      "loss": 0.5191,
      "step": 3750
    },
    {
      "epoch": 4.72,
      "grad_norm": 1.0188932418823242,
      "learning_rate": 7.92442721355996e-08,
      "loss": 0.5277,
      "step": 3760
    },
    {
      "epoch": 4.73,
      "grad_norm": 1.0666061639785767,
      "learning_rate": 7.237353485846865e-08,
      "loss": 0.5272,
      "step": 3770
    },
    {
      "epoch": 4.74,
      "grad_norm": 0.9185099601745605,
      "learning_rate": 6.58121466900552e-08,
      "loss": 0.5114,
      "step": 3780
    },
    {
      "epoch": 4.76,
      "grad_norm": 1.1004579067230225,
      "learning_rate": 5.9560519544614725e-08,
      "loss": 0.5088,
      "step": 3790
    },
    {
      "epoch": 4.77,
      "grad_norm": 1.0939706563949585,
      "learning_rate": 5.36190458900615e-08,
      "loss": 0.5371,
      "step": 3800
    },
    {
      "epoch": 4.77,
      "eval_loss": 0.5705362558364868,
      "eval_runtime": 72.0164,
      "eval_samples_per_second": 62.486,
      "eval_steps_per_second": 1.958,
      "step": 3800
    },
    {
      "epoch": 4.78,
      "grad_norm": 1.0084924697875977,
      "learning_rate": 4.798809872333276e-08,
      "loss": 0.5561,
      "step": 3810
    },
    {
      "epoch": 4.79,
      "grad_norm": 1.118712067604065,
      "learning_rate": 4.266803154697241e-08,
      "loss": 0.5511,
      "step": 3820
    },
    {
      "epoch": 4.81,
      "grad_norm": 1.0696793794631958,
      "learning_rate": 3.765917834693711e-08,
      "loss": 0.4753,
      "step": 3830
    },
    {
      "epoch": 4.82,
      "grad_norm": 0.9693195223808289,
      "learning_rate": 3.296185357162973e-08,
      "loss": 0.5308,
      "step": 3840
    },
    {
      "epoch": 4.83,
      "grad_norm": 1.145647406578064,
      "learning_rate": 2.8576352112158478e-08,
      "loss": 0.5174,
      "step": 3850
    },
    {
      "epoch": 4.84,
      "grad_norm": 0.9660029411315918,
      "learning_rate": 2.4502949283826683e-08,
      "loss": 0.4866,
      "step": 3860
    },
    {
      "epoch": 4.86,
      "grad_norm": 0.9987422823905945,
      "learning_rate": 2.074190080884386e-08,
      "loss": 0.5169,
      "step": 3870
    },
    {
      "epoch": 4.87,
      "grad_norm": 0.9800373911857605,
      "learning_rate": 1.729344280027745e-08,
      "loss": 0.5065,
      "step": 3880
    },
    {
      "epoch": 4.88,
      "grad_norm": 1.1359976530075073,
      "learning_rate": 1.4157791747225202e-08,
      "loss": 0.5496,
      "step": 3890
    },
    {
      "epoch": 4.89,
      "grad_norm": 0.9586722254753113,
      "learning_rate": 1.133514450122719e-08,
      "loss": 0.5138,
      "step": 3900
    },
    {
      "epoch": 4.91,
      "grad_norm": 1.1620326042175293,
      "learning_rate": 8.825678263906789e-09,
      "loss": 0.5573,
      "step": 3910
    },
    {
      "epoch": 4.92,
      "grad_norm": 0.8677152395248413,
      "learning_rate": 6.629550575847355e-09,
      "loss": 0.4891,
      "step": 3920
    },
    {
      "epoch": 4.93,
      "grad_norm": 1.0054067373275757,
      "learning_rate": 4.746899306699582e-09,
      "loss": 0.5325,
      "step": 3930
    },
    {
      "epoch": 4.94,
      "grad_norm": 1.2144650220870972,
      "learning_rate": 3.1778426465289838e-09,
      "loss": 0.4983,
      "step": 3940
    },
    {
      "epoch": 4.96,
      "grad_norm": 1.1614487171173096,
      "learning_rate": 1.9224790983940477e-09,
      "loss": 0.5108,
      "step": 3950
    },
    {
      "epoch": 4.97,
      "grad_norm": 0.9379814267158508,
      "learning_rate": 9.808874721628503e-10,
      "loss": 0.5122,
      "step": 3960
    },
    {
      "epoch": 4.98,
      "grad_norm": 1.0148200988769531,
      "learning_rate": 3.5312687956756684e-10,
      "loss": 0.5337,
      "step": 3970
    },
    {
      "epoch": 4.99,
      "grad_norm": 1.089341640472412,
      "learning_rate": 3.923673048966681e-11,
      "loss": 0.529,
      "step": 3980
    },
    {
      "epoch": 5.0,
      "step": 3985,
      "total_flos": 2.2824671149817856e+18,
      "train_loss": 0.569244503227052,
      "train_runtime": 8741.9073,
      "train_samples_per_second": 14.585,
      "train_steps_per_second": 0.456
    }
  ],
  "logging_steps": 10,
  "max_steps": 3985,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 1000,
  "total_flos": 2.2824671149817856e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}