{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.36,
  "eval_steps": 1000,
  "global_step": 18000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002,
      "grad_norm": 0.5271317362785339,
      "learning_rate": 0.00066,
      "loss": 7.779569091796875,
      "step": 100
    },
    {
      "epoch": 0.004,
      "grad_norm": 0.6387752890586853,
      "learning_rate": 0.0009999976160274871,
      "loss": 6.24430908203125,
      "step": 200
    },
    {
      "epoch": 0.006,
      "grad_norm": 0.42202863097190857,
      "learning_rate": 0.000999977956590203,
      "loss": 5.698394165039063,
      "step": 300
    },
    {
      "epoch": 0.008,
      "grad_norm": 0.42532625794410706,
      "learning_rate": 0.0009999384399149713,
      "loss": 5.405557861328125,
      "step": 400
    },
    {
      "epoch": 0.01,
      "grad_norm": 0.4566122889518738,
      "learning_rate": 0.0009998790675712448,
      "loss": 5.189588623046875,
      "step": 500
    },
    {
      "epoch": 0.012,
      "grad_norm": 0.4048032760620117,
      "learning_rate": 0.000999799841917069,
      "loss": 5.048131713867187,
      "step": 600
    },
    {
      "epoch": 0.014,
      "grad_norm": 0.33951064944267273,
      "learning_rate": 0.0009997007660989881,
      "loss": 4.923414611816407,
      "step": 700
    },
    {
      "epoch": 0.016,
      "grad_norm": 0.3192354440689087,
      "learning_rate": 0.0009995818440519195,
      "loss": 4.7589013671875,
      "step": 800
    },
    {
      "epoch": 0.018,
      "grad_norm": 0.4026377201080322,
      "learning_rate": 0.0009994430804989986,
      "loss": 4.643058166503907,
      "step": 900
    },
    {
      "epoch": 0.02,
      "grad_norm": 0.3453679084777832,
      "learning_rate": 0.0009992844809513896,
      "loss": 4.579913330078125,
      "step": 1000
    },
    {
      "epoch": 0.02,
      "eval_accuracy": 0.269652397386301,
      "eval_loss": 4.505138874053955,
      "eval_runtime": 6.375,
      "eval_samples_per_second": 304.469,
      "eval_steps_per_second": 19.137,
      "step": 1000
    },
    {
      "epoch": 0.022,
      "grad_norm": 0.3148439824581146,
      "learning_rate": 0.0009991060517080686,
      "loss": 4.5075772094726565,
      "step": 1100
    },
    {
      "epoch": 0.024,
      "grad_norm": 0.3168841600418091,
      "learning_rate": 0.0009989077998555718,
      "loss": 4.395697326660156,
      "step": 1200
    },
    {
      "epoch": 0.026,
      "grad_norm": 0.277462899684906,
      "learning_rate": 0.0009986897332677144,
      "loss": 4.3901171875,
      "step": 1300
    },
    {
      "epoch": 0.028,
      "grad_norm": 0.27992936968803406,
      "learning_rate": 0.0009984518606052791,
      "loss": 4.348124084472656,
      "step": 1400
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.2656252980232239,
      "learning_rate": 0.0009981941913156701,
      "loss": 4.3058538818359375,
      "step": 1500
    },
    {
      "epoch": 0.032,
      "grad_norm": 0.36868757009506226,
      "learning_rate": 0.0009979167356325394,
      "loss": 4.28273681640625,
      "step": 1600
    },
    {
      "epoch": 0.034,
      "grad_norm": 0.25333118438720703,
      "learning_rate": 0.0009976195045753797,
      "loss": 4.249351501464844,
      "step": 1700
    },
    {
      "epoch": 0.036,
      "grad_norm": 0.2491893470287323,
      "learning_rate": 0.0009973025099490876,
      "loss": 4.22731689453125,
      "step": 1800
    },
    {
      "epoch": 0.038,
      "grad_norm": 0.28424108028411865,
      "learning_rate": 0.0009969657643434933,
      "loss": 4.169223022460938,
      "step": 1900
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.2666165232658386,
      "learning_rate": 0.0009966092811328612,
      "loss": 4.170299682617188,
      "step": 2000
    },
    {
      "epoch": 0.04,
      "eval_accuracy": 0.3056205014664501,
      "eval_loss": 4.103638648986816,
      "eval_runtime": 6.566,
      "eval_samples_per_second": 295.612,
      "eval_steps_per_second": 18.58,
      "step": 2000
    },
    {
      "epoch": 0.042,
      "grad_norm": 0.3338668942451477,
      "learning_rate": 0.0009962330744753604,
      "loss": 4.163348388671875,
      "step": 2100
    },
    {
      "epoch": 0.044,
      "grad_norm": 0.25480538606643677,
      "learning_rate": 0.0009958371593124988,
      "loss": 4.118557434082032,
      "step": 2200
    },
    {
      "epoch": 0.046,
      "grad_norm": 0.2351720631122589,
      "learning_rate": 0.0009954215513685335,
      "loss": 4.1006707763671875,
      "step": 2300
    },
    {
      "epoch": 0.048,
      "grad_norm": 0.23063364624977112,
      "learning_rate": 0.0009949862671498437,
      "loss": 4.084933166503906,
      "step": 2400
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.270161896944046,
      "learning_rate": 0.000994531323944276,
      "loss": 4.043960876464844,
      "step": 2500
    },
    {
      "epoch": 0.052,
      "grad_norm": 0.23590487241744995,
      "learning_rate": 0.0009940567398204583,
      "loss": 4.043964233398437,
      "step": 2600
    },
    {
      "epoch": 0.054,
      "grad_norm": 0.2604995369911194,
      "learning_rate": 0.000993562533627081,
      "loss": 4.023601379394531,
      "step": 2700
    },
    {
      "epoch": 0.056,
      "grad_norm": 0.21301409602165222,
      "learning_rate": 0.0009930487249921495,
      "loss": 4.006760864257813,
      "step": 2800
    },
    {
      "epoch": 0.058,
      "grad_norm": 0.21575306355953217,
      "learning_rate": 0.0009925153343222043,
      "loss": 3.99057861328125,
      "step": 2900
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.2405770868062973,
      "learning_rate": 0.00099196238280151,
      "loss": 4.004396667480469,
      "step": 3000
    },
    {
      "epoch": 0.06,
      "eval_accuracy": 0.3194532243250246,
      "eval_loss": 3.945106267929077,
      "eval_runtime": 6.283,
      "eval_samples_per_second": 308.927,
      "eval_steps_per_second": 19.417,
      "step": 3000
    },
    {
      "epoch": 0.062,
      "grad_norm": 0.22380003333091736,
      "learning_rate": 0.0009913898923912148,
      "loss": 3.9952774047851562,
      "step": 3100
    },
    {
      "epoch": 0.064,
      "grad_norm": 0.21692751348018646,
      "learning_rate": 0.0009907978858284775,
      "loss": 3.9829229736328124,
      "step": 3200
    },
    {
      "epoch": 0.066,
      "grad_norm": 0.22457344830036163,
      "learning_rate": 0.0009901863866255652,
      "loss": 3.9637200927734373,
      "step": 3300
    },
    {
      "epoch": 0.068,
      "grad_norm": 0.20820939540863037,
      "learning_rate": 0.0009895554190689191,
      "loss": 3.930269470214844,
      "step": 3400
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.1961025595664978,
      "learning_rate": 0.0009889050082181896,
      "loss": 3.91483154296875,
      "step": 3500
    },
    {
      "epoch": 0.072,
      "grad_norm": 0.23347830772399902,
      "learning_rate": 0.0009882351799052416,
      "loss": 3.8973043823242186,
      "step": 3600
    },
    {
      "epoch": 0.074,
      "grad_norm": 0.21384640038013458,
      "learning_rate": 0.0009875459607331285,
      "loss": 3.9072760009765624,
      "step": 3700
    },
    {
      "epoch": 0.076,
      "grad_norm": 0.21380040049552917,
      "learning_rate": 0.000986837378075035,
      "loss": 3.8974929809570313,
      "step": 3800
    },
    {
      "epoch": 0.078,
      "grad_norm": 0.2199064940214157,
      "learning_rate": 0.0009861094600731911,
      "loss": 3.884527587890625,
      "step": 3900
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.2245541661977768,
      "learning_rate": 0.000985362235637753,
      "loss": 3.863822021484375,
      "step": 4000
    },
    {
      "epoch": 0.08,
      "eval_accuracy": 0.3280885939521158,
      "eval_loss": 3.8499362468719482,
      "eval_runtime": 6.3172,
      "eval_samples_per_second": 307.255,
      "eval_steps_per_second": 19.312,
      "step": 4000
    },
    {
      "epoch": 0.082,
      "grad_norm": 0.20376908779144287,
      "learning_rate": 0.000984595734445656,
      "loss": 3.8685012817382813,
      "step": 4100
    },
    {
      "epoch": 0.084,
      "grad_norm": 0.18422065675258636,
      "learning_rate": 0.0009838099869394353,
      "loss": 3.853851318359375,
      "step": 4200
    },
    {
      "epoch": 0.086,
      "grad_norm": 0.19115814566612244,
      "learning_rate": 0.0009830050243260172,
      "loss": 3.8678164672851563,
      "step": 4300
    },
    {
      "epoch": 0.088,
      "grad_norm": 0.20425236225128174,
      "learning_rate": 0.0009821808785754791,
      "loss": 3.8287237548828124,
      "step": 4400
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.1944635510444641,
      "learning_rate": 0.0009813375824197811,
      "loss": 3.8451242065429687,
      "step": 4500
    },
    {
      "epoch": 0.092,
      "grad_norm": 0.17947296798229218,
      "learning_rate": 0.0009804751693514645,
      "loss": 3.845400695800781,
      "step": 4600
    },
    {
      "epoch": 0.094,
      "grad_norm": 0.20076608657836914,
      "learning_rate": 0.0009795936736223222,
      "loss": 3.8257626342773436,
      "step": 4700
    },
    {
      "epoch": 0.096,
      "grad_norm": 0.18554596602916718,
      "learning_rate": 0.0009786931302420386,
      "loss": 3.808670959472656,
      "step": 4800
    },
    {
      "epoch": 0.098,
      "grad_norm": 0.23913995921611786,
      "learning_rate": 0.000977773574976799,
      "loss": 3.810527038574219,
      "step": 4900
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.18414103984832764,
      "learning_rate": 0.0009768350443478686,
      "loss": 3.7990948486328127,
      "step": 5000
    },
    {
      "epoch": 0.1,
      "eval_accuracy": 0.33500293894949945,
      "eval_loss": 3.7702300548553467,
      "eval_runtime": 6.4137,
      "eval_samples_per_second": 302.634,
      "eval_steps_per_second": 19.022,
      "step": 5000
    },
    {
      "epoch": 0.102,
      "grad_norm": 0.20944379270076752,
      "learning_rate": 0.0009758775756301432,
      "loss": 3.798116455078125,
      "step": 5100
    },
    {
      "epoch": 0.104,
      "grad_norm": 0.18056762218475342,
      "learning_rate": 0.0009749012068506673,
      "loss": 3.7997445678710937,
      "step": 5200
    },
    {
      "epoch": 0.106,
      "grad_norm": 0.20922255516052246,
      "learning_rate": 0.000973905976787125,
      "loss": 3.7765731811523438,
      "step": 5300
    },
    {
      "epoch": 0.108,
      "grad_norm": 0.19328580796718597,
      "learning_rate": 0.0009728919249662993,
      "loss": 3.729337463378906,
      "step": 5400
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.2031419277191162,
      "learning_rate": 0.0009718590916625022,
      "loss": 3.744057922363281,
      "step": 5500
    },
    {
      "epoch": 0.112,
      "grad_norm": 0.19743800163269043,
      "learning_rate": 0.0009708075178959756,
      "loss": 3.7054071044921875,
      "step": 5600
    },
    {
      "epoch": 0.114,
      "grad_norm": 0.1741236299276352,
      "learning_rate": 0.0009697372454312618,
      "loss": 3.699843444824219,
      "step": 5700
    },
    {
      "epoch": 0.116,
      "grad_norm": 0.1799750030040741,
      "learning_rate": 0.0009686483167755448,
      "loss": 3.7227572631835937,
      "step": 5800
    },
    {
      "epoch": 0.118,
      "grad_norm": 0.2036087065935135,
      "learning_rate": 0.000967540775176962,
      "loss": 3.6742709350585936,
      "step": 5900
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.21463191509246826,
      "learning_rate": 0.0009664146646228871,
      "loss": 3.6995980834960935,
      "step": 6000
    },
    {
      "epoch": 0.12,
      "eval_accuracy": 0.34141821705074654,
      "eval_loss": 3.711214303970337,
      "eval_runtime": 6.3349,
      "eval_samples_per_second": 306.4,
      "eval_steps_per_second": 19.259,
      "step": 6000
    },
    {
      "epoch": 0.122,
      "grad_norm": 0.18666808307170868,
      "learning_rate": 0.000965270029838182,
      "loss": 3.6821221923828125,
      "step": 6100
    },
    {
      "epoch": 0.124,
      "grad_norm": 0.15751489996910095,
      "learning_rate": 0.0009641069162834215,
      "loss": 3.671753234863281,
      "step": 6200
    },
    {
      "epoch": 0.126,
      "grad_norm": 0.22766178846359253,
      "learning_rate": 0.0009629253701530875,
      "loss": 3.6878753662109376,
      "step": 6300
    },
    {
      "epoch": 0.128,
      "grad_norm": 0.1677054762840271,
      "learning_rate": 0.0009617254383737343,
      "loss": 3.6619857788085937,
      "step": 6400
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.17945168912410736,
      "learning_rate": 0.0009605071686021245,
      "loss": 3.6484750366210936,
      "step": 6500
    },
    {
      "epoch": 0.132,
      "grad_norm": 0.17799174785614014,
      "learning_rate": 0.0009592706092233365,
      "loss": 3.6635867309570314,
      "step": 6600
    },
    {
      "epoch": 0.134,
      "grad_norm": 0.5586189031600952,
      "learning_rate": 0.0009580158093488436,
      "loss": 3.6489688110351564,
      "step": 6700
    },
    {
      "epoch": 0.136,
      "grad_norm": 0.2055712342262268,
      "learning_rate": 0.000956742818814562,
      "loss": 3.6523666381835938,
      "step": 6800
    },
    {
      "epoch": 0.138,
      "grad_norm": 0.1812128871679306,
      "learning_rate": 0.0009554516881788724,
      "loss": 3.6479287719726563,
      "step": 6900
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.1664743423461914,
      "learning_rate": 0.0009541424687206124,
      "loss": 3.6319317626953125,
      "step": 7000
    },
    {
      "epoch": 0.14,
      "eval_accuracy": 0.3448148965923309,
      "eval_loss": 3.6666343212127686,
      "eval_runtime": 6.38,
      "eval_samples_per_second": 304.231,
      "eval_steps_per_second": 19.122,
      "step": 7000
    },
    {
      "epoch": 0.142,
      "grad_norm": 0.15217284858226776,
      "learning_rate": 0.0009528152124370387,
      "loss": 3.6308517456054688,
      "step": 7100
    },
    {
      "epoch": 0.144,
      "grad_norm": 0.16507816314697266,
      "learning_rate": 0.0009514699720417631,
      "loss": 3.6217156982421876,
      "step": 7200
    },
    {
      "epoch": 0.146,
      "grad_norm": 0.2040642648935318,
      "learning_rate": 0.0009501068009626583,
      "loss": 3.632703857421875,
      "step": 7300
    },
    {
      "epoch": 0.148,
      "grad_norm": 0.18437133729457855,
      "learning_rate": 0.0009487257533397361,
      "loss": 3.6346307373046876,
      "step": 7400
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.18760915100574493,
      "learning_rate": 0.0009473268840229971,
      "loss": 3.5973056030273436,
      "step": 7500
    },
    {
      "epoch": 0.152,
      "grad_norm": 0.17388536036014557,
      "learning_rate": 0.0009459102485702528,
      "loss": 3.611794738769531,
      "step": 7600
    },
    {
      "epoch": 0.154,
      "grad_norm": 0.18797264993190765,
      "learning_rate": 0.0009444759032449177,
      "loss": 3.609403381347656,
      "step": 7700
    },
    {
      "epoch": 0.156,
      "grad_norm": 0.18140241503715515,
      "learning_rate": 0.0009430239050137766,
      "loss": 3.6027340698242187,
      "step": 7800
    },
    {
      "epoch": 0.158,
      "grad_norm": 0.17433640360832214,
      "learning_rate": 0.0009415543115447203,
      "loss": 3.6214892578125,
      "step": 7900
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.16605013608932495,
      "learning_rate": 0.0009400671812044565,
      "loss": 3.6061587524414063,
      "step": 8000
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.34762277801806923,
      "eval_loss": 3.640489339828491,
      "eval_runtime": 8.0453,
      "eval_samples_per_second": 241.259,
      "eval_steps_per_second": 15.164,
      "step": 8000
    },
    {
      "epoch": 0.162,
      "grad_norm": 0.16548939049243927,
      "learning_rate": 0.0009385625730561914,
      "loss": 3.588617248535156,
      "step": 8100
    },
    {
      "epoch": 0.164,
      "grad_norm": 0.16542939841747284,
      "learning_rate": 0.0009370405468572831,
      "loss": 3.590567626953125,
      "step": 8200
    },
    {
      "epoch": 0.166,
      "grad_norm": 0.18206581473350525,
      "learning_rate": 0.0009355011630568697,
      "loss": 3.636882629394531,
      "step": 8300
    },
    {
      "epoch": 0.168,
      "grad_norm": 0.2654706835746765,
      "learning_rate": 0.0009339444827934676,
      "loss": 3.6797540283203123,
      "step": 8400
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.16825009882450104,
      "learning_rate": 0.0009323705678925427,
      "loss": 3.686376953125,
      "step": 8500
    },
    {
      "epoch": 0.172,
      "grad_norm": 0.18937945365905762,
      "learning_rate": 0.000930779480864057,
      "loss": 3.6942425537109376,
      "step": 8600
    },
    {
      "epoch": 0.174,
      "grad_norm": 0.19819994270801544,
      "learning_rate": 0.0009291712848999833,
      "loss": 3.6582196044921873,
      "step": 8700
    },
    {
      "epoch": 0.176,
      "grad_norm": 0.1557057648897171,
      "learning_rate": 0.0009275460438717975,
      "loss": 3.6659271240234377,
      "step": 8800
    },
    {
      "epoch": 0.178,
      "grad_norm": 0.16359280049800873,
      "learning_rate": 0.0009259038223279413,
      "loss": 3.66092529296875,
      "step": 8900
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.17897452414035797,
      "learning_rate": 0.0009242446854912575,
      "loss": 3.6642529296875,
      "step": 9000
    },
    {
      "epoch": 0.18,
      "eval_accuracy": 0.3497692697794326,
      "eval_loss": 3.617546319961548,
      "eval_runtime": 6.279,
      "eval_samples_per_second": 309.124,
      "eval_steps_per_second": 19.43,
      "step": 9000
    },
    {
      "epoch": 0.182,
      "grad_norm": 0.20515520870685577,
      "learning_rate": 0.0009225686992564015,
      "loss": 3.6445053100585936,
      "step": 9100
    },
    {
      "epoch": 0.184,
      "grad_norm": 0.16788582503795624,
      "learning_rate": 0.0009208759301872229,
      "loss": 3.657896728515625,
      "step": 9200
    },
    {
      "epoch": 0.186,
      "grad_norm": 0.1988508403301239,
      "learning_rate": 0.0009191664455141217,
      "loss": 3.616723937988281,
      "step": 9300
    },
    {
      "epoch": 0.188,
      "grad_norm": 0.1822570413351059,
      "learning_rate": 0.0009174403131313792,
      "loss": 3.6103390502929686,
      "step": 9400
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.15679553151130676,
      "learning_rate": 0.0009156976015944607,
      "loss": 3.624302978515625,
      "step": 9500
    },
    {
      "epoch": 0.192,
      "grad_norm": 0.15683937072753906,
      "learning_rate": 0.0009139383801172929,
      "loss": 3.6558917236328123,
      "step": 9600
    },
    {
      "epoch": 0.194,
      "grad_norm": 0.1553466022014618,
      "learning_rate": 0.0009121627185695152,
      "loss": 3.6198284912109373,
      "step": 9700
    },
    {
      "epoch": 0.196,
      "grad_norm": 0.14881721138954163,
      "learning_rate": 0.0009103706874737043,
      "loss": 3.60408203125,
      "step": 9800
    },
    {
      "epoch": 0.198,
      "grad_norm": 0.1628589779138565,
      "learning_rate": 0.0009085623580025737,
      "loss": 3.6047940063476562,
      "step": 9900
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.15589991211891174,
      "learning_rate": 0.0009067378019761472,
      "loss": 3.599223327636719,
      "step": 10000
    },
    {
      "epoch": 0.2,
      "eval_accuracy": 0.3534593401629882,
      "eval_loss": 3.579674005508423,
      "eval_runtime": 6.3176,
      "eval_samples_per_second": 307.239,
      "eval_steps_per_second": 19.311,
      "step": 10000
    },
    {
      "epoch": 0.202,
      "grad_norm": 0.18158847093582153,
      "learning_rate": 0.0009048970918589055,
      "loss": 3.607147521972656,
      "step": 10100
    },
    {
      "epoch": 0.204,
      "grad_norm": 0.15705344080924988,
      "learning_rate": 0.0009030403007569094,
      "loss": 3.5972610473632813,
      "step": 10200
    },
    {
      "epoch": 0.206,
      "grad_norm": 0.18561947345733643,
      "learning_rate": 0.0009011675024148955,
      "loss": 3.5994192504882814,
      "step": 10300
    },
    {
      "epoch": 0.208,
      "grad_norm": 0.17375363409519196,
      "learning_rate": 0.0008992787712133476,
      "loss": 3.592384948730469,
      "step": 10400
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.14704184234142303,
      "learning_rate": 0.0008973741821655428,
      "loss": 3.5763116455078126,
      "step": 10500
    },
    {
      "epoch": 0.212,
      "grad_norm": 0.1663106381893158,
      "learning_rate": 0.0008954538109145713,
      "loss": 3.59366455078125,
      "step": 10600
    },
    {
      "epoch": 0.214,
      "grad_norm": 0.15191592276096344,
      "learning_rate": 0.0008935177337303337,
      "loss": 3.5883197021484374,
      "step": 10700
    },
    {
      "epoch": 0.216,
      "grad_norm": 0.15566915273666382,
      "learning_rate": 0.0008915660275065107,
      "loss": 3.5549887084960936,
      "step": 10800
    },
    {
      "epoch": 0.218,
      "grad_norm": 0.17490389943122864,
      "learning_rate": 0.0008895987697575093,
      "loss": 3.558960266113281,
      "step": 10900
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.17277678847312927,
      "learning_rate": 0.0008876160386153846,
      "loss": 3.5671484375,
      "step": 11000
    },
    {
      "epoch": 0.22,
      "eval_accuracy": 0.3558709927196726,
      "eval_loss": 3.554290771484375,
      "eval_runtime": 6.8937,
      "eval_samples_per_second": 281.559,
      "eval_steps_per_second": 17.697,
      "step": 11000
    },
    {
      "epoch": 0.222,
      "grad_norm": 0.15808428823947906,
      "learning_rate": 0.0008856179128267363,
      "loss": 3.578575439453125,
      "step": 11100
    },
    {
      "epoch": 0.224,
      "grad_norm": 0.14474111795425415,
      "learning_rate": 0.0008836044717495819,
      "loss": 3.5772747802734375,
      "step": 11200
    },
    {
      "epoch": 0.226,
      "grad_norm": 0.1413886845111847,
      "learning_rate": 0.0008815757953502035,
      "loss": 3.556199645996094,
      "step": 11300
    },
    {
      "epoch": 0.228,
      "grad_norm": 0.1502675712108612,
      "learning_rate": 0.0008795319641999735,
      "loss": 3.570118103027344,
      "step": 11400
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.1517445147037506,
      "learning_rate": 0.0008774730594721534,
      "loss": 3.5563198852539064,
      "step": 11500
    },
    {
      "epoch": 0.232,
      "grad_norm": 0.16406166553497314,
      "learning_rate": 0.0008753991629386706,
      "loss": 3.5616604614257814,
      "step": 11600
    },
    {
      "epoch": 0.234,
      "grad_norm": 0.14764074981212616,
      "learning_rate": 0.0008733103569668701,
      "loss": 3.5495458984375,
      "step": 11700
    },
    {
      "epoch": 0.236,
      "grad_norm": 0.17058512568473816,
      "learning_rate": 0.0008712067245162439,
      "loss": 3.5244882202148435,
      "step": 11800
    },
    {
      "epoch": 0.238,
      "grad_norm": 0.16127121448516846,
      "learning_rate": 0.0008690883491351356,
      "loss": 3.5396826171875,
      "step": 11900
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.14211590588092804,
      "learning_rate": 0.0008669553149574224,
      "loss": 3.522979431152344,
      "step": 12000
    },
    {
      "epoch": 0.24,
      "eval_accuracy": 0.3578128166428224,
      "eval_loss": 3.5307533740997314,
      "eval_runtime": 6.6725,
      "eval_samples_per_second": 290.897,
      "eval_steps_per_second": 18.284,
      "step": 12000
    },
    {
      "epoch": 0.242,
      "grad_norm": 0.1405150443315506,
      "learning_rate": 0.0008648077066991736,
      "loss": 3.527177734375,
      "step": 12100
    },
    {
      "epoch": 0.244,
      "grad_norm": 0.16479545831680298,
      "learning_rate": 0.0008626456096552858,
      "loss": 3.5393338012695312,
      "step": 12200
    },
    {
      "epoch": 0.246,
      "grad_norm": 0.1583888679742813,
      "learning_rate": 0.0008604691096960961,
      "loss": 3.5796047973632814,
      "step": 12300
    },
    {
      "epoch": 0.248,
      "grad_norm": 0.17304493486881256,
      "learning_rate": 0.0008582782932639702,
      "loss": 3.591195068359375,
      "step": 12400
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.15857121348381042,
      "learning_rate": 0.0008560732473698707,
      "loss": 3.584400634765625,
      "step": 12500
    },
    {
      "epoch": 0.252,
      "grad_norm": 0.14658208191394806,
      "learning_rate": 0.0008538540595899008,
      "loss": 3.589564208984375,
      "step": 12600
    },
    {
      "epoch": 0.254,
      "grad_norm": 0.143401101231575,
      "learning_rate": 0.0008516208180618252,
      "loss": 3.5824600219726563,
      "step": 12700
    },
    {
      "epoch": 0.256,
      "grad_norm": 0.16244202852249146,
      "learning_rate": 0.0008493736114815718,
      "loss": 3.574875793457031,
      "step": 12800
    },
    {
      "epoch": 0.258,
      "grad_norm": 0.1520756632089615,
      "learning_rate": 0.0008471125290997063,
      "loss": 3.568282165527344,
      "step": 12900
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.1582706868648529,
      "learning_rate": 0.00084483766071789,
      "loss": 3.5930340576171873,
      "step": 13000
    },
    {
      "epoch": 0.26,
      "eval_accuracy": 0.35929186944409996,
      "eval_loss": 3.5190227031707764,
      "eval_runtime": 6.3791,
      "eval_samples_per_second": 304.275,
      "eval_steps_per_second": 19.125,
      "step": 13000
    },
    {
      "epoch": 0.262,
      "grad_norm": 0.17478469014167786,
      "learning_rate": 0.0008425490966853113,
      "loss": 3.5665899658203126,
      "step": 13100
    },
    {
      "epoch": 0.264,
      "grad_norm": 0.15228988230228424,
      "learning_rate": 0.0008402469278950989,
      "loss": 3.5444100952148436,
      "step": 13200
    },
    {
      "epoch": 0.266,
      "grad_norm": 0.15244163572788239,
      "learning_rate": 0.0008379312457807102,
      "loss": 3.5695867919921875,
      "step": 13300
    },
    {
      "epoch": 0.268,
      "grad_norm": 0.15935847163200378,
      "learning_rate": 0.0008356021423123017,
      "loss": 3.5604925537109375,
      "step": 13400
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.15347428619861603,
      "learning_rate": 0.000833259709993075,
      "loss": 3.5551889038085935,
      "step": 13500
    },
    {
      "epoch": 0.272,
      "grad_norm": 0.21459928154945374,
      "learning_rate": 0.0008309040418556038,
      "loss": 3.550736389160156,
      "step": 13600
    },
    {
      "epoch": 0.274,
      "grad_norm": 0.17246893048286438,
      "learning_rate": 0.0008285352314581378,
      "loss": 3.5650180053710936,
      "step": 13700
    },
    {
      "epoch": 0.276,
      "grad_norm": 0.16454122960567474,
      "learning_rate": 0.0008261533728808883,
      "loss": 3.549154968261719,
      "step": 13800
    },
    {
      "epoch": 0.278,
      "grad_norm": 0.14885519444942474,
      "learning_rate": 0.000823758560722291,
      "loss": 3.5430950927734375,
      "step": 13900
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.1521039605140686,
      "learning_rate": 0.0008213508900952489,
      "loss": 3.547527160644531,
      "step": 14000
    },
    {
      "epoch": 0.28,
      "eval_accuracy": 0.36159261824608735,
      "eval_loss": 3.4961202144622803,
      "eval_runtime": 6.3742,
      "eval_samples_per_second": 304.507,
      "eval_steps_per_second": 19.14,
      "step": 14000
    },
    {
      "epoch": 0.282,
      "grad_norm": 0.1499979943037033,
      "learning_rate": 0.0008189304566233547,
      "loss": 3.5474667358398437,
      "step": 14100
    },
    {
      "epoch": 0.284,
      "grad_norm": 0.1552744209766388,
      "learning_rate": 0.0008164973564370934,
      "loss": 3.5074334716796876,
      "step": 14200
    },
    {
      "epoch": 0.286,
      "grad_norm": 0.13462179899215698,
      "learning_rate": 0.0008140516861700238,
      "loss": 3.5272781372070314,
      "step": 14300
    },
    {
      "epoch": 0.288,
      "grad_norm": 0.1381780356168747,
      "learning_rate": 0.000811593542954941,
      "loss": 3.5094314575195313,
      "step": 14400
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.13239678740501404,
      "learning_rate": 0.000809123024420019,
      "loss": 3.5180905151367186,
      "step": 14500
    },
    {
      "epoch": 0.292,
      "grad_norm": 0.15930698812007904,
      "learning_rate": 0.0008066402286849317,
      "loss": 3.5262945556640624,
      "step": 14600
    },
    {
      "epoch": 0.294,
      "grad_norm": 0.1524200737476349,
      "learning_rate": 0.0008041452543569583,
      "loss": 3.541943359375,
      "step": 14700
    },
    {
      "epoch": 0.296,
      "grad_norm": 0.14040064811706543,
      "learning_rate": 0.0008016382005270651,
      "loss": 3.543092041015625,
      "step": 14800
    },
    {
      "epoch": 0.298,
      "grad_norm": 0.14799822866916656,
      "learning_rate": 0.0007991191667659708,
      "loss": 3.5315762329101563,
      "step": 14900
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.1449238359928131,
      "learning_rate": 0.0007965882531201913,
      "loss": 3.522008361816406,
      "step": 15000
    },
    {
      "epoch": 0.3,
      "eval_accuracy": 0.36332372503531274,
      "eval_loss": 3.4778010845184326,
      "eval_runtime": 6.6567,
      "eval_samples_per_second": 291.588,
      "eval_steps_per_second": 18.328,
      "step": 15000
    },
    {
      "epoch": 0.302,
      "grad_norm": 0.13593044877052307,
      "learning_rate": 0.0007940455601080678,
      "loss": 3.5114361572265627,
      "step": 15100
    },
    {
      "epoch": 0.304,
      "grad_norm": 0.1423630267381668,
      "learning_rate": 0.0007914911887157724,
      "loss": 3.5139959716796874,
      "step": 15200
    },
    {
      "epoch": 0.306,
      "grad_norm": 0.15380671620368958,
      "learning_rate": 0.0007889252403932986,
      "loss": 3.515833435058594,
      "step": 15300
    },
    {
      "epoch": 0.308,
      "grad_norm": 0.13759265840053558,
      "learning_rate": 0.0007863478170504326,
      "loss": 3.528351135253906,
      "step": 15400
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.13887037336826324,
      "learning_rate": 0.0007837590210527042,
      "loss": 3.5276870727539062,
      "step": 15500
    },
    {
      "epoch": 0.312,
      "grad_norm": 0.15678516030311584,
      "learning_rate": 0.0007811589552173225,
      "loss": 3.515467224121094,
      "step": 15600
    },
    {
      "epoch": 0.314,
      "grad_norm": 0.2133788764476776,
      "learning_rate": 0.0007785477228090924,
      "loss": 3.5256777954101564,
      "step": 15700
    },
    {
      "epoch": 0.316,
      "grad_norm": 0.17379774153232574,
      "learning_rate": 0.0007759254275363124,
      "loss": 3.5354928588867187,
      "step": 15800
    },
    {
      "epoch": 0.318,
      "grad_norm": 0.15530315041542053,
      "learning_rate": 0.0007732921735466564,
      "loss": 3.5182159423828123,
      "step": 15900
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.14711181819438934,
      "learning_rate": 0.0007706480654230373,
      "loss": 3.5050119018554686,
      "step": 16000
    },
    {
      "epoch": 0.32,
      "eval_accuracy": 0.3644458693896563,
      "eval_loss": 3.4653918743133545,
      "eval_runtime": 6.2866,
      "eval_samples_per_second": 308.75,
      "eval_steps_per_second": 19.406,
      "step": 16000
    },
    {
      "epoch": 0.322,
      "grad_norm": 0.16945767402648926,
      "learning_rate": 0.0007679932081794532,
      "loss": 3.5018475341796873,
      "step": 16100
    },
    {
      "epoch": 0.324,
      "grad_norm": 0.14158344268798828,
      "learning_rate": 0.0007653277072568166,
      "loss": 3.52821044921875,
      "step": 16200
    },
    {
      "epoch": 0.326,
      "grad_norm": 0.15132898092269897,
      "learning_rate": 0.0007626516685187669,
      "loss": 3.5037826538085937,
      "step": 16300
    },
    {
      "epoch": 0.328,
      "grad_norm": 0.1444397270679474,
      "learning_rate": 0.0007599651982474656,
      "loss": 3.5029248046875,
      "step": 16400
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.15066049993038177,
      "learning_rate": 0.0007572684031393755,
      "loss": 3.5109619140625,
      "step": 16500
    },
    {
      "epoch": 0.332,
      "grad_norm": 0.15420712530612946,
      "learning_rate": 0.0007545613903010231,
      "loss": 3.498372802734375,
      "step": 16600
    },
    {
      "epoch": 0.334,
      "grad_norm": 0.15669448673725128,
      "learning_rate": 0.000751844267244744,
      "loss": 3.504073181152344,
      "step": 16700
    },
    {
      "epoch": 0.336,
      "grad_norm": 0.14457182586193085,
      "learning_rate": 0.000749117141884414,
      "loss": 3.5009725952148436,
      "step": 16800
    },
    {
      "epoch": 0.338,
      "grad_norm": 0.1625896543264389,
      "learning_rate": 0.0007463801225311631,
      "loss": 3.5116473388671876,
      "step": 16900
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.1407516449689865,
      "learning_rate": 0.0007436333178890721,
      "loss": 3.4775790405273437,
      "step": 17000
    },
    {
      "epoch": 0.34,
      "eval_accuracy": 0.3666105090381519,
      "eval_loss": 3.4458343982696533,
      "eval_runtime": 6.6119,
      "eval_samples_per_second": 293.56,
      "eval_steps_per_second": 18.451,
      "step": 17000
    },
    {
      "epoch": 0.342,
      "grad_norm": 0.15488570928573608,
      "learning_rate": 0.0007408768370508576,
      "loss": 3.5057693481445313,
      "step": 17100
    },
    {
      "epoch": 0.344,
      "grad_norm": 0.13969258964061737,
      "learning_rate": 0.0007381107894935382,
      "loss": 3.4655438232421876,
      "step": 17200
    },
    {
      "epoch": 0.346,
      "grad_norm": 0.15081284940242767,
      "learning_rate": 0.0007353352850740858,
      "loss": 3.490411376953125,
      "step": 17300
    },
    {
      "epoch": 0.348,
      "grad_norm": 0.1477256566286087,
      "learning_rate": 0.0007325504340250635,
      "loss": 3.48463134765625,
      "step": 17400
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.15759490430355072,
      "learning_rate": 0.0007297563469502479,
      "loss": 3.5031024169921876,
      "step": 17500
    },
    {
      "epoch": 0.352,
      "grad_norm": 0.14924432337284088,
      "learning_rate": 0.0007269531348202344,
      "loss": 3.5010330200195314,
      "step": 17600
    },
    {
      "epoch": 0.354,
      "grad_norm": 0.21179653704166412,
      "learning_rate": 0.0007241409089680324,
      "loss": 3.484757385253906,
      "step": 17700
    },
    {
      "epoch": 0.356,
      "grad_norm": 0.1406019777059555,
      "learning_rate": 0.0007213197810846416,
      "loss": 3.4941912841796876,
      "step": 17800
    },
    {
      "epoch": 0.358,
      "grad_norm": 0.18723711371421814,
      "learning_rate": 0.0007184898632146171,
      "loss": 3.4815142822265623,
      "step": 17900
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.15316542983055115,
      "learning_rate": 0.0007156512677516189,
      "loss": 3.4828509521484374,
      "step": 18000
    },
    {
      "epoch": 0.36,
      "eval_accuracy": 0.36743422147076527,
      "eval_loss": 3.434631824493408,
      "eval_runtime": 6.7452,
      "eval_samples_per_second": 287.759,
      "eval_steps_per_second": 18.087,
      "step": 18000
    }
  ],
  "logging_steps": 100,
  "max_steps": 50000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 2000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7.0559295602688e+17,
  "train_batch_size": 120,
  "trial_name": null,
  "trial_params": null
}