{
  "best_metric": 3.3074893951416016,
  "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_8397/checkpoint-90000",
  "epoch": 10.0,
  "eval_steps": 1000,
  "global_step": 92910,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.005381552039608223,
      "grad_norm": 1.0923771858215332,
      "learning_rate": 0.000276,
      "loss": 8.905,
      "step": 50
    },
    {
      "epoch": 0.010763104079216447,
      "grad_norm": 5.010653972625732,
      "learning_rate": 0.0005759999999999999,
      "loss": 6.9301,
      "step": 100
    },
    {
      "epoch": 0.01614465611882467,
      "grad_norm": 3.7069694995880127,
      "learning_rate": 0.0005997026182523434,
      "loss": 6.4972,
      "step": 150
    },
    {
      "epoch": 0.021526208158432893,
      "grad_norm": 1.9403975009918213,
      "learning_rate": 0.000599379377222282,
      "loss": 6.2279,
      "step": 200
    },
    {
      "epoch": 0.026907760198041114,
      "grad_norm": 0.8380448818206787,
      "learning_rate": 0.0005990561361922206,
      "loss": 6.0994,
      "step": 250
    },
    {
      "epoch": 0.03228931223764934,
      "grad_norm": 0.9163997769355774,
      "learning_rate": 0.0005987328951621592,
      "loss": 6.0177,
      "step": 300
    },
    {
      "epoch": 0.03767086427725756,
      "grad_norm": 1.83314049243927,
      "learning_rate": 0.0005984096541320978,
      "loss": 5.8949,
      "step": 350
    },
    {
      "epoch": 0.04305241631686579,
      "grad_norm": 1.4096308946609497,
      "learning_rate": 0.0005980864131020364,
      "loss": 5.8067,
      "step": 400
    },
    {
      "epoch": 0.048433968356474004,
      "grad_norm": 1.6748238801956177,
      "learning_rate": 0.0005977631720719749,
      "loss": 5.7313,
      "step": 450
    },
    {
      "epoch": 0.05381552039608223,
      "grad_norm": 1.4383798837661743,
      "learning_rate": 0.0005974399310419136,
      "loss": 5.654,
      "step": 500
    },
    {
      "epoch": 0.05919707243569045,
      "grad_norm": 1.6079118251800537,
      "learning_rate": 0.0005971166900118521,
      "loss": 5.6001,
      "step": 550
    },
    {
      "epoch": 0.06457862447529868,
      "grad_norm": 1.9025360345840454,
      "learning_rate": 0.0005967934489817907,
      "loss": 5.5244,
      "step": 600
    },
    {
      "epoch": 0.0699601765149069,
      "grad_norm": 1.3215116262435913,
      "learning_rate": 0.0005964702079517293,
      "loss": 5.4419,
      "step": 650
    },
    {
      "epoch": 0.07534172855451512,
      "grad_norm": 1.4640989303588867,
      "learning_rate": 0.0005961469669216679,
      "loss": 5.3806,
      "step": 700
    },
    {
      "epoch": 0.08072328059412334,
      "grad_norm": 1.0821962356567383,
      "learning_rate": 0.0005958237258916065,
      "loss": 5.339,
      "step": 750
    },
    {
      "epoch": 0.08610483263373157,
      "grad_norm": 0.9703434705734253,
      "learning_rate": 0.000595500484861545,
      "loss": 5.2583,
      "step": 800
    },
    {
      "epoch": 0.09148638467333979,
      "grad_norm": 1.449822187423706,
      "learning_rate": 0.0005951772438314836,
      "loss": 5.2537,
      "step": 850
    },
    {
      "epoch": 0.09686793671294801,
      "grad_norm": 0.968447208404541,
      "learning_rate": 0.0005948540028014222,
      "loss": 5.1838,
      "step": 900
    },
    {
      "epoch": 0.10224948875255624,
      "grad_norm": 1.200282096862793,
      "learning_rate": 0.0005945307617713608,
      "loss": 5.1603,
      "step": 950
    },
    {
      "epoch": 0.10763104079216446,
      "grad_norm": 0.853746235370636,
      "learning_rate": 0.0005942075207412994,
      "loss": 5.0938,
      "step": 1000
    },
    {
      "epoch": 0.10763104079216446,
      "eval_accuracy": 0.22625899089065055,
      "eval_loss": 5.027560710906982,
      "eval_runtime": 185.6101,
      "eval_samples_per_second": 97.037,
      "eval_steps_per_second": 6.066,
      "step": 1000
    },
    {
      "epoch": 0.11301259283177269,
      "grad_norm": 1.0276198387145996,
      "learning_rate": 0.0005938842797112379,
      "loss": 5.0735,
      "step": 1050
    },
    {
      "epoch": 0.1183941448713809,
      "grad_norm": 1.47368586063385,
      "learning_rate": 0.0005935610386811766,
      "loss": 5.0273,
      "step": 1100
    },
    {
      "epoch": 0.12377569691098914,
      "grad_norm": 1.2419272661209106,
      "learning_rate": 0.0005932377976511151,
      "loss": 4.9866,
      "step": 1150
    },
    {
      "epoch": 0.12915724895059735,
      "grad_norm": 1.1619503498077393,
      "learning_rate": 0.0005929145566210538,
      "loss": 4.952,
      "step": 1200
    },
    {
      "epoch": 0.13453880099020557,
      "grad_norm": 1.1115535497665405,
      "learning_rate": 0.0005925913155909923,
      "loss": 4.9194,
      "step": 1250
    },
    {
      "epoch": 0.1399203530298138,
      "grad_norm": 1.3335832357406616,
      "learning_rate": 0.0005922680745609308,
      "loss": 4.8757,
      "step": 1300
    },
    {
      "epoch": 0.14530190506942203,
      "grad_norm": 1.134364366531372,
      "learning_rate": 0.0005919448335308695,
      "loss": 4.8749,
      "step": 1350
    },
    {
      "epoch": 0.15068345710903025,
      "grad_norm": 1.3366910219192505,
      "learning_rate": 0.000591621592500808,
      "loss": 4.8819,
      "step": 1400
    },
    {
      "epoch": 0.15606500914863847,
      "grad_norm": 1.0884933471679688,
      "learning_rate": 0.0005912983514707467,
      "loss": 4.8362,
      "step": 1450
    },
    {
      "epoch": 0.16144656118824668,
      "grad_norm": 1.1018588542938232,
      "learning_rate": 0.0005909751104406852,
      "loss": 4.8289,
      "step": 1500
    },
    {
      "epoch": 0.1668281132278549,
      "grad_norm": 0.9973709583282471,
      "learning_rate": 0.0005906518694106238,
      "loss": 4.7902,
      "step": 1550
    },
    {
      "epoch": 0.17220966526746315,
      "grad_norm": 1.046265959739685,
      "learning_rate": 0.0005903286283805624,
      "loss": 4.7524,
      "step": 1600
    },
    {
      "epoch": 0.17759121730707136,
      "grad_norm": 0.9931600689888,
      "learning_rate": 0.000590005387350501,
      "loss": 4.71,
      "step": 1650
    },
    {
      "epoch": 0.18297276934667958,
      "grad_norm": 1.0001553297042847,
      "learning_rate": 0.0005896821463204395,
      "loss": 4.7059,
      "step": 1700
    },
    {
      "epoch": 0.1883543213862878,
      "grad_norm": 0.9167148470878601,
      "learning_rate": 0.0005893589052903781,
      "loss": 4.7046,
      "step": 1750
    },
    {
      "epoch": 0.19373587342589602,
      "grad_norm": 0.9704087972640991,
      "learning_rate": 0.0005890356642603167,
      "loss": 4.6749,
      "step": 1800
    },
    {
      "epoch": 0.19911742546550426,
      "grad_norm": 1.140143632888794,
      "learning_rate": 0.0005887124232302553,
      "loss": 4.6602,
      "step": 1850
    },
    {
      "epoch": 0.20449897750511248,
      "grad_norm": 0.8475252389907837,
      "learning_rate": 0.0005883891822001939,
      "loss": 4.6277,
      "step": 1900
    },
    {
      "epoch": 0.2098805295447207,
      "grad_norm": 0.8329743146896362,
      "learning_rate": 0.0005880659411701324,
      "loss": 4.6096,
      "step": 1950
    },
    {
      "epoch": 0.2152620815843289,
      "grad_norm": 0.9684931635856628,
      "learning_rate": 0.0005877427001400711,
      "loss": 4.5878,
      "step": 2000
    },
    {
      "epoch": 0.2152620815843289,
      "eval_accuracy": 0.2699155039087333,
      "eval_loss": 4.517221450805664,
      "eval_runtime": 184.3843,
      "eval_samples_per_second": 97.682,
      "eval_steps_per_second": 6.107,
      "step": 2000
    },
    {
      "epoch": 0.22064363362393713,
      "grad_norm": 1.2581902742385864,
      "learning_rate": 0.0005874194591100097,
      "loss": 4.5739,
      "step": 2050
    },
    {
      "epoch": 0.22602518566354537,
      "grad_norm": 0.7552979588508606,
      "learning_rate": 0.0005870962180799483,
      "loss": 4.5653,
      "step": 2100
    },
    {
      "epoch": 0.2314067377031536,
      "grad_norm": 1.023279070854187,
      "learning_rate": 0.0005867729770498868,
      "loss": 4.5489,
      "step": 2150
    },
    {
      "epoch": 0.2367882897427618,
      "grad_norm": 0.9059367179870605,
      "learning_rate": 0.0005864497360198253,
      "loss": 4.517,
      "step": 2200
    },
    {
      "epoch": 0.24216984178237003,
      "grad_norm": 1.0197012424468994,
      "learning_rate": 0.000586126494989764,
      "loss": 4.5106,
      "step": 2250
    },
    {
      "epoch": 0.24755139382197827,
      "grad_norm": 0.8007754683494568,
      "learning_rate": 0.0005858032539597025,
      "loss": 4.4997,
      "step": 2300
    },
    {
      "epoch": 0.2529329458615865,
      "grad_norm": 1.3180443048477173,
      "learning_rate": 0.0005854800129296412,
      "loss": 4.4656,
      "step": 2350
    },
    {
      "epoch": 0.2583144979011947,
      "grad_norm": 0.8712721467018127,
      "learning_rate": 0.0005851567718995797,
      "loss": 4.4512,
      "step": 2400
    },
    {
      "epoch": 0.2636960499408029,
      "grad_norm": 0.7161034345626831,
      "learning_rate": 0.0005848335308695184,
      "loss": 4.4308,
      "step": 2450
    },
    {
      "epoch": 0.26907760198041114,
      "grad_norm": 1.078532099723816,
      "learning_rate": 0.0005845102898394569,
      "loss": 4.4226,
      "step": 2500
    },
    {
      "epoch": 0.27445915402001936,
      "grad_norm": 0.917757511138916,
      "learning_rate": 0.0005841870488093954,
      "loss": 4.4269,
      "step": 2550
    },
    {
      "epoch": 0.2798407060596276,
      "grad_norm": 0.7378865480422974,
      "learning_rate": 0.0005838638077793341,
      "loss": 4.4106,
      "step": 2600
    },
    {
      "epoch": 0.2852222580992358,
      "grad_norm": 0.6821770071983337,
      "learning_rate": 0.0005835405667492726,
      "loss": 4.4227,
      "step": 2650
    },
    {
      "epoch": 0.29060381013884407,
      "grad_norm": 0.6910375952720642,
      "learning_rate": 0.0005832173257192113,
      "loss": 4.3887,
      "step": 2700
    },
    {
      "epoch": 0.2959853621784523,
      "grad_norm": 0.7855560183525085,
      "learning_rate": 0.0005828940846891498,
      "loss": 4.3653,
      "step": 2750
    },
    {
      "epoch": 0.3013669142180605,
      "grad_norm": 0.7701571583747864,
      "learning_rate": 0.0005825708436590884,
      "loss": 4.3616,
      "step": 2800
    },
    {
      "epoch": 0.3067484662576687,
      "grad_norm": 0.775539755821228,
      "learning_rate": 0.000582247602629027,
      "loss": 4.3327,
      "step": 2850
    },
    {
      "epoch": 0.31213001829727693,
      "grad_norm": 0.8155066967010498,
      "learning_rate": 0.0005819243615989656,
      "loss": 4.3324,
      "step": 2900
    },
    {
      "epoch": 0.31751157033688515,
      "grad_norm": 0.9817063212394714,
      "learning_rate": 0.0005816011205689042,
      "loss": 4.3248,
      "step": 2950
    },
    {
      "epoch": 0.32289312237649337,
      "grad_norm": 0.7922750115394592,
      "learning_rate": 0.0005812778795388427,
      "loss": 4.3159,
      "step": 3000
    },
    {
      "epoch": 0.32289312237649337,
      "eval_accuracy": 0.2971647789495026,
      "eval_loss": 4.248455047607422,
      "eval_runtime": 184.2879,
      "eval_samples_per_second": 97.733,
      "eval_steps_per_second": 6.11,
      "step": 3000
    },
    {
      "epoch": 0.3282746744161016,
      "grad_norm": 0.8070213198661804,
      "learning_rate": 0.0005809546385087813,
      "loss": 4.3187,
      "step": 3050
    },
    {
      "epoch": 0.3336562264557098,
      "grad_norm": 0.8484339118003845,
      "learning_rate": 0.0005806313974787199,
      "loss": 4.2951,
      "step": 3100
    },
    {
      "epoch": 0.3390377784953181,
      "grad_norm": 0.8407900333404541,
      "learning_rate": 0.0005803081564486585,
      "loss": 4.2776,
      "step": 3150
    },
    {
      "epoch": 0.3444193305349263,
      "grad_norm": 0.6488985419273376,
      "learning_rate": 0.0005799849154185971,
      "loss": 4.2513,
      "step": 3200
    },
    {
      "epoch": 0.3498008825745345,
      "grad_norm": 0.7860519289970398,
      "learning_rate": 0.0005796616743885357,
      "loss": 4.2804,
      "step": 3250
    },
    {
      "epoch": 0.35518243461414273,
      "grad_norm": 0.8071337938308716,
      "learning_rate": 0.0005793384333584742,
      "loss": 4.2652,
      "step": 3300
    },
    {
      "epoch": 0.36056398665375095,
      "grad_norm": 0.9089975953102112,
      "learning_rate": 0.0005790151923284129,
      "loss": 4.2523,
      "step": 3350
    },
    {
      "epoch": 0.36594553869335916,
      "grad_norm": 0.6199256181716919,
      "learning_rate": 0.0005786919512983514,
      "loss": 4.2457,
      "step": 3400
    },
    {
      "epoch": 0.3713270907329674,
      "grad_norm": 0.865071177482605,
      "learning_rate": 0.00057836871026829,
      "loss": 4.2515,
      "step": 3450
    },
    {
      "epoch": 0.3767086427725756,
      "grad_norm": 1.0260653495788574,
      "learning_rate": 0.0005780454692382286,
      "loss": 4.2156,
      "step": 3500
    },
    {
      "epoch": 0.3820901948121838,
      "grad_norm": 0.7370710372924805,
      "learning_rate": 0.0005777222282081672,
      "loss": 4.224,
      "step": 3550
    },
    {
      "epoch": 0.38747174685179203,
      "grad_norm": 0.7235073447227478,
      "learning_rate": 0.0005773989871781058,
      "loss": 4.2312,
      "step": 3600
    },
    {
      "epoch": 0.3928532988914003,
      "grad_norm": 0.7964087724685669,
      "learning_rate": 0.0005770757461480443,
      "loss": 4.2306,
      "step": 3650
    },
    {
      "epoch": 0.3982348509310085,
      "grad_norm": 0.7269555330276489,
      "learning_rate": 0.0005767525051179829,
      "loss": 4.1817,
      "step": 3700
    },
    {
      "epoch": 0.40361640297061674,
      "grad_norm": 0.8491329550743103,
      "learning_rate": 0.0005764292640879215,
      "loss": 4.2062,
      "step": 3750
    },
    {
      "epoch": 0.40899795501022496,
      "grad_norm": 0.7043264508247375,
      "learning_rate": 0.0005761060230578601,
      "loss": 4.1914,
      "step": 3800
    },
    {
      "epoch": 0.4143795070498332,
      "grad_norm": 0.6347329616546631,
      "learning_rate": 0.0005757827820277987,
      "loss": 4.1934,
      "step": 3850
    },
    {
      "epoch": 0.4197610590894414,
      "grad_norm": 0.6858643293380737,
      "learning_rate": 0.0005754595409977372,
      "loss": 4.1993,
      "step": 3900
    },
    {
      "epoch": 0.4251426111290496,
      "grad_norm": 0.7558616995811462,
      "learning_rate": 0.0005751362999676759,
      "loss": 4.1659,
      "step": 3950
    },
    {
      "epoch": 0.4305241631686578,
      "grad_norm": 0.6548260450363159,
      "learning_rate": 0.0005748130589376144,
      "loss": 4.1508,
      "step": 4000
    },
    {
      "epoch": 0.4305241631686578,
      "eval_accuracy": 0.31135604127983973,
      "eval_loss": 4.0979228019714355,
      "eval_runtime": 184.2288,
      "eval_samples_per_second": 97.764,
      "eval_steps_per_second": 6.112,
      "step": 4000
    },
    {
      "epoch": 0.43590571520826604,
      "grad_norm": 0.8841599225997925,
      "learning_rate": 0.0005744898179075531,
      "loss": 4.1773,
      "step": 4050
    },
    {
      "epoch": 0.44128726724787426,
      "grad_norm": 0.6886371970176697,
      "learning_rate": 0.0005741665768774916,
      "loss": 4.1583,
      "step": 4100
    },
    {
      "epoch": 0.44666881928748253,
      "grad_norm": 0.7568600177764893,
      "learning_rate": 0.0005738433358474301,
      "loss": 4.1538,
      "step": 4150
    },
    {
      "epoch": 0.45205037132709075,
      "grad_norm": 0.654498815536499,
      "learning_rate": 0.0005735200948173688,
      "loss": 4.1466,
      "step": 4200
    },
    {
      "epoch": 0.45743192336669897,
      "grad_norm": 0.64454185962677,
      "learning_rate": 0.0005731968537873073,
      "loss": 4.1449,
      "step": 4250
    },
    {
      "epoch": 0.4628134754063072,
      "grad_norm": 0.6529496312141418,
      "learning_rate": 0.0005728736127572459,
      "loss": 4.1347,
      "step": 4300
    },
    {
      "epoch": 0.4681950274459154,
      "grad_norm": 0.6523901224136353,
      "learning_rate": 0.0005725503717271845,
      "loss": 4.1423,
      "step": 4350
    },
    {
      "epoch": 0.4735765794855236,
      "grad_norm": 0.6438787579536438,
      "learning_rate": 0.0005722271306971231,
      "loss": 4.1317,
      "step": 4400
    },
    {
      "epoch": 0.47895813152513184,
      "grad_norm": 0.6913219690322876,
      "learning_rate": 0.0005719038896670617,
      "loss": 4.1008,
      "step": 4450
    },
    {
      "epoch": 0.48433968356474005,
      "grad_norm": 0.5881590843200684,
      "learning_rate": 0.0005715806486370003,
      "loss": 4.1066,
      "step": 4500
    },
    {
      "epoch": 0.48972123560434827,
      "grad_norm": 0.7720493078231812,
      "learning_rate": 0.0005712574076069388,
      "loss": 4.1066,
      "step": 4550
    },
    {
      "epoch": 0.49510278764395654,
      "grad_norm": 0.6393665671348572,
      "learning_rate": 0.0005709341665768774,
      "loss": 4.1018,
      "step": 4600
    },
    {
      "epoch": 0.5004843396835648,
      "grad_norm": 0.6110801100730896,
      "learning_rate": 0.0005706109255468161,
      "loss": 4.0953,
      "step": 4650
    },
    {
      "epoch": 0.505865891723173,
      "grad_norm": 0.676651120185852,
      "learning_rate": 0.0005702876845167546,
      "loss": 4.1018,
      "step": 4700
    },
    {
      "epoch": 0.5112474437627812,
      "grad_norm": 0.5868623852729797,
      "learning_rate": 0.0005699644434866932,
      "loss": 4.0973,
      "step": 4750
    },
    {
      "epoch": 0.5166289958023894,
      "grad_norm": 0.7219673991203308,
      "learning_rate": 0.0005696412024566317,
      "loss": 4.0898,
      "step": 4800
    },
    {
      "epoch": 0.5220105478419976,
      "grad_norm": 0.5670786499977112,
      "learning_rate": 0.0005693179614265704,
      "loss": 4.0783,
      "step": 4850
    },
    {
      "epoch": 0.5273920998816058,
      "grad_norm": 0.6825320720672607,
      "learning_rate": 0.000568994720396509,
      "loss": 4.0794,
      "step": 4900
    },
    {
      "epoch": 0.5327736519212141,
      "grad_norm": 0.5827348828315735,
      "learning_rate": 0.0005686714793664476,
      "loss": 4.0609,
      "step": 4950
    },
    {
      "epoch": 0.5381552039608223,
      "grad_norm": 0.6638315320014954,
      "learning_rate": 0.0005683482383363861,
      "loss": 4.0837,
      "step": 5000
    },
    {
      "epoch": 0.5381552039608223,
      "eval_accuracy": 0.32000948322404843,
      "eval_loss": 3.999183177947998,
      "eval_runtime": 184.338,
      "eval_samples_per_second": 97.706,
      "eval_steps_per_second": 6.108,
      "step": 5000
    },
    {
      "epoch": 0.5435367560004305,
      "grad_norm": 0.7330002784729004,
      "learning_rate": 0.0005680249973063247,
      "loss": 4.0738,
      "step": 5050
    },
    {
      "epoch": 0.5489183080400387,
      "grad_norm": 0.6824637055397034,
      "learning_rate": 0.0005677017562762633,
      "loss": 4.0366,
      "step": 5100
    },
    {
      "epoch": 0.5542998600796469,
      "grad_norm": 0.5739036798477173,
      "learning_rate": 0.0005673785152462018,
      "loss": 4.0466,
      "step": 5150
    },
    {
      "epoch": 0.5596814121192552,
      "grad_norm": 0.6992231011390686,
      "learning_rate": 0.0005670552742161405,
      "loss": 4.0659,
      "step": 5200
    },
    {
      "epoch": 0.5650629641588634,
      "grad_norm": 0.5668877363204956,
      "learning_rate": 0.000566732033186079,
      "loss": 4.0459,
      "step": 5250
    },
    {
      "epoch": 0.5704445161984716,
      "grad_norm": 0.6363801956176758,
      "learning_rate": 0.0005664087921560177,
      "loss": 4.0489,
      "step": 5300
    },
    {
      "epoch": 0.5758260682380799,
      "grad_norm": 0.5494595170021057,
      "learning_rate": 0.0005660855511259562,
      "loss": 4.042,
      "step": 5350
    },
    {
      "epoch": 0.5812076202776881,
      "grad_norm": 0.6047319173812866,
      "learning_rate": 0.0005657623100958947,
      "loss": 4.0271,
      "step": 5400
    },
    {
      "epoch": 0.5865891723172963,
      "grad_norm": 0.6150631308555603,
      "learning_rate": 0.0005654390690658334,
      "loss": 4.0221,
      "step": 5450
    },
    {
      "epoch": 0.5919707243569046,
      "grad_norm": 0.688296377658844,
      "learning_rate": 0.0005651158280357719,
      "loss": 4.0468,
      "step": 5500
    },
    {
      "epoch": 0.5973522763965128,
      "grad_norm": 0.6616494059562683,
      "learning_rate": 0.0005647925870057106,
      "loss": 4.018,
      "step": 5550
    },
    {
      "epoch": 0.602733828436121,
      "grad_norm": 0.707185685634613,
      "learning_rate": 0.0005644693459756491,
      "loss": 4.024,
      "step": 5600
    },
    {
      "epoch": 0.6081153804757292,
      "grad_norm": 0.6183993816375732,
      "learning_rate": 0.0005641461049455877,
      "loss": 4.0125,
      "step": 5650
    },
    {
      "epoch": 0.6134969325153374,
      "grad_norm": 0.643642008304596,
      "learning_rate": 0.0005638228639155263,
      "loss": 3.9979,
      "step": 5700
    },
    {
      "epoch": 0.6188784845549457,
      "grad_norm": 0.5784419775009155,
      "learning_rate": 0.000563499622885465,
      "loss": 4.0152,
      "step": 5750
    },
    {
      "epoch": 0.6242600365945539,
      "grad_norm": 0.576373815536499,
      "learning_rate": 0.0005631763818554035,
      "loss": 3.9974,
      "step": 5800
    },
    {
      "epoch": 0.6296415886341621,
      "grad_norm": 0.6600840091705322,
      "learning_rate": 0.000562853140825342,
      "loss": 3.9927,
      "step": 5850
    },
    {
      "epoch": 0.6350231406737703,
      "grad_norm": 0.5999715328216553,
      "learning_rate": 0.0005625298997952806,
      "loss": 4.0025,
      "step": 5900
    },
    {
      "epoch": 0.6404046927133785,
      "grad_norm": 0.7674629092216492,
      "learning_rate": 0.0005622066587652192,
      "loss": 3.9989,
      "step": 5950
    },
    {
      "epoch": 0.6457862447529867,
      "grad_norm": 0.587006688117981,
      "learning_rate": 0.0005618834177351578,
      "loss": 3.9879,
      "step": 6000
    },
    {
      "epoch": 0.6457862447529867,
      "eval_accuracy": 0.3270059686290863,
      "eval_loss": 3.927816152572632,
      "eval_runtime": 184.1499,
      "eval_samples_per_second": 97.806,
      "eval_steps_per_second": 6.115,
      "step": 6000
    },
    {
      "epoch": 0.651167796792595,
      "grad_norm": 0.5483689904212952,
      "learning_rate": 0.0005615601767050964,
      "loss": 3.9963,
      "step": 6050
    },
    {
      "epoch": 0.6565493488322032,
      "grad_norm": 0.6261682510375977,
      "learning_rate": 0.000561236935675035,
      "loss": 3.9999,
      "step": 6100
    },
    {
      "epoch": 0.6619309008718114,
      "grad_norm": 0.678266167640686,
      "learning_rate": 0.0005609136946449735,
      "loss": 4.0082,
      "step": 6150
    },
    {
      "epoch": 0.6673124529114196,
      "grad_norm": 0.6246299743652344,
      "learning_rate": 0.0005605904536149122,
      "loss": 3.979,
      "step": 6200
    },
    {
      "epoch": 0.6726940049510278,
      "grad_norm": 0.6304299235343933,
      "learning_rate": 0.0005602672125848507,
      "loss": 3.9852,
      "step": 6250
    },
    {
      "epoch": 0.6780755569906362,
      "grad_norm": 0.5472633838653564,
      "learning_rate": 0.0005599439715547892,
      "loss": 3.9831,
      "step": 6300
    },
    {
      "epoch": 0.6834571090302444,
      "grad_norm": 0.6037302613258362,
      "learning_rate": 0.0005596207305247279,
      "loss": 3.9646,
      "step": 6350
    },
    {
      "epoch": 0.6888386610698526,
      "grad_norm": 0.6150529384613037,
      "learning_rate": 0.0005592974894946665,
      "loss": 3.995,
      "step": 6400
    },
    {
      "epoch": 0.6942202131094608,
      "grad_norm": 0.6179707646369934,
      "learning_rate": 0.0005589742484646051,
      "loss": 3.9371,
      "step": 6450
    },
    {
      "epoch": 0.699601765149069,
      "grad_norm": 0.6564136743545532,
      "learning_rate": 0.0005586510074345436,
      "loss": 3.9702,
      "step": 6500
    },
    {
      "epoch": 0.7049833171886772,
      "grad_norm": 0.5408684015274048,
      "learning_rate": 0.0005583277664044822,
      "loss": 3.9481,
      "step": 6550
    },
    {
      "epoch": 0.7103648692282855,
      "grad_norm": 0.5830994248390198,
      "learning_rate": 0.0005580045253744208,
      "loss": 3.9577,
      "step": 6600
    },
    {
      "epoch": 0.7157464212678937,
      "grad_norm": 0.5456255078315735,
      "learning_rate": 0.0005576812843443594,
      "loss": 3.9669,
      "step": 6650
    },
    {
      "epoch": 0.7211279733075019,
      "grad_norm": 0.6891123652458191,
      "learning_rate": 0.000557358043314298,
      "loss": 3.9589,
      "step": 6700
    },
    {
      "epoch": 0.7265095253471101,
      "grad_norm": 0.5847958922386169,
      "learning_rate": 0.0005570348022842365,
      "loss": 3.9368,
      "step": 6750
    },
    {
      "epoch": 0.7318910773867183,
      "grad_norm": 0.5667703151702881,
      "learning_rate": 0.0005567115612541752,
      "loss": 3.9593,
      "step": 6800
    },
    {
      "epoch": 0.7372726294263265,
      "grad_norm": 0.541790783405304,
      "learning_rate": 0.0005563883202241137,
      "loss": 3.9328,
      "step": 6850
    },
    {
      "epoch": 0.7426541814659348,
      "grad_norm": 0.5813462734222412,
      "learning_rate": 0.0005560650791940524,
      "loss": 3.9225,
      "step": 6900
    },
    {
      "epoch": 0.748035733505543,
      "grad_norm": 0.6490563154220581,
      "learning_rate": 0.0005557418381639909,
      "loss": 3.9485,
      "step": 6950
    },
    {
      "epoch": 0.7534172855451512,
      "grad_norm": 0.5413534045219421,
      "learning_rate": 0.0005554185971339294,
      "loss": 3.9299,
      "step": 7000
    },
    {
      "epoch": 0.7534172855451512,
      "eval_accuracy": 0.33278380324439694,
      "eval_loss": 3.8707778453826904,
      "eval_runtime": 184.3815,
      "eval_samples_per_second": 97.683,
      "eval_steps_per_second": 6.107,
      "step": 7000
    },
    {
      "epoch": 0.7587988375847594,
      "grad_norm": 0.6043950319290161,
      "learning_rate": 0.0005550953561038681,
      "loss": 3.9224,
      "step": 7050
    },
    {
      "epoch": 0.7641803896243676,
      "grad_norm": 0.5365579128265381,
      "learning_rate": 0.0005547721150738066,
      "loss": 3.9412,
      "step": 7100
    },
    {
      "epoch": 0.7695619416639758,
      "grad_norm": 0.5504319667816162,
      "learning_rate": 0.0005544488740437452,
      "loss": 3.9377,
      "step": 7150
    },
    {
      "epoch": 0.7749434937035841,
      "grad_norm": 0.6742015480995178,
      "learning_rate": 0.0005541256330136838,
      "loss": 3.9204,
      "step": 7200
    },
    {
      "epoch": 0.7803250457431924,
      "grad_norm": 0.5779771208763123,
      "learning_rate": 0.0005538023919836224,
      "loss": 3.9493,
      "step": 7250
    },
    {
      "epoch": 0.7857065977828006,
      "grad_norm": 0.5889744758605957,
      "learning_rate": 0.000553479150953561,
      "loss": 3.9188,
      "step": 7300
    },
    {
      "epoch": 0.7910881498224088,
      "grad_norm": 0.5811043381690979,
      "learning_rate": 0.0005531559099234996,
      "loss": 3.9044,
      "step": 7350
    },
    {
      "epoch": 0.796469701862017,
      "grad_norm": 0.5543528199195862,
      "learning_rate": 0.0005528326688934381,
      "loss": 3.9059,
      "step": 7400
    },
    {
      "epoch": 0.8018512539016253,
      "grad_norm": 0.5994428396224976,
      "learning_rate": 0.0005525094278633767,
      "loss": 3.9127,
      "step": 7450
    },
    {
      "epoch": 0.8072328059412335,
      "grad_norm": 0.5440976619720459,
      "learning_rate": 0.0005521861868333154,
      "loss": 3.9111,
      "step": 7500
    },
    {
      "epoch": 0.8126143579808417,
      "grad_norm": 0.6140932440757751,
      "learning_rate": 0.0005518629458032539,
      "loss": 3.8938,
      "step": 7550
    },
    {
      "epoch": 0.8179959100204499,
      "grad_norm": 0.5452330112457275,
      "learning_rate": 0.0005515397047731925,
      "loss": 3.91,
      "step": 7600
    },
    {
      "epoch": 0.8233774620600581,
      "grad_norm": 0.606785237789154,
      "learning_rate": 0.000551216463743131,
      "loss": 3.9099,
      "step": 7650
    },
    {
      "epoch": 0.8287590140996663,
      "grad_norm": 0.7822673320770264,
      "learning_rate": 0.0005508932227130697,
      "loss": 3.9161,
      "step": 7700
    },
    {
      "epoch": 0.8341405661392746,
      "grad_norm": 0.6136397123336792,
      "learning_rate": 0.0005505699816830083,
      "loss": 3.9033,
      "step": 7750
    },
    {
      "epoch": 0.8395221181788828,
      "grad_norm": 0.5778501033782959,
      "learning_rate": 0.0005502467406529469,
      "loss": 3.8885,
      "step": 7800
    },
    {
      "epoch": 0.844903670218491,
      "grad_norm": 0.593637228012085,
      "learning_rate": 0.0005499234996228854,
      "loss": 3.9036,
      "step": 7850
    },
    {
      "epoch": 0.8502852222580992,
      "grad_norm": 0.5076649188995361,
      "learning_rate": 0.000549600258592824,
      "loss": 3.879,
      "step": 7900
    },
    {
      "epoch": 0.8556667742977074,
      "grad_norm": 0.5722057223320007,
      "learning_rate": 0.0005492770175627626,
      "loss": 3.8818,
      "step": 7950
    },
    {
      "epoch": 0.8610483263373157,
      "grad_norm": 0.6967954039573669,
      "learning_rate": 0.0005489537765327011,
      "loss": 3.8842,
      "step": 8000
    },
    {
      "epoch": 0.8610483263373157,
      "eval_accuracy": 0.3366016484164222,
      "eval_loss": 3.825267791748047,
      "eval_runtime": 184.5054,
      "eval_samples_per_second": 97.618,
      "eval_steps_per_second": 6.103,
      "step": 8000
    },
    {
      "epoch": 0.8664298783769239,
      "grad_norm": 0.5275264382362366,
      "learning_rate": 0.0005486305355026398,
      "loss": 3.9018,
      "step": 8050
    },
    {
      "epoch": 0.8718114304165321,
      "grad_norm": 0.5515099167823792,
      "learning_rate": 0.0005483072944725783,
      "loss": 3.8622,
      "step": 8100
    },
    {
      "epoch": 0.8771929824561403,
      "grad_norm": 0.5505463480949402,
      "learning_rate": 0.000547984053442517,
      "loss": 3.878,
      "step": 8150
    },
    {
      "epoch": 0.8825745344957485,
      "grad_norm": 0.5193983316421509,
      "learning_rate": 0.0005476608124124555,
      "loss": 3.8674,
      "step": 8200
    },
    {
      "epoch": 0.8879560865353568,
      "grad_norm": 0.620212197303772,
      "learning_rate": 0.000547337571382394,
      "loss": 3.8707,
      "step": 8250
    },
    {
      "epoch": 0.8933376385749651,
      "grad_norm": 0.5699217915534973,
      "learning_rate": 0.0005470143303523327,
      "loss": 3.8838,
      "step": 8300
    },
    {
      "epoch": 0.8987191906145733,
      "grad_norm": 0.6195424199104309,
      "learning_rate": 0.0005466910893222712,
      "loss": 3.8663,
      "step": 8350
    },
    {
      "epoch": 0.9041007426541815,
      "grad_norm": 0.6764254570007324,
      "learning_rate": 0.0005463678482922099,
      "loss": 3.8717,
      "step": 8400
    },
    {
      "epoch": 0.9094822946937897,
      "grad_norm": 0.6726669669151306,
      "learning_rate": 0.0005460446072621484,
      "loss": 3.8749,
      "step": 8450
    },
    {
      "epoch": 0.9148638467333979,
      "grad_norm": 0.6046283841133118,
      "learning_rate": 0.000545721366232087,
      "loss": 3.8916,
      "step": 8500
    },
    {
      "epoch": 0.9202453987730062,
      "grad_norm": 0.6035982966423035,
      "learning_rate": 0.0005453981252020256,
      "loss": 3.8527,
      "step": 8550
    },
    {
      "epoch": 0.9256269508126144,
      "grad_norm": 0.5406537055969238,
      "learning_rate": 0.0005450748841719643,
      "loss": 3.8683,
      "step": 8600
    },
    {
      "epoch": 0.9310085028522226,
      "grad_norm": 0.6445631980895996,
      "learning_rate": 0.0005447516431419028,
      "loss": 3.8542,
      "step": 8650
    },
    {
      "epoch": 0.9363900548918308,
      "grad_norm": 0.602548360824585,
      "learning_rate": 0.0005444284021118413,
      "loss": 3.8721,
      "step": 8700
    },
    {
      "epoch": 0.941771606931439,
      "grad_norm": 0.610095739364624,
      "learning_rate": 0.0005441051610817799,
      "loss": 3.8594,
      "step": 8750
    },
    {
      "epoch": 0.9471531589710472,
      "grad_norm": 0.6943879127502441,
      "learning_rate": 0.0005437819200517185,
      "loss": 3.8639,
      "step": 8800
    },
    {
      "epoch": 0.9525347110106555,
      "grad_norm": 0.6467054486274719,
      "learning_rate": 0.0005434586790216571,
      "loss": 3.8672,
      "step": 8850
    },
    {
      "epoch": 0.9579162630502637,
      "grad_norm": 0.5887782573699951,
      "learning_rate": 0.0005431354379915957,
      "loss": 3.8524,
      "step": 8900
    },
    {
      "epoch": 0.9632978150898719,
      "grad_norm": 0.5128800868988037,
      "learning_rate": 0.0005428121969615343,
      "loss": 3.8485,
      "step": 8950
    },
    {
      "epoch": 0.9686793671294801,
      "grad_norm": 0.5380699634552002,
      "learning_rate": 0.0005424889559314729,
      "loss": 3.8386,
      "step": 9000
    },
    {
      "epoch": 0.9686793671294801,
      "eval_accuracy": 0.33981255855711573,
      "eval_loss": 3.788262128829956,
      "eval_runtime": 184.0526,
      "eval_samples_per_second": 97.858,
      "eval_steps_per_second": 6.118,
      "step": 9000
    },
    {
      "epoch": 0.9740609191690883,
      "grad_norm": 0.6042304635047913,
      "learning_rate": 0.0005421657149014114,
      "loss": 3.8379,
      "step": 9050
    },
    {
      "epoch": 0.9794424712086965,
      "grad_norm": 0.617558479309082,
      "learning_rate": 0.00054184247387135,
      "loss": 3.8482,
      "step": 9100
    },
    {
      "epoch": 0.9848240232483048,
      "grad_norm": 0.5932952761650085,
      "learning_rate": 0.0005415192328412885,
      "loss": 3.8522,
      "step": 9150
    },
    {
      "epoch": 0.9902055752879131,
      "grad_norm": 0.6659948229789734,
      "learning_rate": 0.0005411959918112272,
      "loss": 3.8387,
      "step": 9200
    },
    {
      "epoch": 0.9955871273275213,
      "grad_norm": 0.5533868074417114,
      "learning_rate": 0.0005408727507811658,
      "loss": 3.8382,
      "step": 9250
    },
    {
      "epoch": 1.0009686793671295,
      "grad_norm": 0.5350527167320251,
      "learning_rate": 0.0005405495097511044,
      "loss": 3.8408,
      "step": 9300
    },
    {
      "epoch": 1.0063502314067376,
      "grad_norm": 0.5588667988777161,
      "learning_rate": 0.0005402262687210429,
      "loss": 3.7594,
      "step": 9350
    },
    {
      "epoch": 1.011731783446346,
      "grad_norm": 0.5812219381332397,
      "learning_rate": 0.0005399030276909816,
      "loss": 3.7775,
      "step": 9400
    },
    {
      "epoch": 1.017113335485954,
      "grad_norm": 0.5945906043052673,
      "learning_rate": 0.0005395797866609201,
      "loss": 3.7699,
      "step": 9450
    },
    {
      "epoch": 1.0224948875255624,
      "grad_norm": 0.574937641620636,
      "learning_rate": 0.0005392565456308587,
      "loss": 3.7631,
      "step": 9500
    },
    {
      "epoch": 1.0278764395651705,
      "grad_norm": 0.6346364617347717,
      "learning_rate": 0.0005389333046007973,
      "loss": 3.7746,
      "step": 9550
    },
    {
      "epoch": 1.0332579916047788,
      "grad_norm": 0.5427976250648499,
      "learning_rate": 0.0005386100635707358,
      "loss": 3.7821,
      "step": 9600
    },
    {
      "epoch": 1.0386395436443872,
      "grad_norm": 0.5465062856674194,
      "learning_rate": 0.0005382868225406745,
      "loss": 3.7803,
      "step": 9650
    },
    {
      "epoch": 1.0440210956839953,
      "grad_norm": 0.5530688762664795,
      "learning_rate": 0.000537963581510613,
      "loss": 3.7915,
      "step": 9700
    },
    {
      "epoch": 1.0494026477236036,
      "grad_norm": 0.5546349287033081,
      "learning_rate": 0.0005376403404805517,
      "loss": 3.7977,
      "step": 9750
    },
    {
      "epoch": 1.0547841997632117,
      "grad_norm": 0.5490013360977173,
      "learning_rate": 0.0005373170994504902,
      "loss": 3.76,
      "step": 9800
    },
    {
      "epoch": 1.06016575180282,
      "grad_norm": 0.5702437162399292,
      "learning_rate": 0.0005369938584204287,
      "loss": 3.7829,
      "step": 9850
    },
    {
      "epoch": 1.0655473038424281,
      "grad_norm": 0.6038593053817749,
      "learning_rate": 0.0005366706173903674,
      "loss": 3.7752,
      "step": 9900
    },
    {
      "epoch": 1.0709288558820365,
      "grad_norm": 0.6309857368469238,
      "learning_rate": 0.0005363473763603059,
      "loss": 3.7587,
      "step": 9950
    },
    {
      "epoch": 1.0763104079216446,
      "grad_norm": 0.5484387874603271,
      "learning_rate": 0.0005360241353302445,
      "loss": 3.7785,
      "step": 10000
    },
    {
      "epoch": 1.0763104079216446,
      "eval_accuracy": 0.34385216427317034,
      "eval_loss": 3.7527427673339844,
      "eval_runtime": 184.4909,
      "eval_samples_per_second": 97.625,
      "eval_steps_per_second": 6.103,
      "step": 10000
    },
    {
      "epoch": 1.081691959961253,
      "grad_norm": 0.5536912083625793,
      "learning_rate": 0.0005357008943001831,
      "loss": 3.794,
      "step": 10050
    },
    {
      "epoch": 1.087073512000861,
      "grad_norm": 0.59341961145401,
      "learning_rate": 0.0005353776532701217,
      "loss": 3.7755,
      "step": 10100
    },
    {
      "epoch": 1.0924550640404693,
      "grad_norm": 0.5561048984527588,
      "learning_rate": 0.0005350544122400603,
      "loss": 3.7789,
      "step": 10150
    },
    {
      "epoch": 1.0978366160800774,
      "grad_norm": 0.5674111247062683,
      "learning_rate": 0.0005347311712099989,
      "loss": 3.7766,
      "step": 10200
    },
    {
      "epoch": 1.1032181681196858,
      "grad_norm": 0.5893213748931885,
      "learning_rate": 0.0005344079301799374,
      "loss": 3.775,
      "step": 10250
    },
    {
      "epoch": 1.1085997201592939,
      "grad_norm": 0.6958675384521484,
      "learning_rate": 0.000534084689149876,
      "loss": 3.7703,
      "step": 10300
    },
    {
      "epoch": 1.1139812721989022,
      "grad_norm": 0.5321887135505676,
      "learning_rate": 0.0005337614481198147,
      "loss": 3.7636,
      "step": 10350
    },
    {
      "epoch": 1.1193628242385103,
      "grad_norm": 0.5547024607658386,
      "learning_rate": 0.0005334382070897532,
      "loss": 3.7686,
      "step": 10400
    },
    {
      "epoch": 1.1247443762781186,
      "grad_norm": 0.5801669359207153,
      "learning_rate": 0.0005331149660596918,
      "loss": 3.7487,
      "step": 10450
    },
    {
      "epoch": 1.1301259283177267,
      "grad_norm": 0.5799009799957275,
      "learning_rate": 0.0005327917250296303,
      "loss": 3.761,
      "step": 10500
    },
    {
      "epoch": 1.135507480357335,
      "grad_norm": 0.5902183651924133,
      "learning_rate": 0.000532468483999569,
      "loss": 3.7607,
      "step": 10550
    },
    {
      "epoch": 1.1408890323969434,
      "grad_norm": 0.5588002800941467,
      "learning_rate": 0.0005321452429695076,
      "loss": 3.7727,
      "step": 10600
    },
    {
      "epoch": 1.1462705844365515,
      "grad_norm": 0.633770227432251,
      "learning_rate": 0.0005318220019394462,
      "loss": 3.7646,
      "step": 10650
    },
    {
      "epoch": 1.1516521364761596,
      "grad_norm": 0.5668737292289734,
      "learning_rate": 0.0005314987609093847,
      "loss": 3.7847,
      "step": 10700
    },
    {
      "epoch": 1.157033688515768,
      "grad_norm": 0.5726791620254517,
      "learning_rate": 0.0005311755198793233,
      "loss": 3.7619,
      "step": 10750
    },
    {
      "epoch": 1.1624152405553763,
      "grad_norm": 0.6115407347679138,
      "learning_rate": 0.0005308522788492619,
      "loss": 3.7672,
      "step": 10800
    },
    {
      "epoch": 1.1677967925949844,
      "grad_norm": 0.6202018857002258,
      "learning_rate": 0.0005305290378192004,
      "loss": 3.7785,
      "step": 10850
    },
    {
      "epoch": 1.1731783446345927,
      "grad_norm": 0.5638216733932495,
      "learning_rate": 0.0005302057967891391,
      "loss": 3.7813,
      "step": 10900
    },
    {
      "epoch": 1.1785598966742008,
      "grad_norm": 0.5763724446296692,
      "learning_rate": 0.0005298825557590776,
      "loss": 3.7563,
      "step": 10950
    },
    {
      "epoch": 1.1839414487138091,
      "grad_norm": 0.52304607629776,
      "learning_rate": 0.0005295593147290163,
      "loss": 3.748,
      "step": 11000
    },
    {
      "epoch": 1.1839414487138091,
      "eval_accuracy": 0.3460101192780537,
      "eval_loss": 3.728248119354248,
      "eval_runtime": 184.4731,
      "eval_samples_per_second": 97.635,
      "eval_steps_per_second": 6.104,
      "step": 11000
    },
    {
      "epoch": 1.1893230007534172,
      "grad_norm": 0.5795536041259766,
      "learning_rate": 0.0005292360736989548,
      "loss": 3.7537,
      "step": 11050
    },
    {
      "epoch": 1.1947045527930256,
      "grad_norm": 0.5533538460731506,
      "learning_rate": 0.0005289128326688933,
      "loss": 3.7474,
      "step": 11100
    },
    {
      "epoch": 1.2000861048326337,
      "grad_norm": 0.5671941637992859,
      "learning_rate": 0.000528589591638832,
      "loss": 3.749,
      "step": 11150
    },
    {
      "epoch": 1.205467656872242,
      "grad_norm": 0.5548419952392578,
      "learning_rate": 0.0005282663506087705,
      "loss": 3.7523,
      "step": 11200
    },
    {
      "epoch": 1.21084920891185,
      "grad_norm": 0.601361095905304,
      "learning_rate": 0.0005279431095787092,
      "loss": 3.7471,
      "step": 11250
    },
    {
      "epoch": 1.2162307609514584,
      "grad_norm": 0.6410704255104065,
      "learning_rate": 0.0005276198685486477,
      "loss": 3.7612,
      "step": 11300
    },
    {
      "epoch": 1.2216123129910665,
      "grad_norm": 0.5552122592926025,
      "learning_rate": 0.0005272966275185863,
      "loss": 3.7434,
      "step": 11350
    },
    {
      "epoch": 1.2269938650306749,
      "grad_norm": 0.5526418685913086,
      "learning_rate": 0.0005269733864885249,
      "loss": 3.7377,
      "step": 11400
    },
    {
      "epoch": 1.232375417070283,
      "grad_norm": 0.6743530631065369,
      "learning_rate": 0.0005266501454584636,
      "loss": 3.7469,
      "step": 11450
    },
    {
      "epoch": 1.2377569691098913,
      "grad_norm": 0.5308992862701416,
      "learning_rate": 0.0005263269044284021,
      "loss": 3.743,
      "step": 11500
    },
    {
      "epoch": 1.2431385211494996,
      "grad_norm": 0.6078556776046753,
      "learning_rate": 0.0005260036633983406,
      "loss": 3.7301,
      "step": 11550
    },
    {
      "epoch": 1.2485200731891077,
      "grad_norm": 0.6298527121543884,
      "learning_rate": 0.0005256804223682792,
      "loss": 3.738,
      "step": 11600
    },
    {
      "epoch": 1.2539016252287158,
      "grad_norm": 0.6837873458862305,
      "learning_rate": 0.0005253571813382178,
      "loss": 3.7637,
      "step": 11650
    },
    {
      "epoch": 1.2592831772683242,
      "grad_norm": 0.5754698514938354,
      "learning_rate": 0.0005250339403081564,
      "loss": 3.7493,
      "step": 11700
    },
    {
      "epoch": 1.2646647293079325,
      "grad_norm": 0.5734676122665405,
      "learning_rate": 0.000524710699278095,
      "loss": 3.7391,
      "step": 11750
    },
    {
      "epoch": 1.2700462813475406,
      "grad_norm": 0.5261704325675964,
      "learning_rate": 0.0005243874582480336,
      "loss": 3.75,
      "step": 11800
    },
    {
      "epoch": 1.275427833387149,
      "grad_norm": 0.5651998519897461,
      "learning_rate": 0.0005240642172179722,
      "loss": 3.7287,
      "step": 11850
    },
    {
      "epoch": 1.280809385426757,
      "grad_norm": 0.6202215552330017,
      "learning_rate": 0.0005237409761879107,
      "loss": 3.7528,
      "step": 11900
    },
    {
      "epoch": 1.2861909374663654,
      "grad_norm": 0.5493283271789551,
      "learning_rate": 0.0005234177351578493,
      "loss": 3.7276,
      "step": 11950
    },
    {
      "epoch": 1.2915724895059735,
      "grad_norm": 0.6286284327507019,
      "learning_rate": 0.0005230944941277878,
      "loss": 3.7604,
      "step": 12000
    },
    {
      "epoch": 1.2915724895059735,
      "eval_accuracy": 0.34891256387024194,
      "eval_loss": 3.701730966567993,
      "eval_runtime": 184.0502,
      "eval_samples_per_second": 97.859,
      "eval_steps_per_second": 6.118,
      "step": 12000
    },
    {
      "epoch": 1.2969540415455818,
      "grad_norm": 0.6389520764350891,
      "learning_rate": 0.0005227712530977265,
      "loss": 3.7321,
      "step": 12050
    },
    {
      "epoch": 1.30233559358519,
      "grad_norm": 0.5742055177688599,
      "learning_rate": 0.0005224480120676651,
      "loss": 3.7233,
      "step": 12100
    },
    {
      "epoch": 1.3077171456247982,
      "grad_norm": 0.5082948803901672,
      "learning_rate": 0.0005221247710376037,
      "loss": 3.7462,
      "step": 12150
    },
    {
      "epoch": 1.3130986976644063,
      "grad_norm": 0.5427027344703674,
      "learning_rate": 0.0005218015300075422,
      "loss": 3.7197,
      "step": 12200
    },
    {
      "epoch": 1.3184802497040147,
      "grad_norm": 0.5587579607963562,
      "learning_rate": 0.0005214782889774809,
      "loss": 3.7339,
      "step": 12250
    },
    {
      "epoch": 1.3238618017436228,
      "grad_norm": 0.61777263879776,
      "learning_rate": 0.0005211550479474194,
      "loss": 3.7294,
      "step": 12300
    },
    {
      "epoch": 1.329243353783231,
      "grad_norm": 0.6232154369354248,
      "learning_rate": 0.000520831806917358,
      "loss": 3.7185,
      "step": 12350
    },
    {
      "epoch": 1.3346249058228392,
      "grad_norm": 0.6058857440948486,
      "learning_rate": 0.0005205085658872966,
      "loss": 3.7318,
      "step": 12400
    },
    {
      "epoch": 1.3400064578624475,
      "grad_norm": 0.5883937478065491,
      "learning_rate": 0.0005201853248572351,
      "loss": 3.7151,
      "step": 12450
    },
    {
      "epoch": 1.3453880099020559,
      "grad_norm": 0.6621643900871277,
      "learning_rate": 0.0005198620838271738,
      "loss": 3.7424,
      "step": 12500
    },
    {
      "epoch": 1.350769561941664,
      "grad_norm": 0.5513103008270264,
      "learning_rate": 0.0005195388427971123,
      "loss": 3.7475,
      "step": 12550
    },
    {
      "epoch": 1.356151113981272,
      "grad_norm": 0.6361138820648193,
      "learning_rate": 0.000519215601767051,
      "loss": 3.7202,
      "step": 12600
    },
    {
      "epoch": 1.3615326660208804,
      "grad_norm": 0.5991111397743225,
      "learning_rate": 0.0005188923607369895,
      "loss": 3.7252,
      "step": 12650
    },
    {
      "epoch": 1.3669142180604887,
      "grad_norm": 0.6061519384384155,
      "learning_rate": 0.000518569119706928,
      "loss": 3.73,
      "step": 12700
    },
    {
      "epoch": 1.3722957701000968,
      "grad_norm": 0.5958523750305176,
      "learning_rate": 0.0005182458786768667,
      "loss": 3.7286,
      "step": 12750
    },
    {
      "epoch": 1.3776773221397052,
      "grad_norm": 0.5368340611457825,
      "learning_rate": 0.0005179226376468052,
      "loss": 3.7402,
      "step": 12800
    },
    {
      "epoch": 1.3830588741793133,
      "grad_norm": 0.5887826085090637,
      "learning_rate": 0.0005175993966167438,
      "loss": 3.7141,
      "step": 12850
    },
    {
      "epoch": 1.3884404262189216,
      "grad_norm": 0.5436802506446838,
      "learning_rate": 0.0005172761555866824,
      "loss": 3.7147,
      "step": 12900
    },
    {
      "epoch": 1.3938219782585297,
      "grad_norm": 0.5853438377380371,
      "learning_rate": 0.0005169529145566211,
      "loss": 3.7338,
      "step": 12950
    },
    {
      "epoch": 1.399203530298138,
      "grad_norm": 0.5979605317115784,
      "learning_rate": 0.0005166296735265596,
      "loss": 3.7174,
      "step": 13000
    },
    {
      "epoch": 1.399203530298138,
      "eval_accuracy": 0.3508014943248967,
      "eval_loss": 3.6826083660125732,
      "eval_runtime": 184.2248,
      "eval_samples_per_second": 97.766,
      "eval_steps_per_second": 6.112,
      "step": 13000
    },
    {
      "epoch": 1.4045850823377461,
      "grad_norm": 0.5936151742935181,
      "learning_rate": 0.0005163064324964982,
      "loss": 3.7369,
      "step": 13050
    },
    {
      "epoch": 1.4099666343773545,
      "grad_norm": 0.6222744584083557,
      "learning_rate": 0.0005159831914664367,
      "loss": 3.7154,
      "step": 13100
    },
    {
      "epoch": 1.4153481864169626,
      "grad_norm": 0.5368512272834778,
      "learning_rate": 0.0005156599504363753,
      "loss": 3.6889,
      "step": 13150
    },
    {
      "epoch": 1.420729738456571,
      "grad_norm": 0.5740714073181152,
      "learning_rate": 0.000515336709406314,
      "loss": 3.7157,
      "step": 13200
    },
    {
      "epoch": 1.426111290496179,
      "grad_norm": 0.5489821434020996,
      "learning_rate": 0.0005150134683762525,
      "loss": 3.7377,
      "step": 13250
    },
    {
      "epoch": 1.4314928425357873,
      "grad_norm": 0.5233475565910339,
      "learning_rate": 0.0005146902273461911,
      "loss": 3.7081,
      "step": 13300
    },
    {
      "epoch": 1.4368743945753955,
      "grad_norm": 0.5411694049835205,
      "learning_rate": 0.0005143669863161297,
      "loss": 3.7221,
      "step": 13350
    },
    {
      "epoch": 1.4422559466150038,
      "grad_norm": 0.5820092558860779,
      "learning_rate": 0.0005140437452860683,
      "loss": 3.708,
      "step": 13400
    },
    {
      "epoch": 1.447637498654612,
      "grad_norm": 0.6159219145774841,
      "learning_rate": 0.0005137205042560069,
      "loss": 3.7193,
      "step": 13450
    },
    {
      "epoch": 1.4530190506942202,
      "grad_norm": 0.5878141522407532,
      "learning_rate": 0.0005133972632259455,
      "loss": 3.7209,
      "step": 13500
    },
    {
      "epoch": 1.4584006027338283,
      "grad_norm": 0.574815571308136,
      "learning_rate": 0.000513074022195884,
      "loss": 3.721,
      "step": 13550
    },
    {
      "epoch": 1.4637821547734367,
      "grad_norm": 0.5873903632164001,
      "learning_rate": 0.0005127507811658226,
      "loss": 3.7024,
      "step": 13600
    },
    {
      "epoch": 1.469163706813045,
      "grad_norm": 0.9297853708267212,
      "learning_rate": 0.0005124275401357612,
      "loss": 3.7091,
      "step": 13650
    },
    {
      "epoch": 1.474545258852653,
      "grad_norm": 0.6174495220184326,
      "learning_rate": 0.0005121042991056997,
      "loss": 3.6912,
      "step": 13700
    },
    {
      "epoch": 1.4799268108922612,
      "grad_norm": 0.5224736332893372,
      "learning_rate": 0.0005117810580756384,
      "loss": 3.6908,
      "step": 13750
    },
    {
      "epoch": 1.4853083629318695,
      "grad_norm": 0.623096764087677,
      "learning_rate": 0.0005114578170455769,
      "loss": 3.675,
      "step": 13800
    },
    {
      "epoch": 1.4906899149714778,
      "grad_norm": 0.5642491579055786,
      "learning_rate": 0.0005111345760155156,
      "loss": 3.7033,
      "step": 13850
    },
    {
      "epoch": 1.496071467011086,
      "grad_norm": 0.5480178594589233,
      "learning_rate": 0.0005108113349854541,
      "loss": 3.7064,
      "step": 13900
    },
    {
      "epoch": 1.501453019050694,
      "grad_norm": 0.5862089395523071,
      "learning_rate": 0.0005104880939553926,
      "loss": 3.6888,
      "step": 13950
    },
    {
      "epoch": 1.5068345710903024,
      "grad_norm": 0.5604771375656128,
      "learning_rate": 0.0005101648529253313,
      "loss": 3.6922,
      "step": 14000
    },
    {
      "epoch": 1.5068345710903024,
      "eval_accuracy": 0.3528002728491319,
      "eval_loss": 3.6652770042419434,
      "eval_runtime": 184.2712,
      "eval_samples_per_second": 97.742,
      "eval_steps_per_second": 6.111,
      "step": 14000
    },
    {
      "epoch": 1.5122161231299107,
      "grad_norm": 0.5487949252128601,
      "learning_rate": 0.0005098416118952699,
      "loss": 3.7047,
      "step": 14050
    },
    {
      "epoch": 1.5175976751695188,
      "grad_norm": 0.6444543600082397,
      "learning_rate": 0.0005095183708652085,
      "loss": 3.6964,
      "step": 14100
    },
    {
      "epoch": 1.5229792272091272,
      "grad_norm": 0.5607266426086426,
      "learning_rate": 0.000509195129835147,
      "loss": 3.7153,
      "step": 14150
    },
    {
      "epoch": 1.5283607792487355,
      "grad_norm": 0.5527377128601074,
      "learning_rate": 0.0005088718888050856,
      "loss": 3.7026,
      "step": 14200
    },
    {
      "epoch": 1.5337423312883436,
      "grad_norm": 0.556549608707428,
      "learning_rate": 0.0005085486477750242,
      "loss": 3.6994,
      "step": 14250
    },
    {
      "epoch": 1.5391238833279517,
      "grad_norm": 0.5716401934623718,
      "learning_rate": 0.0005082254067449629,
      "loss": 3.6973,
      "step": 14300
    },
    {
      "epoch": 1.54450543536756,
      "grad_norm": 0.7053285837173462,
      "learning_rate": 0.0005079086305355026,
      "loss": 3.684,
      "step": 14350
    },
    {
      "epoch": 1.5498869874071683,
      "grad_norm": 0.6206086277961731,
      "learning_rate": 0.0005075853895054412,
      "loss": 3.7008,
      "step": 14400
    },
    {
      "epoch": 1.5552685394467765,
      "grad_norm": 0.5947590470314026,
      "learning_rate": 0.0005072621484753797,
      "loss": 3.6984,
      "step": 14450
    },
    {
      "epoch": 1.5606500914863846,
      "grad_norm": 0.5644803047180176,
      "learning_rate": 0.0005069389074453184,
      "loss": 3.692,
      "step": 14500
    },
    {
      "epoch": 1.566031643525993,
      "grad_norm": 0.512422502040863,
      "learning_rate": 0.0005066156664152569,
      "loss": 3.6855,
      "step": 14550
    },
    {
      "epoch": 1.5714131955656012,
      "grad_norm": 0.564999520778656,
      "learning_rate": 0.0005062924253851955,
      "loss": 3.6967,
      "step": 14600
    },
    {
      "epoch": 1.5767947476052093,
      "grad_norm": 0.5477495789527893,
      "learning_rate": 0.0005059691843551341,
      "loss": 3.6851,
      "step": 14650
    },
    {
      "epoch": 1.5821762996448174,
      "grad_norm": 0.5937101244926453,
      "learning_rate": 0.0005056459433250727,
      "loss": 3.6844,
      "step": 14700
    },
    {
      "epoch": 1.5875578516844258,
      "grad_norm": 0.6070849895477295,
      "learning_rate": 0.0005053227022950113,
      "loss": 3.6785,
      "step": 14750
    },
    {
      "epoch": 1.592939403724034,
      "grad_norm": 0.587536096572876,
      "learning_rate": 0.0005049994612649499,
      "loss": 3.6996,
      "step": 14800
    },
    {
      "epoch": 1.5983209557636422,
      "grad_norm": 0.6087131500244141,
      "learning_rate": 0.0005046762202348884,
      "loss": 3.6795,
      "step": 14850
    },
    {
      "epoch": 1.6037025078032503,
      "grad_norm": 0.6166165471076965,
      "learning_rate": 0.000504352979204827,
      "loss": 3.678,
      "step": 14900
    },
    {
      "epoch": 1.6090840598428586,
      "grad_norm": 0.56649249792099,
      "learning_rate": 0.0005040297381747656,
      "loss": 3.6812,
      "step": 14950
    },
    {
      "epoch": 1.614465611882467,
      "grad_norm": 0.6390756964683533,
      "learning_rate": 0.0005037064971447042,
      "loss": 3.6811,
      "step": 15000
    },
    {
      "epoch": 1.614465611882467,
      "eval_accuracy": 0.3544349555463007,
      "eval_loss": 3.6462361812591553,
      "eval_runtime": 184.0934,
      "eval_samples_per_second": 97.836,
      "eval_steps_per_second": 6.116,
      "step": 15000
    },
    {
      "epoch": 1.619847163922075,
      "grad_norm": 0.5951522588729858,
      "learning_rate": 0.0005033832561146428,
      "loss": 3.6841,
      "step": 15050
    },
    {
      "epoch": 1.6252287159616834,
      "grad_norm": 0.5375372171401978,
      "learning_rate": 0.0005030600150845813,
      "loss": 3.6811,
      "step": 15100
    },
    {
      "epoch": 1.6306102680012917,
      "grad_norm": 0.5611913204193115,
      "learning_rate": 0.00050273677405452,
      "loss": 3.688,
      "step": 15150
    },
    {
      "epoch": 1.6359918200408998,
      "grad_norm": 0.631302535533905,
      "learning_rate": 0.0005024135330244585,
      "loss": 3.6667,
      "step": 15200
    },
    {
      "epoch": 1.641373372080508,
      "grad_norm": 0.623928964138031,
      "learning_rate": 0.0005020902919943972,
      "loss": 3.6895,
      "step": 15250
    },
    {
      "epoch": 1.6467549241201163,
      "grad_norm": 0.5371923446655273,
      "learning_rate": 0.0005017670509643357,
      "loss": 3.6985,
      "step": 15300
    },
    {
      "epoch": 1.6521364761597246,
      "grad_norm": 0.5388199687004089,
      "learning_rate": 0.0005014438099342743,
      "loss": 3.6888,
      "step": 15350
    },
    {
      "epoch": 1.6575180281993327,
      "grad_norm": 0.5434895753860474,
      "learning_rate": 0.0005011205689042129,
      "loss": 3.681,
      "step": 15400
    },
    {
      "epoch": 1.6628995802389408,
      "grad_norm": 0.5811678767204285,
      "learning_rate": 0.0005007973278741514,
      "loss": 3.691,
      "step": 15450
    },
    {
      "epoch": 1.6682811322785491,
      "grad_norm": 0.5489638447761536,
      "learning_rate": 0.00050047408684409,
      "loss": 3.6756,
      "step": 15500
    },
    {
      "epoch": 1.6736626843181575,
      "grad_norm": 0.5695392489433289,
      "learning_rate": 0.0005001508458140286,
      "loss": 3.666,
      "step": 15550
    },
    {
      "epoch": 1.6790442363577656,
      "grad_norm": 0.6463865041732788,
      "learning_rate": 0.0004998276047839673,
      "loss": 3.6623,
      "step": 15600
    },
    {
      "epoch": 1.6844257883973737,
      "grad_norm": 0.5845815539360046,
      "learning_rate": 0.0004995043637539058,
      "loss": 3.6865,
      "step": 15650
    },
    {
      "epoch": 1.689807340436982,
      "grad_norm": 0.7483472228050232,
      "learning_rate": 0.0004991811227238443,
      "loss": 3.6742,
      "step": 15700
    },
    {
      "epoch": 1.6951888924765903,
      "grad_norm": 0.5330408215522766,
      "learning_rate": 0.0004988578816937829,
      "loss": 3.6795,
      "step": 15750
    },
    {
      "epoch": 1.7005704445161984,
      "grad_norm": 0.6368387937545776,
      "learning_rate": 0.0004985346406637215,
      "loss": 3.6749,
      "step": 15800
    },
    {
      "epoch": 1.7059519965558065,
      "grad_norm": 0.5391474366188049,
      "learning_rate": 0.0004982113996336602,
      "loss": 3.6582,
      "step": 15850
    },
    {
      "epoch": 1.7113335485954149,
      "grad_norm": 0.5919294953346252,
      "learning_rate": 0.0004978881586035987,
      "loss": 3.66,
      "step": 15900
    },
    {
      "epoch": 1.7167151006350232,
      "grad_norm": 0.5829312205314636,
      "learning_rate": 0.0004975649175735373,
      "loss": 3.6612,
      "step": 15950
    },
    {
      "epoch": 1.7220966526746313,
      "grad_norm": 0.5279169082641602,
      "learning_rate": 0.0004972416765434759,
      "loss": 3.6861,
      "step": 16000
    },
    {
      "epoch": 1.7220966526746313,
      "eval_accuracy": 0.3568658466053741,
      "eval_loss": 3.6250016689300537,
      "eval_runtime": 184.2201,
      "eval_samples_per_second": 97.769,
      "eval_steps_per_second": 6.112,
      "step": 16000
    },
    {
      "epoch": 1.7274782047142396,
      "grad_norm": 0.54816073179245,
      "learning_rate": 0.0004969184355134145,
      "loss": 3.667,
      "step": 16050
    },
    {
      "epoch": 1.732859756753848,
      "grad_norm": 0.5427775979042053,
      "learning_rate": 0.0004965951944833531,
      "loss": 3.6573,
      "step": 16100
    },
    {
      "epoch": 1.738241308793456,
      "grad_norm": 0.5446629524230957,
      "learning_rate": 0.0004962719534532916,
      "loss": 3.6872,
      "step": 16150
    },
    {
      "epoch": 1.7436228608330642,
      "grad_norm": 0.6108595728874207,
      "learning_rate": 0.0004959487124232302,
      "loss": 3.6735,
      "step": 16200
    },
    {
      "epoch": 1.7490044128726725,
      "grad_norm": 0.5399534106254578,
      "learning_rate": 0.0004956254713931688,
      "loss": 3.6574,
      "step": 16250
    },
    {
      "epoch": 1.7543859649122808,
      "grad_norm": 0.6165593266487122,
      "learning_rate": 0.0004953022303631074,
      "loss": 3.6531,
      "step": 16300
    },
    {
      "epoch": 1.759767516951889,
      "grad_norm": 0.5562168955802917,
      "learning_rate": 0.0004949854541536472,
      "loss": 3.6588,
      "step": 16350
    },
    {
      "epoch": 1.765149068991497,
      "grad_norm": 0.5855176448822021,
      "learning_rate": 0.0004946622131235857,
      "loss": 3.665,
      "step": 16400
    },
    {
      "epoch": 1.7705306210311054,
      "grad_norm": 0.5627042055130005,
      "learning_rate": 0.0004943389720935244,
      "loss": 3.6703,
      "step": 16450
    },
    {
      "epoch": 1.7759121730707137,
      "grad_norm": 0.5681560635566711,
      "learning_rate": 0.000494015731063463,
      "loss": 3.6605,
      "step": 16500
    },
    {
      "epoch": 1.7812937251103218,
      "grad_norm": 0.5521273016929626,
      "learning_rate": 0.0004936924900334016,
      "loss": 3.6669,
      "step": 16550
    },
    {
      "epoch": 1.78667527714993,
      "grad_norm": 0.5735337734222412,
      "learning_rate": 0.0004933692490033401,
      "loss": 3.6618,
      "step": 16600
    },
    {
      "epoch": 1.7920568291895382,
      "grad_norm": 0.5497586727142334,
      "learning_rate": 0.0004930460079732786,
      "loss": 3.6593,
      "step": 16650
    },
    {
      "epoch": 1.7974383812291466,
      "grad_norm": 0.5310925841331482,
      "learning_rate": 0.0004927227669432173,
      "loss": 3.6708,
      "step": 16700
    },
    {
      "epoch": 1.8028199332687547,
      "grad_norm": 0.648143470287323,
      "learning_rate": 0.0004923995259131558,
      "loss": 3.6626,
      "step": 16750
    },
    {
      "epoch": 1.8082014853083628,
      "grad_norm": 0.5871980786323547,
      "learning_rate": 0.0004920762848830945,
      "loss": 3.6488,
      "step": 16800
    },
    {
      "epoch": 1.813583037347971,
      "grad_norm": 0.5406138896942139,
      "learning_rate": 0.000491753043853033,
      "loss": 3.6708,
      "step": 16850
    },
    {
      "epoch": 1.8189645893875794,
      "grad_norm": 0.5458976626396179,
      "learning_rate": 0.0004914298028229717,
      "loss": 3.6545,
      "step": 16900
    },
    {
      "epoch": 1.8243461414271875,
      "grad_norm": 0.5447959303855896,
      "learning_rate": 0.0004911065617929102,
      "loss": 3.6692,
      "step": 16950
    },
    {
      "epoch": 1.8297276934667959,
      "grad_norm": 0.6151652932167053,
      "learning_rate": 0.0004907833207628487,
      "loss": 3.6621,
      "step": 17000
    },
    {
      "epoch": 1.8297276934667959,
      "eval_accuracy": 0.3579900780355906,
      "eval_loss": 3.6100103855133057,
      "eval_runtime": 184.3956,
      "eval_samples_per_second": 97.676,
      "eval_steps_per_second": 6.106,
      "step": 17000
    },
    {
      "epoch": 1.8351092455064042,
      "grad_norm": 0.5729044079780579,
      "learning_rate": 0.0004904600797327874,
      "loss": 3.6535,
      "step": 17050
    },
    {
      "epoch": 1.8404907975460123,
      "grad_norm": 0.6575093269348145,
      "learning_rate": 0.0004901368387027259,
      "loss": 3.657,
      "step": 17100
    },
    {
      "epoch": 1.8458723495856204,
      "grad_norm": 0.5465648174285889,
      "learning_rate": 0.0004898135976726646,
      "loss": 3.6279,
      "step": 17150
    },
    {
      "epoch": 1.8512539016252287,
      "grad_norm": 0.5691052079200745,
      "learning_rate": 0.0004894903566426031,
      "loss": 3.663,
      "step": 17200
    },
    {
      "epoch": 1.856635453664837,
      "grad_norm": 0.6024985313415527,
      "learning_rate": 0.0004891671156125417,
      "loss": 3.646,
      "step": 17250
    },
    {
      "epoch": 1.8620170057044452,
      "grad_norm": 0.5859248042106628,
      "learning_rate": 0.0004888438745824803,
      "loss": 3.6554,
      "step": 17300
    },
    {
      "epoch": 1.8673985577440533,
      "grad_norm": 0.586715817451477,
      "learning_rate": 0.0004885206335524189,
      "loss": 3.6427,
      "step": 17350
    },
    {
      "epoch": 1.8727801097836616,
      "grad_norm": 0.5394769310951233,
      "learning_rate": 0.0004881973925223575,
      "loss": 3.6445,
      "step": 17400
    },
    {
      "epoch": 1.87816166182327,
      "grad_norm": 0.6003996133804321,
      "learning_rate": 0.00048787415149229604,
      "loss": 3.6574,
      "step": 17450
    },
    {
      "epoch": 1.883543213862878,
      "grad_norm": 0.5910128951072693,
      "learning_rate": 0.00048755091046223464,
      "loss": 3.6666,
      "step": 17500
    },
    {
      "epoch": 1.8889247659024861,
      "grad_norm": 0.5640808939933777,
      "learning_rate": 0.0004872276694321732,
      "loss": 3.6436,
      "step": 17550
    },
    {
      "epoch": 1.8943063179420945,
      "grad_norm": 0.5623183250427246,
      "learning_rate": 0.00048690442840211177,
      "loss": 3.6493,
      "step": 17600
    },
    {
      "epoch": 1.8996878699817028,
      "grad_norm": 0.5348426103591919,
      "learning_rate": 0.0004865811873720504,
      "loss": 3.6458,
      "step": 17650
    },
    {
      "epoch": 1.905069422021311,
      "grad_norm": 0.5619913339614868,
      "learning_rate": 0.00048625794634198896,
      "loss": 3.643,
      "step": 17700
    },
    {
      "epoch": 1.910450974060919,
      "grad_norm": 0.6076251268386841,
      "learning_rate": 0.00048593470531192756,
      "loss": 3.647,
      "step": 17750
    },
    {
      "epoch": 1.9158325261005273,
      "grad_norm": 0.6116069555282593,
      "learning_rate": 0.00048561146428186615,
      "loss": 3.6481,
      "step": 17800
    },
    {
      "epoch": 1.9212140781401357,
      "grad_norm": 0.5479802489280701,
      "learning_rate": 0.0004852882232518047,
      "loss": 3.6412,
      "step": 17850
    },
    {
      "epoch": 1.9265956301797438,
      "grad_norm": 0.5557817816734314,
      "learning_rate": 0.00048496498222174334,
      "loss": 3.6561,
      "step": 17900
    },
    {
      "epoch": 1.931977182219352,
      "grad_norm": 0.5599830150604248,
      "learning_rate": 0.00048464174119168193,
      "loss": 3.6369,
      "step": 17950
    },
    {
      "epoch": 1.9373587342589604,
      "grad_norm": 0.5717028975486755,
      "learning_rate": 0.0004843185001616205,
      "loss": 3.6516,
      "step": 18000
    },
    {
      "epoch": 1.9373587342589604,
      "eval_accuracy": 0.3590484658157914,
      "eval_loss": 3.5976455211639404,
      "eval_runtime": 184.3209,
      "eval_samples_per_second": 97.715,
      "eval_steps_per_second": 6.109,
      "step": 18000
    },
    {
      "epoch": 1.9427402862985685,
      "grad_norm": 0.540344774723053,
      "learning_rate": 0.00048399525913155907,
      "loss": 3.6547,
      "step": 18050
    },
    {
      "epoch": 1.9481218383381766,
      "grad_norm": 0.6229539513587952,
      "learning_rate": 0.0004836720181014976,
      "loss": 3.6394,
      "step": 18100
    },
    {
      "epoch": 1.953503390377785,
      "grad_norm": 0.6036831736564636,
      "learning_rate": 0.0004833487770714362,
      "loss": 3.6598,
      "step": 18150
    },
    {
      "epoch": 1.9588849424173933,
      "grad_norm": 0.6020011305809021,
      "learning_rate": 0.00048302553604137485,
      "loss": 3.627,
      "step": 18200
    },
    {
      "epoch": 1.9642664944570014,
      "grad_norm": 0.549821138381958,
      "learning_rate": 0.0004827022950113134,
      "loss": 3.6389,
      "step": 18250
    },
    {
      "epoch": 1.9696480464966095,
      "grad_norm": 0.5844795107841492,
      "learning_rate": 0.000482379053981252,
      "loss": 3.6501,
      "step": 18300
    },
    {
      "epoch": 1.9750295985362178,
      "grad_norm": 0.5954594612121582,
      "learning_rate": 0.0004820558129511906,
      "loss": 3.6342,
      "step": 18350
    },
    {
      "epoch": 1.9804111505758262,
      "grad_norm": 0.5975394248962402,
      "learning_rate": 0.0004817325719211291,
      "loss": 3.6399,
      "step": 18400
    },
    {
      "epoch": 1.9857927026154343,
      "grad_norm": 0.5545531511306763,
      "learning_rate": 0.0004814093308910677,
      "loss": 3.6376,
      "step": 18450
    },
    {
      "epoch": 1.9911742546550424,
      "grad_norm": 0.6136595010757446,
      "learning_rate": 0.0004810925546816075,
      "loss": 3.6449,
      "step": 18500
    },
    {
      "epoch": 1.9965558066946507,
      "grad_norm": 0.5912418961524963,
      "learning_rate": 0.0004807693136515461,
      "loss": 3.64,
      "step": 18550
    },
    {
      "epoch": 2.001937358734259,
      "grad_norm": 0.594277560710907,
      "learning_rate": 0.0004804460726214847,
      "loss": 3.6059,
      "step": 18600
    },
    {
      "epoch": 2.007318910773867,
      "grad_norm": 0.5589869618415833,
      "learning_rate": 0.0004801228315914233,
      "loss": 3.5623,
      "step": 18650
    },
    {
      "epoch": 2.0127004628134753,
      "grad_norm": 0.595622718334198,
      "learning_rate": 0.0004797995905613619,
      "loss": 3.5509,
      "step": 18700
    },
    {
      "epoch": 2.018082014853084,
      "grad_norm": 0.6202673316001892,
      "learning_rate": 0.00047947634953130044,
      "loss": 3.5577,
      "step": 18750
    },
    {
      "epoch": 2.023463566892692,
      "grad_norm": 0.5669680237770081,
      "learning_rate": 0.00047915310850123904,
      "loss": 3.5644,
      "step": 18800
    },
    {
      "epoch": 2.0288451189323,
      "grad_norm": 0.5390088558197021,
      "learning_rate": 0.0004788298674711777,
      "loss": 3.5522,
      "step": 18850
    },
    {
      "epoch": 2.034226670971908,
      "grad_norm": 0.5642072558403015,
      "learning_rate": 0.00047850662644111623,
      "loss": 3.559,
      "step": 18900
    },
    {
      "epoch": 2.0396082230115167,
      "grad_norm": 0.5674673318862915,
      "learning_rate": 0.0004781833854110548,
      "loss": 3.5566,
      "step": 18950
    },
    {
      "epoch": 2.044989775051125,
      "grad_norm": 0.6004260778427124,
      "learning_rate": 0.00047786014438099336,
      "loss": 3.5667,
      "step": 19000
    },
    {
      "epoch": 2.044989775051125,
      "eval_accuracy": 0.36095152114586204,
      "eval_loss": 3.5886144638061523,
      "eval_runtime": 184.1971,
      "eval_samples_per_second": 97.781,
      "eval_steps_per_second": 6.113,
      "step": 19000
    },
    {
      "epoch": 2.050371327090733,
      "grad_norm": 0.666540265083313,
      "learning_rate": 0.00047753690335093196,
      "loss": 3.569,
      "step": 19050
    },
    {
      "epoch": 2.055752879130341,
      "grad_norm": 0.6222928762435913,
      "learning_rate": 0.00047721366232087055,
      "loss": 3.5529,
      "step": 19100
    },
    {
      "epoch": 2.0611344311699495,
      "grad_norm": 0.5838487148284912,
      "learning_rate": 0.00047689042129080915,
      "loss": 3.5546,
      "step": 19150
    },
    {
      "epoch": 2.0665159832095576,
      "grad_norm": 0.6417465806007385,
      "learning_rate": 0.00047656718026074774,
      "loss": 3.5665,
      "step": 19200
    },
    {
      "epoch": 2.0718975352491658,
      "grad_norm": 0.5686841011047363,
      "learning_rate": 0.00047624393923068634,
      "loss": 3.5465,
      "step": 19250
    },
    {
      "epoch": 2.0772790872887743,
      "grad_norm": 0.5918363332748413,
      "learning_rate": 0.0004759206982006249,
      "loss": 3.5714,
      "step": 19300
    },
    {
      "epoch": 2.0826606393283824,
      "grad_norm": 0.6635754704475403,
      "learning_rate": 0.00047559745717056347,
      "loss": 3.5606,
      "step": 19350
    },
    {
      "epoch": 2.0880421913679905,
      "grad_norm": 0.5866388082504272,
      "learning_rate": 0.000475274216140502,
      "loss": 3.5583,
      "step": 19400
    },
    {
      "epoch": 2.0934237434075986,
      "grad_norm": 0.6153836250305176,
      "learning_rate": 0.00047495097511044066,
      "loss": 3.5632,
      "step": 19450
    },
    {
      "epoch": 2.098805295447207,
      "grad_norm": 0.5819552540779114,
      "learning_rate": 0.00047462773408037925,
      "loss": 3.5766,
      "step": 19500
    },
    {
      "epoch": 2.1041868474868153,
      "grad_norm": 0.5855319499969482,
      "learning_rate": 0.0004743044930503178,
      "loss": 3.5731,
      "step": 19550
    },
    {
      "epoch": 2.1095683995264234,
      "grad_norm": 0.583659291267395,
      "learning_rate": 0.0004739812520202564,
      "loss": 3.5702,
      "step": 19600
    },
    {
      "epoch": 2.1149499515660315,
      "grad_norm": 0.5505194067955017,
      "learning_rate": 0.000473658010990195,
      "loss": 3.5549,
      "step": 19650
    },
    {
      "epoch": 2.12033150360564,
      "grad_norm": 0.6187375783920288,
      "learning_rate": 0.0004733347699601336,
      "loss": 3.5626,
      "step": 19700
    },
    {
      "epoch": 2.125713055645248,
      "grad_norm": 0.6520795226097107,
      "learning_rate": 0.0004730115289300722,
      "loss": 3.5627,
      "step": 19750
    },
    {
      "epoch": 2.1310946076848563,
      "grad_norm": 0.574247419834137,
      "learning_rate": 0.00047268828790001077,
      "loss": 3.5748,
      "step": 19800
    },
    {
      "epoch": 2.1364761597244644,
      "grad_norm": 0.5729595422744751,
      "learning_rate": 0.0004723650468699493,
      "loss": 3.5586,
      "step": 19850
    },
    {
      "epoch": 2.141857711764073,
      "grad_norm": 0.5469658970832825,
      "learning_rate": 0.0004720418058398879,
      "loss": 3.5544,
      "step": 19900
    },
    {
      "epoch": 2.147239263803681,
      "grad_norm": 0.5846270322799683,
      "learning_rate": 0.00047171856480982644,
      "loss": 3.5631,
      "step": 19950
    },
    {
      "epoch": 2.152620815843289,
      "grad_norm": 0.5745915770530701,
      "learning_rate": 0.0004713953237797651,
      "loss": 3.5658,
      "step": 20000
    },
    {
      "epoch": 2.152620815843289,
      "eval_accuracy": 0.3618274807274224,
      "eval_loss": 3.5798401832580566,
      "eval_runtime": 184.5167,
      "eval_samples_per_second": 97.612,
      "eval_steps_per_second": 6.102,
      "step": 20000
    },
    {
      "epoch": 2.1580023678828972,
      "grad_norm": 0.5974456667900085,
      "learning_rate": 0.0004710720827497037,
      "loss": 3.5724,
      "step": 20050
    },
    {
      "epoch": 2.163383919922506,
      "grad_norm": 0.5742269158363342,
      "learning_rate": 0.00047074884171964223,
      "loss": 3.5592,
      "step": 20100
    },
    {
      "epoch": 2.168765471962114,
      "grad_norm": 0.6028411984443665,
      "learning_rate": 0.0004704256006895808,
      "loss": 3.5481,
      "step": 20150
    },
    {
      "epoch": 2.174147024001722,
      "grad_norm": 0.5908669829368591,
      "learning_rate": 0.00047010235965951936,
      "loss": 3.5879,
      "step": 20200
    },
    {
      "epoch": 2.1795285760413305,
      "grad_norm": 0.6563610434532166,
      "learning_rate": 0.00046977911862945796,
      "loss": 3.5475,
      "step": 20250
    },
    {
      "epoch": 2.1849101280809387,
      "grad_norm": 0.579626739025116,
      "learning_rate": 0.0004694558775993966,
      "loss": 3.5658,
      "step": 20300
    },
    {
      "epoch": 2.1902916801205468,
      "grad_norm": 0.5718281865119934,
      "learning_rate": 0.00046913263656933515,
      "loss": 3.5387,
      "step": 20350
    },
    {
      "epoch": 2.195673232160155,
      "grad_norm": 0.5388026237487793,
      "learning_rate": 0.00046880939553927374,
      "loss": 3.5557,
      "step": 20400
    },
    {
      "epoch": 2.2010547841997634,
      "grad_norm": 0.6123741865158081,
      "learning_rate": 0.00046848615450921234,
      "loss": 3.5459,
      "step": 20450
    },
    {
      "epoch": 2.2064363362393715,
      "grad_norm": 0.5772309303283691,
      "learning_rate": 0.0004681629134791509,
      "loss": 3.5662,
      "step": 20500
    },
    {
      "epoch": 2.2118178882789796,
      "grad_norm": 0.6064363121986389,
      "learning_rate": 0.00046784613726969074,
      "loss": 3.5674,
      "step": 20550
    },
    {
      "epoch": 2.2171994403185877,
      "grad_norm": 0.5686280131340027,
      "learning_rate": 0.0004675228962396293,
      "loss": 3.5523,
      "step": 20600
    },
    {
      "epoch": 2.2225809923581963,
      "grad_norm": 0.5950042605400085,
      "learning_rate": 0.0004671996552095679,
      "loss": 3.5816,
      "step": 20650
    },
    {
      "epoch": 2.2279625443978044,
      "grad_norm": 0.5808351635932922,
      "learning_rate": 0.0004668764141795065,
      "loss": 3.5671,
      "step": 20700
    },
    {
      "epoch": 2.2333440964374125,
      "grad_norm": 0.5976179838180542,
      "learning_rate": 0.00046655317314944506,
      "loss": 3.5787,
      "step": 20750
    },
    {
      "epoch": 2.2387256484770206,
      "grad_norm": 0.5799880623817444,
      "learning_rate": 0.00046622993211938366,
      "loss": 3.5677,
      "step": 20800
    },
    {
      "epoch": 2.244107200516629,
      "grad_norm": 0.6059759855270386,
      "learning_rate": 0.0004659066910893222,
      "loss": 3.5668,
      "step": 20850
    },
    {
      "epoch": 2.2494887525562373,
      "grad_norm": 0.6136220693588257,
      "learning_rate": 0.0004655834500592608,
      "loss": 3.5688,
      "step": 20900
    },
    {
      "epoch": 2.2548703045958454,
      "grad_norm": 0.6185546517372131,
      "learning_rate": 0.00046526020902919944,
      "loss": 3.5477,
      "step": 20950
    },
    {
      "epoch": 2.2602518566354535,
      "grad_norm": 0.5652609467506409,
      "learning_rate": 0.000464936967999138,
      "loss": 3.5641,
      "step": 21000
    },
    {
      "epoch": 2.2602518566354535,
      "eval_accuracy": 0.36298387341243193,
      "eval_loss": 3.567474365234375,
      "eval_runtime": 184.044,
      "eval_samples_per_second": 97.862,
      "eval_steps_per_second": 6.118,
      "step": 21000
    },
    {
      "epoch": 2.265633408675062,
      "grad_norm": 0.6556390523910522,
      "learning_rate": 0.0004646137269690766,
      "loss": 3.5542,
      "step": 21050
    },
    {
      "epoch": 2.27101496071467,
      "grad_norm": 0.5754619836807251,
      "learning_rate": 0.00046429048593901517,
      "loss": 3.5957,
      "step": 21100
    },
    {
      "epoch": 2.2763965127542782,
      "grad_norm": 0.6319244503974915,
      "learning_rate": 0.0004639672449089537,
      "loss": 3.5726,
      "step": 21150
    },
    {
      "epoch": 2.281778064793887,
      "grad_norm": 0.609137237071991,
      "learning_rate": 0.0004636440038788923,
      "loss": 3.561,
      "step": 21200
    },
    {
      "epoch": 2.287159616833495,
      "grad_norm": 0.6190239787101746,
      "learning_rate": 0.00046332076284883095,
      "loss": 3.5578,
      "step": 21250
    },
    {
      "epoch": 2.292541168873103,
      "grad_norm": 0.5698543787002563,
      "learning_rate": 0.0004629975218187695,
      "loss": 3.5682,
      "step": 21300
    },
    {
      "epoch": 2.297922720912711,
      "grad_norm": 0.5980182886123657,
      "learning_rate": 0.0004626742807887081,
      "loss": 3.586,
      "step": 21350
    },
    {
      "epoch": 2.303304272952319,
      "grad_norm": 0.6205169558525085,
      "learning_rate": 0.00046235103975864663,
      "loss": 3.5561,
      "step": 21400
    },
    {
      "epoch": 2.3086858249919278,
      "grad_norm": 0.6614342927932739,
      "learning_rate": 0.0004620277987285852,
      "loss": 3.5696,
      "step": 21450
    },
    {
      "epoch": 2.314067377031536,
      "grad_norm": 0.5894066691398621,
      "learning_rate": 0.0004617045576985239,
      "loss": 3.5809,
      "step": 21500
    },
    {
      "epoch": 2.319448929071144,
      "grad_norm": 0.743106484413147,
      "learning_rate": 0.0004613813166684624,
      "loss": 3.5723,
      "step": 21550
    },
    {
      "epoch": 2.3248304811107525,
      "grad_norm": 0.5671776533126831,
      "learning_rate": 0.000461058075638401,
      "loss": 3.5553,
      "step": 21600
    },
    {
      "epoch": 2.3302120331503606,
      "grad_norm": 0.5718901753425598,
      "learning_rate": 0.00046073483460833955,
      "loss": 3.559,
      "step": 21650
    },
    {
      "epoch": 2.3355935851899687,
      "grad_norm": 0.5780794024467468,
      "learning_rate": 0.00046041159357827814,
      "loss": 3.5487,
      "step": 21700
    },
    {
      "epoch": 2.340975137229577,
      "grad_norm": 0.6014887094497681,
      "learning_rate": 0.00046008835254821674,
      "loss": 3.5651,
      "step": 21750
    },
    {
      "epoch": 2.3463566892691854,
      "grad_norm": 0.5638294816017151,
      "learning_rate": 0.0004597651115181554,
      "loss": 3.5426,
      "step": 21800
    },
    {
      "epoch": 2.3517382413087935,
      "grad_norm": 0.6568518877029419,
      "learning_rate": 0.00045944187048809393,
      "loss": 3.5666,
      "step": 21850
    },
    {
      "epoch": 2.3571197933484016,
      "grad_norm": 0.5806652903556824,
      "learning_rate": 0.0004591186294580325,
      "loss": 3.5579,
      "step": 21900
    },
    {
      "epoch": 2.3625013453880097,
      "grad_norm": 0.5711958408355713,
      "learning_rate": 0.00045879538842797106,
      "loss": 3.5675,
      "step": 21950
    },
    {
      "epoch": 2.3678828974276183,
      "grad_norm": 0.6058608889579773,
      "learning_rate": 0.00045847214739790966,
      "loss": 3.5568,
      "step": 22000
    },
    {
      "epoch": 2.3678828974276183,
      "eval_accuracy": 0.36439299271449793,
      "eval_loss": 3.554220676422119,
      "eval_runtime": 184.5501,
      "eval_samples_per_second": 97.594,
      "eval_steps_per_second": 6.101,
      "step": 22000
    },
    {
      "epoch": 2.3732644494672264,
      "grad_norm": 0.6164901256561279,
      "learning_rate": 0.0004581489063678482,
      "loss": 3.5661,
      "step": 22050
    },
    {
      "epoch": 2.3786460015068345,
      "grad_norm": 0.6135861277580261,
      "learning_rate": 0.00045782566533778685,
      "loss": 3.5533,
      "step": 22100
    },
    {
      "epoch": 2.384027553546443,
      "grad_norm": 0.6621028184890747,
      "learning_rate": 0.00045750242430772544,
      "loss": 3.5605,
      "step": 22150
    },
    {
      "epoch": 2.389409105586051,
      "grad_norm": 0.6456934213638306,
      "learning_rate": 0.000457179183277664,
      "loss": 3.5674,
      "step": 22200
    },
    {
      "epoch": 2.3947906576256592,
      "grad_norm": 0.6146150827407837,
      "learning_rate": 0.0004568559422476026,
      "loss": 3.5687,
      "step": 22250
    },
    {
      "epoch": 2.4001722096652673,
      "grad_norm": 0.6114180088043213,
      "learning_rate": 0.00045653270121754117,
      "loss": 3.5443,
      "step": 22300
    },
    {
      "epoch": 2.4055537617048754,
      "grad_norm": 0.6257768273353577,
      "learning_rate": 0.0004562094601874797,
      "loss": 3.5545,
      "step": 22350
    },
    {
      "epoch": 2.410935313744484,
      "grad_norm": 0.5966166257858276,
      "learning_rate": 0.00045588621915741836,
      "loss": 3.5347,
      "step": 22400
    },
    {
      "epoch": 2.416316865784092,
      "grad_norm": 0.5521513819694519,
      "learning_rate": 0.00045556297812735696,
      "loss": 3.5613,
      "step": 22450
    },
    {
      "epoch": 2.4216984178237,
      "grad_norm": 0.5702809691429138,
      "learning_rate": 0.0004552397370972955,
      "loss": 3.5733,
      "step": 22500
    },
    {
      "epoch": 2.4270799698633088,
      "grad_norm": 0.5816466808319092,
      "learning_rate": 0.0004549164960672341,
      "loss": 3.574,
      "step": 22550
    },
    {
      "epoch": 2.432461521902917,
      "grad_norm": 0.5726392865180969,
      "learning_rate": 0.0004545997198577739,
      "loss": 3.5588,
      "step": 22600
    },
    {
      "epoch": 2.437843073942525,
      "grad_norm": 0.6315518021583557,
      "learning_rate": 0.0004542764788277125,
      "loss": 3.5446,
      "step": 22650
    },
    {
      "epoch": 2.443224625982133,
      "grad_norm": 0.5915260910987854,
      "learning_rate": 0.00045395323779765103,
      "loss": 3.5475,
      "step": 22700
    },
    {
      "epoch": 2.4486061780217416,
      "grad_norm": 0.6051799058914185,
      "learning_rate": 0.0004536299967675897,
      "loss": 3.5626,
      "step": 22750
    },
    {
      "epoch": 2.4539877300613497,
      "grad_norm": 0.5841783881187439,
      "learning_rate": 0.0004533067557375283,
      "loss": 3.5664,
      "step": 22800
    },
    {
      "epoch": 2.459369282100958,
      "grad_norm": 0.597858726978302,
      "learning_rate": 0.0004529835147074668,
      "loss": 3.5616,
      "step": 22850
    },
    {
      "epoch": 2.464750834140566,
      "grad_norm": 0.6230478882789612,
      "learning_rate": 0.0004526602736774054,
      "loss": 3.5564,
      "step": 22900
    },
    {
      "epoch": 2.4701323861801745,
      "grad_norm": 0.6594014167785645,
      "learning_rate": 0.00045233703264734395,
      "loss": 3.5613,
      "step": 22950
    },
    {
      "epoch": 2.4755139382197826,
      "grad_norm": 0.6744344830513,
      "learning_rate": 0.00045201379161728255,
      "loss": 3.5627,
      "step": 23000
    },
    {
      "epoch": 2.4755139382197826,
      "eval_accuracy": 0.3652093018606481,
      "eval_loss": 3.546725034713745,
      "eval_runtime": 184.3568,
      "eval_samples_per_second": 97.696,
      "eval_steps_per_second": 6.108,
      "step": 23000
    },
    {
      "epoch": 2.4808954902593907,
      "grad_norm": 0.5974053740501404,
      "learning_rate": 0.0004516905505872212,
      "loss": 3.5591,
      "step": 23050
    },
    {
      "epoch": 2.4862770422989993,
      "grad_norm": 0.5970222353935242,
      "learning_rate": 0.00045136730955715973,
      "loss": 3.5434,
      "step": 23100
    },
    {
      "epoch": 2.4916585943386074,
      "grad_norm": 0.6050614714622498,
      "learning_rate": 0.00045104406852709833,
      "loss": 3.5672,
      "step": 23150
    },
    {
      "epoch": 2.4970401463782155,
      "grad_norm": 0.6158961653709412,
      "learning_rate": 0.0004507208274970369,
      "loss": 3.544,
      "step": 23200
    },
    {
      "epoch": 2.5024216984178236,
      "grad_norm": 1.0708165168762207,
      "learning_rate": 0.00045039758646697546,
      "loss": 3.5555,
      "step": 23250
    },
    {
      "epoch": 2.5078032504574317,
      "grad_norm": 0.6259361505508423,
      "learning_rate": 0.0004500743454369141,
      "loss": 3.5452,
      "step": 23300
    },
    {
      "epoch": 2.5131848024970402,
      "grad_norm": 0.5771409273147583,
      "learning_rate": 0.0004497511044068527,
      "loss": 3.5669,
      "step": 23350
    },
    {
      "epoch": 2.5185663545366483,
      "grad_norm": 0.6256680488586426,
      "learning_rate": 0.00044942786337679125,
      "loss": 3.5385,
      "step": 23400
    },
    {
      "epoch": 2.5239479065762565,
      "grad_norm": 0.5991699695587158,
      "learning_rate": 0.00044910462234672984,
      "loss": 3.5458,
      "step": 23450
    },
    {
      "epoch": 2.529329458615865,
      "grad_norm": 0.6118124127388,
      "learning_rate": 0.0004487813813166684,
      "loss": 3.5564,
      "step": 23500
    },
    {
      "epoch": 2.534711010655473,
      "grad_norm": 0.5802284479141235,
      "learning_rate": 0.000448458140286607,
      "loss": 3.5391,
      "step": 23550
    },
    {
      "epoch": 2.540092562695081,
      "grad_norm": 0.5789219737052917,
      "learning_rate": 0.00044813489925654563,
      "loss": 3.5725,
      "step": 23600
    },
    {
      "epoch": 2.5454741147346893,
      "grad_norm": 0.5526463985443115,
      "learning_rate": 0.00044781165822648417,
      "loss": 3.5537,
      "step": 23650
    },
    {
      "epoch": 2.550855666774298,
      "grad_norm": 0.6136961579322815,
      "learning_rate": 0.00044748841719642276,
      "loss": 3.5335,
      "step": 23700
    },
    {
      "epoch": 2.556237218813906,
      "grad_norm": 0.5691003799438477,
      "learning_rate": 0.00044716517616636136,
      "loss": 3.5513,
      "step": 23750
    },
    {
      "epoch": 2.561618770853514,
      "grad_norm": 0.6383370161056519,
      "learning_rate": 0.0004468419351362999,
      "loss": 3.5548,
      "step": 23800
    },
    {
      "epoch": 2.567000322893122,
      "grad_norm": 0.5766544938087463,
      "learning_rate": 0.0004465186941062385,
      "loss": 3.5637,
      "step": 23850
    },
    {
      "epoch": 2.5723818749327307,
      "grad_norm": 0.5658783912658691,
      "learning_rate": 0.00044619545307617714,
      "loss": 3.5389,
      "step": 23900
    },
    {
      "epoch": 2.577763426972339,
      "grad_norm": 0.5434797406196594,
      "learning_rate": 0.0004458722120461157,
      "loss": 3.5686,
      "step": 23950
    },
    {
      "epoch": 2.583144979011947,
      "grad_norm": 0.6266382932662964,
      "learning_rate": 0.0004455489710160543,
      "loss": 3.5468,
      "step": 24000
    },
    {
      "epoch": 2.583144979011947,
      "eval_accuracy": 0.36643860063338113,
      "eval_loss": 3.5350229740142822,
      "eval_runtime": 184.2805,
      "eval_samples_per_second": 97.737,
      "eval_steps_per_second": 6.11,
      "step": 24000
    },
    {
      "epoch": 2.5885265310515555,
      "grad_norm": 0.589867353439331,
      "learning_rate": 0.0004452257299859928,
      "loss": 3.5555,
      "step": 24050
    },
    {
      "epoch": 2.5939080830911636,
      "grad_norm": 0.5716838836669922,
      "learning_rate": 0.0004449024889559314,
      "loss": 3.5394,
      "step": 24100
    },
    {
      "epoch": 2.5992896351307717,
      "grad_norm": 0.614301860332489,
      "learning_rate": 0.00044457924792587,
      "loss": 3.5539,
      "step": 24150
    },
    {
      "epoch": 2.60467118717038,
      "grad_norm": 0.5801094174385071,
      "learning_rate": 0.0004442560068958086,
      "loss": 3.5685,
      "step": 24200
    },
    {
      "epoch": 2.610052739209988,
      "grad_norm": 0.6649515628814697,
      "learning_rate": 0.0004439327658657472,
      "loss": 3.5648,
      "step": 24250
    },
    {
      "epoch": 2.6154342912495965,
      "grad_norm": 0.5923357009887695,
      "learning_rate": 0.0004436095248356858,
      "loss": 3.5507,
      "step": 24300
    },
    {
      "epoch": 2.6208158432892046,
      "grad_norm": 0.5791347026824951,
      "learning_rate": 0.00044328628380562433,
      "loss": 3.5521,
      "step": 24350
    },
    {
      "epoch": 2.6261973953288127,
      "grad_norm": 0.6230629086494446,
      "learning_rate": 0.0004429630427755629,
      "loss": 3.5347,
      "step": 24400
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 0.7393829226493835,
      "learning_rate": 0.0004426398017455016,
      "loss": 3.5564,
      "step": 24450
    },
    {
      "epoch": 2.6369604994080293,
      "grad_norm": 0.6273407936096191,
      "learning_rate": 0.0004423165607154401,
      "loss": 3.5454,
      "step": 24500
    },
    {
      "epoch": 2.6423420514476375,
      "grad_norm": 0.6728057265281677,
      "learning_rate": 0.0004419933196853787,
      "loss": 3.5278,
      "step": 24550
    },
    {
      "epoch": 2.6477236034872456,
      "grad_norm": 0.6352829933166504,
      "learning_rate": 0.0004416765434759185,
      "loss": 3.5734,
      "step": 24600
    },
    {
      "epoch": 2.653105155526854,
      "grad_norm": 0.5855279564857483,
      "learning_rate": 0.0004413533024458571,
      "loss": 3.5301,
      "step": 24650
    },
    {
      "epoch": 2.658486707566462,
      "grad_norm": 0.6078464388847351,
      "learning_rate": 0.00044103006141579565,
      "loss": 3.5344,
      "step": 24700
    },
    {
      "epoch": 2.6638682596060703,
      "grad_norm": 0.6252312660217285,
      "learning_rate": 0.00044070682038573425,
      "loss": 3.5497,
      "step": 24750
    },
    {
      "epoch": 2.6692498116456784,
      "grad_norm": 0.5892012715339661,
      "learning_rate": 0.0004403835793556728,
      "loss": 3.5421,
      "step": 24800
    },
    {
      "epoch": 2.674631363685287,
      "grad_norm": 0.6075966358184814,
      "learning_rate": 0.00044006033832561143,
      "loss": 3.5565,
      "step": 24850
    },
    {
      "epoch": 2.680012915724895,
      "grad_norm": 0.6388818621635437,
      "learning_rate": 0.00043973709729555003,
      "loss": 3.5469,
      "step": 24900
    },
    {
      "epoch": 2.685394467764503,
      "grad_norm": 0.5894478559494019,
      "learning_rate": 0.00043941385626548857,
      "loss": 3.548,
      "step": 24950
    },
    {
      "epoch": 2.6907760198041117,
      "grad_norm": 0.6302310228347778,
      "learning_rate": 0.00043909708005602843,
      "loss": 3.5419,
      "step": 25000
    },
    {
      "epoch": 2.6907760198041117,
      "eval_accuracy": 0.36701924166586175,
      "eval_loss": 3.529244899749756,
      "eval_runtime": 184.5117,
      "eval_samples_per_second": 97.614,
      "eval_steps_per_second": 6.103,
      "step": 25000
    },
    {
      "epoch": 2.69615757184372,
      "grad_norm": 0.6369917392730713,
      "learning_rate": 0.00043877383902596697,
      "loss": 3.5538,
      "step": 25050
    },
    {
      "epoch": 2.701539123883328,
      "grad_norm": 0.6358683705329895,
      "learning_rate": 0.00043845059799590557,
      "loss": 3.5417,
      "step": 25100
    },
    {
      "epoch": 2.706920675922936,
      "grad_norm": 0.5954503417015076,
      "learning_rate": 0.00043812735696584416,
      "loss": 3.5361,
      "step": 25150
    },
    {
      "epoch": 2.712302227962544,
      "grad_norm": 0.7064067125320435,
      "learning_rate": 0.00043780411593578275,
      "loss": 3.5368,
      "step": 25200
    },
    {
      "epoch": 2.7176837800021527,
      "grad_norm": 0.6246035099029541,
      "learning_rate": 0.00043748087490572135,
      "loss": 3.5622,
      "step": 25250
    },
    {
      "epoch": 2.723065332041761,
      "grad_norm": 0.6261956095695496,
      "learning_rate": 0.00043715763387565994,
      "loss": 3.542,
      "step": 25300
    },
    {
      "epoch": 2.728446884081369,
      "grad_norm": 0.6844997406005859,
      "learning_rate": 0.0004368343928455985,
      "loss": 3.5278,
      "step": 25350
    },
    {
      "epoch": 2.7338284361209775,
      "grad_norm": 0.6102064847946167,
      "learning_rate": 0.0004365111518155371,
      "loss": 3.5448,
      "step": 25400
    },
    {
      "epoch": 2.7392099881605856,
      "grad_norm": 0.5947371125221252,
      "learning_rate": 0.0004361879107854756,
      "loss": 3.5593,
      "step": 25450
    },
    {
      "epoch": 2.7445915402001937,
      "grad_norm": 0.6366443037986755,
      "learning_rate": 0.00043586466975541427,
      "loss": 3.5478,
      "step": 25500
    },
    {
      "epoch": 2.749973092239802,
      "grad_norm": 0.6193564534187317,
      "learning_rate": 0.00043554142872535286,
      "loss": 3.5407,
      "step": 25550
    },
    {
      "epoch": 2.7553546442794103,
      "grad_norm": 0.6757341623306274,
      "learning_rate": 0.0004352181876952914,
      "loss": 3.544,
      "step": 25600
    },
    {
      "epoch": 2.7607361963190185,
      "grad_norm": 0.6367349624633789,
      "learning_rate": 0.00043489494666523,
      "loss": 3.5596,
      "step": 25650
    },
    {
      "epoch": 2.7661177483586266,
      "grad_norm": 0.5869446992874146,
      "learning_rate": 0.00043457170563516854,
      "loss": 3.5491,
      "step": 25700
    },
    {
      "epoch": 2.7714993003982347,
      "grad_norm": 0.629715621471405,
      "learning_rate": 0.00043424846460510713,
      "loss": 3.5429,
      "step": 25750
    },
    {
      "epoch": 2.776880852437843,
      "grad_norm": 0.6242320537567139,
      "learning_rate": 0.0004339252235750458,
      "loss": 3.5401,
      "step": 25800
    },
    {
      "epoch": 2.7822624044774513,
      "grad_norm": 0.5985417366027832,
      "learning_rate": 0.0004336019825449843,
      "loss": 3.5372,
      "step": 25850
    },
    {
      "epoch": 2.7876439565170594,
      "grad_norm": 0.6011677980422974,
      "learning_rate": 0.0004332787415149229,
      "loss": 3.5463,
      "step": 25900
    },
    {
      "epoch": 2.793025508556668,
      "grad_norm": 0.6060757040977478,
      "learning_rate": 0.0004329555004848615,
      "loss": 3.5339,
      "step": 25950
    },
    {
      "epoch": 2.798407060596276,
      "grad_norm": 0.624178946018219,
      "learning_rate": 0.00043263225945480005,
      "loss": 3.5478,
      "step": 26000
    },
    {
      "epoch": 2.798407060596276,
      "eval_accuracy": 0.3681773727970763,
      "eval_loss": 3.51682186126709,
      "eval_runtime": 184.4015,
      "eval_samples_per_second": 97.673,
      "eval_steps_per_second": 6.106,
      "step": 26000
    },
    {
      "epoch": 2.803788612635884,
      "grad_norm": 0.5964826345443726,
      "learning_rate": 0.0004323090184247387,
      "loss": 3.5344,
      "step": 26050
    },
    {
      "epoch": 2.8091701646754923,
      "grad_norm": 0.6394344568252563,
      "learning_rate": 0.0004319857773946773,
      "loss": 3.5364,
      "step": 26100
    },
    {
      "epoch": 2.8145517167151004,
      "grad_norm": 0.5882453918457031,
      "learning_rate": 0.00043166253636461584,
      "loss": 3.5581,
      "step": 26150
    },
    {
      "epoch": 2.819933268754709,
      "grad_norm": 0.6245861053466797,
      "learning_rate": 0.00043133929533455443,
      "loss": 3.5436,
      "step": 26200
    },
    {
      "epoch": 2.825314820794317,
      "grad_norm": 0.583153247833252,
      "learning_rate": 0.00043101605430449297,
      "loss": 3.5279,
      "step": 26250
    },
    {
      "epoch": 2.830696372833925,
      "grad_norm": 0.5886926054954529,
      "learning_rate": 0.00043069281327443157,
      "loss": 3.5384,
      "step": 26300
    },
    {
      "epoch": 2.8360779248735337,
      "grad_norm": 0.6126932501792908,
      "learning_rate": 0.0004303695722443702,
      "loss": 3.5414,
      "step": 26350
    },
    {
      "epoch": 2.841459476913142,
      "grad_norm": 0.6343358755111694,
      "learning_rate": 0.00043004633121430876,
      "loss": 3.5225,
      "step": 26400
    },
    {
      "epoch": 2.84684102895275,
      "grad_norm": 0.6595734357833862,
      "learning_rate": 0.00042972309018424735,
      "loss": 3.5383,
      "step": 26450
    },
    {
      "epoch": 2.852222580992358,
      "grad_norm": 0.5940672159194946,
      "learning_rate": 0.00042939984915418594,
      "loss": 3.5451,
      "step": 26500
    },
    {
      "epoch": 2.857604133031966,
      "grad_norm": 0.5987943410873413,
      "learning_rate": 0.0004290766081241245,
      "loss": 3.5264,
      "step": 26550
    },
    {
      "epoch": 2.8629856850715747,
      "grad_norm": 0.6584435701370239,
      "learning_rate": 0.0004287533670940631,
      "loss": 3.5501,
      "step": 26600
    },
    {
      "epoch": 2.868367237111183,
      "grad_norm": 0.5722735524177551,
      "learning_rate": 0.00042843012606400173,
      "loss": 3.5263,
      "step": 26650
    },
    {
      "epoch": 2.873748789150791,
      "grad_norm": 0.6471678018569946,
      "learning_rate": 0.00042810688503394027,
      "loss": 3.5315,
      "step": 26700
    },
    {
      "epoch": 2.8791303411903995,
      "grad_norm": 0.6302046775817871,
      "learning_rate": 0.00042778364400387886,
      "loss": 3.5396,
      "step": 26750
    },
    {
      "epoch": 2.8845118932300076,
      "grad_norm": 0.6113825440406799,
      "learning_rate": 0.0004274604029738174,
      "loss": 3.5223,
      "step": 26800
    },
    {
      "epoch": 2.8898934452696157,
      "grad_norm": 0.5867438316345215,
      "learning_rate": 0.000427137161943756,
      "loss": 3.544,
      "step": 26850
    },
    {
      "epoch": 2.895274997309224,
      "grad_norm": 0.5938653349876404,
      "learning_rate": 0.00042681392091369465,
      "loss": 3.544,
      "step": 26900
    },
    {
      "epoch": 2.9006565493488323,
      "grad_norm": 0.5596200227737427,
      "learning_rate": 0.0004264906798836332,
      "loss": 3.5352,
      "step": 26950
    },
    {
      "epoch": 2.9060381013884404,
      "grad_norm": 0.5883537530899048,
      "learning_rate": 0.0004261674388535718,
      "loss": 3.5405,
      "step": 27000
    },
    {
      "epoch": 2.9060381013884404,
      "eval_accuracy": 0.36874953890430734,
      "eval_loss": 3.509413957595825,
      "eval_runtime": 184.4187,
      "eval_samples_per_second": 97.664,
      "eval_steps_per_second": 6.106,
      "step": 27000
    },
    {
      "epoch": 2.9114196534280485,
      "grad_norm": 0.5866814255714417,
      "learning_rate": 0.0004258441978235104,
      "loss": 3.5313,
      "step": 27050
    },
    {
      "epoch": 2.9168012054676566,
      "grad_norm": 0.6360952258110046,
      "learning_rate": 0.0004255209567934489,
      "loss": 3.5328,
      "step": 27100
    },
    {
      "epoch": 2.922182757507265,
      "grad_norm": 0.595851480960846,
      "learning_rate": 0.0004251977157633875,
      "loss": 3.5268,
      "step": 27150
    },
    {
      "epoch": 2.9275643095468733,
      "grad_norm": 0.6052762866020203,
      "learning_rate": 0.00042487447473332616,
      "loss": 3.555,
      "step": 27200
    },
    {
      "epoch": 2.9329458615864814,
      "grad_norm": 0.6358559131622314,
      "learning_rate": 0.0004245512337032647,
      "loss": 3.5173,
      "step": 27250
    },
    {
      "epoch": 2.93832741362609,
      "grad_norm": 0.5946629643440247,
      "learning_rate": 0.0004242279926732033,
      "loss": 3.5134,
      "step": 27300
    },
    {
      "epoch": 2.943708965665698,
      "grad_norm": 0.6441041827201843,
      "learning_rate": 0.00042390475164314184,
      "loss": 3.5411,
      "step": 27350
    },
    {
      "epoch": 2.949090517705306,
      "grad_norm": 0.6287612915039062,
      "learning_rate": 0.00042358151061308043,
      "loss": 3.5131,
      "step": 27400
    },
    {
      "epoch": 2.9544720697449143,
      "grad_norm": 0.5959669947624207,
      "learning_rate": 0.000423258269583019,
      "loss": 3.5199,
      "step": 27450
    },
    {
      "epoch": 2.9598536217845224,
      "grad_norm": 0.6317761540412903,
      "learning_rate": 0.0004229350285529576,
      "loss": 3.519,
      "step": 27500
    },
    {
      "epoch": 2.965235173824131,
      "grad_norm": 0.559985339641571,
      "learning_rate": 0.0004226117875228962,
      "loss": 3.5441,
      "step": 27550
    },
    {
      "epoch": 2.970616725863739,
      "grad_norm": 0.6218299269676208,
      "learning_rate": 0.0004222885464928348,
      "loss": 3.525,
      "step": 27600
    },
    {
      "epoch": 2.975998277903347,
      "grad_norm": 0.6129732131958008,
      "learning_rate": 0.00042196530546277335,
      "loss": 3.5182,
      "step": 27650
    },
    {
      "epoch": 2.9813798299429557,
      "grad_norm": 0.5594598650932312,
      "learning_rate": 0.00042164206443271195,
      "loss": 3.5524,
      "step": 27700
    },
    {
      "epoch": 2.986761381982564,
      "grad_norm": 0.560884952545166,
      "learning_rate": 0.0004213188234026505,
      "loss": 3.5301,
      "step": 27750
    },
    {
      "epoch": 2.992142934022172,
      "grad_norm": 0.5810580849647522,
      "learning_rate": 0.00042099558237258914,
      "loss": 3.5274,
      "step": 27800
    },
    {
      "epoch": 2.9975244860617805,
      "grad_norm": 0.6317058205604553,
      "learning_rate": 0.00042067234134252773,
      "loss": 3.5331,
      "step": 27850
    },
    {
      "epoch": 3.0029060381013886,
      "grad_norm": 0.7533642649650574,
      "learning_rate": 0.00042034910031246627,
      "loss": 3.4753,
      "step": 27900
    },
    {
      "epoch": 3.0082875901409967,
      "grad_norm": 0.6032314896583557,
      "learning_rate": 0.00042002585928240486,
      "loss": 3.4192,
      "step": 27950
    },
    {
      "epoch": 3.0136691421806048,
      "grad_norm": 0.5920121669769287,
      "learning_rate": 0.00041970261825234346,
      "loss": 3.4418,
      "step": 28000
    },
    {
      "epoch": 3.0136691421806048,
      "eval_accuracy": 0.3696549434184654,
      "eval_loss": 3.50471830368042,
      "eval_runtime": 184.4783,
      "eval_samples_per_second": 97.632,
      "eval_steps_per_second": 6.104,
      "step": 28000
    },
    {
      "epoch": 3.0190506942202133,
      "grad_norm": 0.5959385633468628,
      "learning_rate": 0.00041937937722228205,
      "loss": 3.4464,
      "step": 28050
    },
    {
      "epoch": 3.0244322462598214,
      "grad_norm": 0.6004893183708191,
      "learning_rate": 0.0004190626010128218,
      "loss": 3.4474,
      "step": 28100
    },
    {
      "epoch": 3.0298137982994295,
      "grad_norm": 0.5674557685852051,
      "learning_rate": 0.00041873935998276046,
      "loss": 3.4291,
      "step": 28150
    },
    {
      "epoch": 3.0351953503390376,
      "grad_norm": 0.6396381855010986,
      "learning_rate": 0.00041841611895269905,
      "loss": 3.4364,
      "step": 28200
    },
    {
      "epoch": 3.040576902378646,
      "grad_norm": 0.6117861270904541,
      "learning_rate": 0.0004180928779226376,
      "loss": 3.4426,
      "step": 28250
    },
    {
      "epoch": 3.0459584544182543,
      "grad_norm": 0.6411035656929016,
      "learning_rate": 0.0004177696368925762,
      "loss": 3.4297,
      "step": 28300
    },
    {
      "epoch": 3.0513400064578624,
      "grad_norm": 0.5998719930648804,
      "learning_rate": 0.0004174463958625148,
      "loss": 3.4609,
      "step": 28350
    },
    {
      "epoch": 3.0567215584974705,
      "grad_norm": 0.6109996438026428,
      "learning_rate": 0.0004171231548324533,
      "loss": 3.4492,
      "step": 28400
    },
    {
      "epoch": 3.062103110537079,
      "grad_norm": 0.6117191910743713,
      "learning_rate": 0.00041679991380239197,
      "loss": 3.4615,
      "step": 28450
    },
    {
      "epoch": 3.067484662576687,
      "grad_norm": 0.6076861619949341,
      "learning_rate": 0.00041647667277233056,
      "loss": 3.4629,
      "step": 28500
    },
    {
      "epoch": 3.0728662146162953,
      "grad_norm": 0.6061896681785583,
      "learning_rate": 0.0004161534317422691,
      "loss": 3.4575,
      "step": 28550
    },
    {
      "epoch": 3.0782477666559034,
      "grad_norm": 0.6356080174446106,
      "learning_rate": 0.0004158301907122077,
      "loss": 3.4633,
      "step": 28600
    },
    {
      "epoch": 3.083629318695512,
      "grad_norm": 0.6221888065338135,
      "learning_rate": 0.00041550694968214624,
      "loss": 3.4469,
      "step": 28650
    },
    {
      "epoch": 3.08901087073512,
      "grad_norm": 0.5875892043113708,
      "learning_rate": 0.0004151837086520849,
      "loss": 3.4509,
      "step": 28700
    },
    {
      "epoch": 3.094392422774728,
      "grad_norm": 0.6204771995544434,
      "learning_rate": 0.0004148604676220235,
      "loss": 3.4541,
      "step": 28750
    },
    {
      "epoch": 3.0997739748143363,
      "grad_norm": 0.6354315876960754,
      "learning_rate": 0.000414537226591962,
      "loss": 3.4589,
      "step": 28800
    },
    {
      "epoch": 3.105155526853945,
      "grad_norm": 0.6152920126914978,
      "learning_rate": 0.0004142139855619006,
      "loss": 3.4623,
      "step": 28850
    },
    {
      "epoch": 3.110537078893553,
      "grad_norm": 0.6353855133056641,
      "learning_rate": 0.0004138907445318392,
      "loss": 3.4665,
      "step": 28900
    },
    {
      "epoch": 3.115918630933161,
      "grad_norm": 0.6515467166900635,
      "learning_rate": 0.00041356750350177775,
      "loss": 3.4618,
      "step": 28950
    },
    {
      "epoch": 3.121300182972769,
      "grad_norm": 0.6271316409111023,
      "learning_rate": 0.0004132442624717164,
      "loss": 3.4607,
      "step": 29000
    },
    {
      "epoch": 3.121300182972769,
      "eval_accuracy": 0.37060891577347654,
      "eval_loss": 3.499715566635132,
      "eval_runtime": 184.1888,
      "eval_samples_per_second": 97.786,
      "eval_steps_per_second": 6.113,
      "step": 29000
    },
    {
      "epoch": 3.1266817350123777,
      "grad_norm": 0.6345009207725525,
      "learning_rate": 0.000412921021441655,
      "loss": 3.4538,
      "step": 29050
    },
    {
      "epoch": 3.132063287051986,
      "grad_norm": 0.5966889262199402,
      "learning_rate": 0.00041259778041159354,
      "loss": 3.4676,
      "step": 29100
    },
    {
      "epoch": 3.137444839091594,
      "grad_norm": 0.6432265639305115,
      "learning_rate": 0.00041227453938153213,
      "loss": 3.4613,
      "step": 29150
    },
    {
      "epoch": 3.1428263911312024,
      "grad_norm": 0.6030633449554443,
      "learning_rate": 0.00041195129835147067,
      "loss": 3.443,
      "step": 29200
    },
    {
      "epoch": 3.1482079431708105,
      "grad_norm": 0.6159685850143433,
      "learning_rate": 0.00041162805732140927,
      "loss": 3.4412,
      "step": 29250
    },
    {
      "epoch": 3.1535894952104186,
      "grad_norm": 0.6083833575248718,
      "learning_rate": 0.0004113048162913479,
      "loss": 3.4484,
      "step": 29300
    },
    {
      "epoch": 3.1589710472500268,
      "grad_norm": 0.614616870880127,
      "learning_rate": 0.00041098157526128646,
      "loss": 3.4547,
      "step": 29350
    },
    {
      "epoch": 3.1643525992896353,
      "grad_norm": 0.5777009129524231,
      "learning_rate": 0.00041065833423122505,
      "loss": 3.4604,
      "step": 29400
    },
    {
      "epoch": 3.1697341513292434,
      "grad_norm": 0.6041639447212219,
      "learning_rate": 0.00041033509320116365,
      "loss": 3.4548,
      "step": 29450
    },
    {
      "epoch": 3.1751157033688515,
      "grad_norm": 0.661534309387207,
      "learning_rate": 0.0004100118521711022,
      "loss": 3.4482,
      "step": 29500
    },
    {
      "epoch": 3.1804972554084596,
      "grad_norm": 0.598143458366394,
      "learning_rate": 0.0004096886111410408,
      "loss": 3.4579,
      "step": 29550
    },
    {
      "epoch": 3.185878807448068,
      "grad_norm": 0.6011090874671936,
      "learning_rate": 0.00040936537011097943,
      "loss": 3.444,
      "step": 29600
    },
    {
      "epoch": 3.1912603594876763,
      "grad_norm": 0.603949785232544,
      "learning_rate": 0.00040904212908091797,
      "loss": 3.452,
      "step": 29650
    },
    {
      "epoch": 3.1966419115272844,
      "grad_norm": 0.6517125964164734,
      "learning_rate": 0.00040871888805085656,
      "loss": 3.457,
      "step": 29700
    },
    {
      "epoch": 3.2020234635668925,
      "grad_norm": 0.6115108132362366,
      "learning_rate": 0.0004083956470207951,
      "loss": 3.4672,
      "step": 29750
    },
    {
      "epoch": 3.207405015606501,
      "grad_norm": 0.5897082686424255,
      "learning_rate": 0.0004080724059907337,
      "loss": 3.4668,
      "step": 29800
    },
    {
      "epoch": 3.212786567646109,
      "grad_norm": 0.6561703681945801,
      "learning_rate": 0.00040774916496067235,
      "loss": 3.4486,
      "step": 29850
    },
    {
      "epoch": 3.2181681196857173,
      "grad_norm": 0.63785320520401,
      "learning_rate": 0.0004074259239306109,
      "loss": 3.4727,
      "step": 29900
    },
    {
      "epoch": 3.2235496717253254,
      "grad_norm": 0.6829721927642822,
      "learning_rate": 0.0004071026829005495,
      "loss": 3.4702,
      "step": 29950
    },
    {
      "epoch": 3.228931223764934,
      "grad_norm": 0.6836195588111877,
      "learning_rate": 0.0004067794418704881,
      "loss": 3.4689,
      "step": 30000
    },
    {
      "epoch": 3.228931223764934,
      "eval_accuracy": 0.37078525941039947,
      "eval_loss": 3.495083808898926,
      "eval_runtime": 184.5676,
      "eval_samples_per_second": 97.585,
      "eval_steps_per_second": 6.101,
      "step": 30000
    },
    {
      "epoch": 3.234312775804542,
      "grad_norm": 0.6209216117858887,
      "learning_rate": 0.0004064562008404266,
      "loss": 3.4579,
      "step": 30050
    },
    {
      "epoch": 3.23969432784415,
      "grad_norm": 0.5718951225280762,
      "learning_rate": 0.0004061329598103652,
      "loss": 3.4574,
      "step": 30100
    },
    {
      "epoch": 3.2450758798837587,
      "grad_norm": 0.5745283961296082,
      "learning_rate": 0.00040580971878030386,
      "loss": 3.4564,
      "step": 30150
    },
    {
      "epoch": 3.250457431923367,
      "grad_norm": 0.6343955993652344,
      "learning_rate": 0.0004054864777502424,
      "loss": 3.4437,
      "step": 30200
    },
    {
      "epoch": 3.255838983962975,
      "grad_norm": 0.6511952877044678,
      "learning_rate": 0.000405163236720181,
      "loss": 3.463,
      "step": 30250
    },
    {
      "epoch": 3.261220536002583,
      "grad_norm": 0.5887660980224609,
      "learning_rate": 0.00040483999569011954,
      "loss": 3.4635,
      "step": 30300
    },
    {
      "epoch": 3.2666020880421915,
      "grad_norm": 0.6006124019622803,
      "learning_rate": 0.00040451675466005813,
      "loss": 3.4757,
      "step": 30350
    },
    {
      "epoch": 3.2719836400817996,
      "grad_norm": 0.6887456178665161,
      "learning_rate": 0.0004041935136299967,
      "loss": 3.4778,
      "step": 30400
    },
    {
      "epoch": 3.2773651921214078,
      "grad_norm": 0.6489464640617371,
      "learning_rate": 0.0004038702725999353,
      "loss": 3.4676,
      "step": 30450
    },
    {
      "epoch": 3.282746744161016,
      "grad_norm": 0.6444425582885742,
      "learning_rate": 0.0004035470315698739,
      "loss": 3.4297,
      "step": 30500
    },
    {
      "epoch": 3.2881282962006244,
      "grad_norm": 0.648106038570404,
      "learning_rate": 0.0004032237905398125,
      "loss": 3.4597,
      "step": 30550
    },
    {
      "epoch": 3.2935098482402325,
      "grad_norm": 0.5875924825668335,
      "learning_rate": 0.00040290054950975105,
      "loss": 3.4601,
      "step": 30600
    },
    {
      "epoch": 3.2988914002798406,
      "grad_norm": 0.639473557472229,
      "learning_rate": 0.00040257730847968965,
      "loss": 3.4641,
      "step": 30650
    },
    {
      "epoch": 3.304272952319449,
      "grad_norm": 0.5909704566001892,
      "learning_rate": 0.0004022540674496283,
      "loss": 3.4688,
      "step": 30700
    },
    {
      "epoch": 3.3096545043590573,
      "grad_norm": 0.6033146977424622,
      "learning_rate": 0.00040193082641956684,
      "loss": 3.4594,
      "step": 30750
    },
    {
      "epoch": 3.3150360563986654,
      "grad_norm": 0.6792746186256409,
      "learning_rate": 0.00040160758538950543,
      "loss": 3.4726,
      "step": 30800
    },
    {
      "epoch": 3.3204176084382735,
      "grad_norm": 0.5983012318611145,
      "learning_rate": 0.00040128434435944397,
      "loss": 3.4517,
      "step": 30850
    },
    {
      "epoch": 3.3257991604778816,
      "grad_norm": 0.6491327881813049,
      "learning_rate": 0.00040096110332938257,
      "loss": 3.4607,
      "step": 30900
    },
    {
      "epoch": 3.33118071251749,
      "grad_norm": 0.5958024859428406,
      "learning_rate": 0.00040063786229932116,
      "loss": 3.4651,
      "step": 30950
    },
    {
      "epoch": 3.3365622645570983,
      "grad_norm": 0.6234532594680786,
      "learning_rate": 0.00040031462126925975,
      "loss": 3.4682,
      "step": 31000
    },
    {
      "epoch": 3.3365622645570983,
      "eval_accuracy": 0.371365357178441,
      "eval_loss": 3.4893696308135986,
      "eval_runtime": 184.582,
      "eval_samples_per_second": 97.577,
      "eval_steps_per_second": 6.1,
      "step": 31000
    },
    {
      "epoch": 3.3419438165967064,
      "grad_norm": 0.6505281925201416,
      "learning_rate": 0.00039999138023919835,
      "loss": 3.4498,
      "step": 31050
    },
    {
      "epoch": 3.347325368636315,
      "grad_norm": 0.6393039226531982,
      "learning_rate": 0.00039966813920913694,
      "loss": 3.4809,
      "step": 31100
    },
    {
      "epoch": 3.352706920675923,
      "grad_norm": 0.557396650314331,
      "learning_rate": 0.0003993448981790755,
      "loss": 3.4634,
      "step": 31150
    },
    {
      "epoch": 3.358088472715531,
      "grad_norm": 0.6200848817825317,
      "learning_rate": 0.0003990216571490141,
      "loss": 3.4779,
      "step": 31200
    },
    {
      "epoch": 3.3634700247551392,
      "grad_norm": 0.5961986780166626,
      "learning_rate": 0.0003986984161189526,
      "loss": 3.4608,
      "step": 31250
    },
    {
      "epoch": 3.368851576794748,
      "grad_norm": 0.6536951065063477,
      "learning_rate": 0.00039837517508889127,
      "loss": 3.4467,
      "step": 31300
    },
    {
      "epoch": 3.374233128834356,
      "grad_norm": 0.6691170930862427,
      "learning_rate": 0.00039805193405882986,
      "loss": 3.4742,
      "step": 31350
    },
    {
      "epoch": 3.379614680873964,
      "grad_norm": 0.663662314414978,
      "learning_rate": 0.0003977286930287684,
      "loss": 3.471,
      "step": 31400
    },
    {
      "epoch": 3.384996232913572,
      "grad_norm": 0.581847071647644,
      "learning_rate": 0.000397405451998707,
      "loss": 3.4687,
      "step": 31450
    },
    {
      "epoch": 3.3903777849531807,
      "grad_norm": 0.6205992102622986,
      "learning_rate": 0.00039708221096864554,
      "loss": 3.4634,
      "step": 31500
    },
    {
      "epoch": 3.3957593369927888,
      "grad_norm": 0.6378187537193298,
      "learning_rate": 0.00039675896993858413,
      "loss": 3.4829,
      "step": 31550
    },
    {
      "epoch": 3.401140889032397,
      "grad_norm": 0.5727849006652832,
      "learning_rate": 0.0003964357289085228,
      "loss": 3.4656,
      "step": 31600
    },
    {
      "epoch": 3.4065224410720054,
      "grad_norm": 0.5948082208633423,
      "learning_rate": 0.0003961124878784613,
      "loss": 3.4606,
      "step": 31650
    },
    {
      "epoch": 3.4119039931116135,
      "grad_norm": 0.5754572153091431,
      "learning_rate": 0.0003957892468483999,
      "loss": 3.4664,
      "step": 31700
    },
    {
      "epoch": 3.4172855451512216,
      "grad_norm": 0.6806990504264832,
      "learning_rate": 0.0003954660058183385,
      "loss": 3.469,
      "step": 31750
    },
    {
      "epoch": 3.4226670971908297,
      "grad_norm": 0.6133432984352112,
      "learning_rate": 0.00039514276478827705,
      "loss": 3.4888,
      "step": 31800
    },
    {
      "epoch": 3.428048649230438,
      "grad_norm": 0.644431471824646,
      "learning_rate": 0.0003948195237582157,
      "loss": 3.4489,
      "step": 31850
    },
    {
      "epoch": 3.4334302012700464,
      "grad_norm": 0.7031443119049072,
      "learning_rate": 0.0003944962827281543,
      "loss": 3.4525,
      "step": 31900
    },
    {
      "epoch": 3.4388117533096545,
      "grad_norm": 0.636039137840271,
      "learning_rate": 0.00039417304169809284,
      "loss": 3.4793,
      "step": 31950
    },
    {
      "epoch": 3.4441933053492626,
      "grad_norm": 0.623345136642456,
      "learning_rate": 0.00039384980066803143,
      "loss": 3.4731,
      "step": 32000
    },
    {
      "epoch": 3.4441933053492626,
      "eval_accuracy": 0.3724262439750616,
      "eval_loss": 3.481135845184326,
      "eval_runtime": 184.4886,
      "eval_samples_per_second": 97.627,
      "eval_steps_per_second": 6.103,
      "step": 32000
    },
    {
      "epoch": 3.449574857388871,
      "grad_norm": 0.6249350309371948,
      "learning_rate": 0.00039352655963796997,
      "loss": 3.4612,
      "step": 32050
    },
    {
      "epoch": 3.4549564094284793,
      "grad_norm": 0.5687639713287354,
      "learning_rate": 0.00039320978342850983,
      "loss": 3.4801,
      "step": 32100
    },
    {
      "epoch": 3.4603379614680874,
      "grad_norm": 0.6019929647445679,
      "learning_rate": 0.00039288654239844837,
      "loss": 3.4634,
      "step": 32150
    },
    {
      "epoch": 3.4657195135076955,
      "grad_norm": 0.6131715774536133,
      "learning_rate": 0.00039256330136838697,
      "loss": 3.4576,
      "step": 32200
    },
    {
      "epoch": 3.471101065547304,
      "grad_norm": 0.6335099935531616,
      "learning_rate": 0.0003922400603383256,
      "loss": 3.4726,
      "step": 32250
    },
    {
      "epoch": 3.476482617586912,
      "grad_norm": 0.6393163204193115,
      "learning_rate": 0.00039191681930826416,
      "loss": 3.4761,
      "step": 32300
    },
    {
      "epoch": 3.4818641696265202,
      "grad_norm": 0.6179066300392151,
      "learning_rate": 0.00039159357827820275,
      "loss": 3.4646,
      "step": 32350
    },
    {
      "epoch": 3.4872457216661283,
      "grad_norm": 0.6377819776535034,
      "learning_rate": 0.00039127033724814135,
      "loss": 3.4636,
      "step": 32400
    },
    {
      "epoch": 3.492627273705737,
      "grad_norm": 0.5949285626411438,
      "learning_rate": 0.0003909470962180799,
      "loss": 3.4832,
      "step": 32450
    },
    {
      "epoch": 3.498008825745345,
      "grad_norm": 0.6022635102272034,
      "learning_rate": 0.00039062385518801854,
      "loss": 3.4621,
      "step": 32500
    },
    {
      "epoch": 3.503390377784953,
      "grad_norm": 0.6100685000419617,
      "learning_rate": 0.00039030061415795713,
      "loss": 3.4737,
      "step": 32550
    },
    {
      "epoch": 3.5087719298245617,
      "grad_norm": 0.654662013053894,
      "learning_rate": 0.00038997737312789567,
      "loss": 3.4695,
      "step": 32600
    },
    {
      "epoch": 3.5141534818641698,
      "grad_norm": 0.6307412385940552,
      "learning_rate": 0.00038965413209783426,
      "loss": 3.452,
      "step": 32650
    },
    {
      "epoch": 3.519535033903778,
      "grad_norm": 0.6365349292755127,
      "learning_rate": 0.0003893308910677728,
      "loss": 3.4632,
      "step": 32700
    },
    {
      "epoch": 3.524916585943386,
      "grad_norm": 0.6419442296028137,
      "learning_rate": 0.0003890076500377114,
      "loss": 3.4537,
      "step": 32750
    },
    {
      "epoch": 3.530298137982994,
      "grad_norm": 0.6395874619483948,
      "learning_rate": 0.00038868440900765005,
      "loss": 3.4574,
      "step": 32800
    },
    {
      "epoch": 3.5356796900226026,
      "grad_norm": 0.6256914138793945,
      "learning_rate": 0.0003883611679775886,
      "loss": 3.4665,
      "step": 32850
    },
    {
      "epoch": 3.5410612420622107,
      "grad_norm": 0.615369975566864,
      "learning_rate": 0.0003880379269475272,
      "loss": 3.4773,
      "step": 32900
    },
    {
      "epoch": 3.546442794101819,
      "grad_norm": 0.5932053327560425,
      "learning_rate": 0.0003877146859174657,
      "loss": 3.4726,
      "step": 32950
    },
    {
      "epoch": 3.5518243461414274,
      "grad_norm": 0.5968548059463501,
      "learning_rate": 0.0003873914448874043,
      "loss": 3.4811,
      "step": 33000
    },
    {
      "epoch": 3.5518243461414274,
      "eval_accuracy": 0.37287335060841814,
      "eval_loss": 3.474228620529175,
      "eval_runtime": 184.1747,
      "eval_samples_per_second": 97.793,
      "eval_steps_per_second": 6.114,
      "step": 33000
    },
    {
      "epoch": 3.5572058981810355,
      "grad_norm": 0.6117073893547058,
      "learning_rate": 0.0003870682038573429,
      "loss": 3.4545,
      "step": 33050
    },
    {
      "epoch": 3.5625874502206436,
      "grad_norm": 0.6557143926620483,
      "learning_rate": 0.0003867449628272815,
      "loss": 3.4441,
      "step": 33100
    },
    {
      "epoch": 3.5679690022602517,
      "grad_norm": 0.627402126789093,
      "learning_rate": 0.0003864217217972201,
      "loss": 3.4705,
      "step": 33150
    },
    {
      "epoch": 3.57335055429986,
      "grad_norm": 0.6005900502204895,
      "learning_rate": 0.0003860984807671587,
      "loss": 3.4812,
      "step": 33200
    },
    {
      "epoch": 3.5787321063394684,
      "grad_norm": 0.7192645072937012,
      "learning_rate": 0.00038577523973709724,
      "loss": 3.4744,
      "step": 33250
    },
    {
      "epoch": 3.5841136583790765,
      "grad_norm": 0.671812117099762,
      "learning_rate": 0.00038545199870703583,
      "loss": 3.4575,
      "step": 33300
    },
    {
      "epoch": 3.5894952104186846,
      "grad_norm": 0.6611122488975525,
      "learning_rate": 0.0003851287576769744,
      "loss": 3.4651,
      "step": 33350
    },
    {
      "epoch": 3.594876762458293,
      "grad_norm": 0.6967386603355408,
      "learning_rate": 0.000384805516646913,
      "loss": 3.4509,
      "step": 33400
    },
    {
      "epoch": 3.6002583144979012,
      "grad_norm": 0.6024479269981384,
      "learning_rate": 0.0003844822756168516,
      "loss": 3.4753,
      "step": 33450
    },
    {
      "epoch": 3.6056398665375093,
      "grad_norm": 0.6339122653007507,
      "learning_rate": 0.00038415903458679016,
      "loss": 3.4743,
      "step": 33500
    },
    {
      "epoch": 3.611021418577118,
      "grad_norm": 0.6279171109199524,
      "learning_rate": 0.00038383579355672875,
      "loss": 3.4574,
      "step": 33550
    },
    {
      "epoch": 3.616402970616726,
      "grad_norm": 0.6288803815841675,
      "learning_rate": 0.00038351255252666735,
      "loss": 3.4601,
      "step": 33600
    },
    {
      "epoch": 3.621784522656334,
      "grad_norm": 0.6291745901107788,
      "learning_rate": 0.00038318931149660594,
      "loss": 3.4573,
      "step": 33650
    },
    {
      "epoch": 3.627166074695942,
      "grad_norm": 0.616725742816925,
      "learning_rate": 0.00038286607046654454,
      "loss": 3.4526,
      "step": 33700
    },
    {
      "epoch": 3.6325476267355503,
      "grad_norm": 0.652170717716217,
      "learning_rate": 0.00038254282943648313,
      "loss": 3.4446,
      "step": 33750
    },
    {
      "epoch": 3.637929178775159,
      "grad_norm": 0.5712392926216125,
      "learning_rate": 0.00038221958840642167,
      "loss": 3.459,
      "step": 33800
    },
    {
      "epoch": 3.643310730814767,
      "grad_norm": 0.645204484462738,
      "learning_rate": 0.00038189634737636027,
      "loss": 3.4547,
      "step": 33850
    },
    {
      "epoch": 3.648692282854375,
      "grad_norm": 0.6385588049888611,
      "learning_rate": 0.0003815731063462988,
      "loss": 3.4552,
      "step": 33900
    },
    {
      "epoch": 3.6540738348939836,
      "grad_norm": 0.5498437881469727,
      "learning_rate": 0.00038124986531623745,
      "loss": 3.46,
      "step": 33950
    },
    {
      "epoch": 3.6594553869335917,
      "grad_norm": 0.6273448467254639,
      "learning_rate": 0.00038092662428617605,
      "loss": 3.4847,
      "step": 34000
    },
    {
      "epoch": 3.6594553869335917,
      "eval_accuracy": 0.37375724185078896,
      "eval_loss": 3.470141887664795,
      "eval_runtime": 184.502,
      "eval_samples_per_second": 97.62,
      "eval_steps_per_second": 6.103,
      "step": 34000
    },
    {
      "epoch": 3.6648369389732,
      "grad_norm": 0.6258223652839661,
      "learning_rate": 0.0003806033832561146,
      "loss": 3.4565,
      "step": 34050
    },
    {
      "epoch": 3.670218491012808,
      "grad_norm": 0.6277667284011841,
      "learning_rate": 0.0003802801422260532,
      "loss": 3.4619,
      "step": 34100
    },
    {
      "epoch": 3.675600043052416,
      "grad_norm": 0.5932169556617737,
      "learning_rate": 0.000379963366016593,
      "loss": 3.4508,
      "step": 34150
    },
    {
      "epoch": 3.6809815950920246,
      "grad_norm": 0.6298589110374451,
      "learning_rate": 0.0003796401249865316,
      "loss": 3.4613,
      "step": 34200
    },
    {
      "epoch": 3.6863631471316327,
      "grad_norm": 0.6247563362121582,
      "learning_rate": 0.0003793168839564701,
      "loss": 3.4507,
      "step": 34250
    },
    {
      "epoch": 3.691744699171241,
      "grad_norm": 0.5867701172828674,
      "learning_rate": 0.0003789936429264088,
      "loss": 3.4696,
      "step": 34300
    },
    {
      "epoch": 3.6971262512108494,
      "grad_norm": 0.6233004927635193,
      "learning_rate": 0.00037867040189634737,
      "loss": 3.4705,
      "step": 34350
    },
    {
      "epoch": 3.7025078032504575,
      "grad_norm": 0.6311918497085571,
      "learning_rate": 0.0003783471608662859,
      "loss": 3.4627,
      "step": 34400
    },
    {
      "epoch": 3.7078893552900656,
      "grad_norm": 0.6299039125442505,
      "learning_rate": 0.0003780239198362245,
      "loss": 3.4684,
      "step": 34450
    },
    {
      "epoch": 3.713270907329674,
      "grad_norm": 0.6793944239616394,
      "learning_rate": 0.0003777006788061631,
      "loss": 3.4433,
      "step": 34500
    },
    {
      "epoch": 3.7186524593692822,
      "grad_norm": 0.6050877571105957,
      "learning_rate": 0.00037737743777610164,
      "loss": 3.4569,
      "step": 34550
    },
    {
      "epoch": 3.7240340114088903,
      "grad_norm": 0.6099978685379028,
      "learning_rate": 0.0003770541967460403,
      "loss": 3.4428,
      "step": 34600
    },
    {
      "epoch": 3.7294155634484984,
      "grad_norm": 0.6048750877380371,
      "learning_rate": 0.0003767309557159789,
      "loss": 3.4715,
      "step": 34650
    },
    {
      "epoch": 3.7347971154881066,
      "grad_norm": 0.6237765550613403,
      "learning_rate": 0.0003764077146859174,
      "loss": 3.4676,
      "step": 34700
    },
    {
      "epoch": 3.740178667527715,
      "grad_norm": 0.7071151733398438,
      "learning_rate": 0.000376084473655856,
      "loss": 3.4621,
      "step": 34750
    },
    {
      "epoch": 3.745560219567323,
      "grad_norm": 0.6428067684173584,
      "learning_rate": 0.00037576123262579456,
      "loss": 3.468,
      "step": 34800
    },
    {
      "epoch": 3.7509417716069313,
      "grad_norm": 0.6198859810829163,
      "learning_rate": 0.00037543799159573315,
      "loss": 3.4467,
      "step": 34850
    },
    {
      "epoch": 3.75632332364654,
      "grad_norm": 0.6678587198257446,
      "learning_rate": 0.0003751147505656718,
      "loss": 3.4583,
      "step": 34900
    },
    {
      "epoch": 3.761704875686148,
      "grad_norm": 0.6549521684646606,
      "learning_rate": 0.00037479150953561034,
      "loss": 3.4339,
      "step": 34950
    },
    {
      "epoch": 3.767086427725756,
      "grad_norm": 0.6431390047073364,
      "learning_rate": 0.00037446826850554894,
      "loss": 3.4697,
      "step": 35000
    },
    {
      "epoch": 3.767086427725756,
      "eval_accuracy": 0.3741979379637645,
      "eval_loss": 3.461988925933838,
      "eval_runtime": 184.229,
      "eval_samples_per_second": 97.764,
      "eval_steps_per_second": 6.112,
      "step": 35000
    },
    {
      "epoch": 3.772467979765364,
      "grad_norm": 0.665264904499054,
      "learning_rate": 0.00037414502747548753,
      "loss": 3.4732,
      "step": 35050
    },
    {
      "epoch": 3.7778495318049723,
      "grad_norm": 0.602967381477356,
      "learning_rate": 0.00037382178644542607,
      "loss": 3.4703,
      "step": 35100
    },
    {
      "epoch": 3.783231083844581,
      "grad_norm": 0.6076384782791138,
      "learning_rate": 0.00037349854541536467,
      "loss": 3.4531,
      "step": 35150
    },
    {
      "epoch": 3.788612635884189,
      "grad_norm": 0.6334365606307983,
      "learning_rate": 0.0003731753043853033,
      "loss": 3.4552,
      "step": 35200
    },
    {
      "epoch": 3.793994187923797,
      "grad_norm": 0.6087632775306702,
      "learning_rate": 0.00037285206335524186,
      "loss": 3.4621,
      "step": 35250
    },
    {
      "epoch": 3.7993757399634056,
      "grad_norm": 0.5887856483459473,
      "learning_rate": 0.00037252882232518045,
      "loss": 3.4462,
      "step": 35300
    },
    {
      "epoch": 3.8047572920030137,
      "grad_norm": 0.6560126543045044,
      "learning_rate": 0.000372205581295119,
      "loss": 3.4574,
      "step": 35350
    },
    {
      "epoch": 3.810138844042622,
      "grad_norm": 0.6371379494667053,
      "learning_rate": 0.0003718823402650576,
      "loss": 3.466,
      "step": 35400
    },
    {
      "epoch": 3.8155203960822304,
      "grad_norm": 0.6066793203353882,
      "learning_rate": 0.00037155909923499624,
      "loss": 3.4624,
      "step": 35450
    },
    {
      "epoch": 3.8209019481218385,
      "grad_norm": 0.6712152361869812,
      "learning_rate": 0.0003712358582049348,
      "loss": 3.4629,
      "step": 35500
    },
    {
      "epoch": 3.8262835001614466,
      "grad_norm": 0.6405826210975647,
      "learning_rate": 0.00037091261717487337,
      "loss": 3.4445,
      "step": 35550
    },
    {
      "epoch": 3.8316650522010547,
      "grad_norm": 0.6559590697288513,
      "learning_rate": 0.00037058937614481197,
      "loss": 3.4596,
      "step": 35600
    },
    {
      "epoch": 3.837046604240663,
      "grad_norm": 0.6620156168937683,
      "learning_rate": 0.0003702661351147505,
      "loss": 3.45,
      "step": 35650
    },
    {
      "epoch": 3.8424281562802713,
      "grad_norm": 0.6478715538978577,
      "learning_rate": 0.0003699428940846891,
      "loss": 3.4644,
      "step": 35700
    },
    {
      "epoch": 3.8478097083198795,
      "grad_norm": 0.6179508566856384,
      "learning_rate": 0.00036961965305462775,
      "loss": 3.4554,
      "step": 35750
    },
    {
      "epoch": 3.8531912603594876,
      "grad_norm": 0.635578989982605,
      "learning_rate": 0.0003692964120245663,
      "loss": 3.4666,
      "step": 35800
    },
    {
      "epoch": 3.858572812399096,
      "grad_norm": 0.6176955103874207,
      "learning_rate": 0.0003689731709945049,
      "loss": 3.459,
      "step": 35850
    },
    {
      "epoch": 3.863954364438704,
      "grad_norm": 0.6287024021148682,
      "learning_rate": 0.0003686499299644434,
      "loss": 3.4637,
      "step": 35900
    },
    {
      "epoch": 3.8693359164783123,
      "grad_norm": 0.5953682661056519,
      "learning_rate": 0.000368326688934382,
      "loss": 3.4471,
      "step": 35950
    },
    {
      "epoch": 3.8747174685179204,
      "grad_norm": 0.6310164332389832,
      "learning_rate": 0.0003680034479043206,
      "loss": 3.4352,
      "step": 36000
    },
    {
      "epoch": 3.8747174685179204,
      "eval_accuracy": 0.37532532032772753,
      "eval_loss": 3.45635986328125,
      "eval_runtime": 184.341,
      "eval_samples_per_second": 97.705,
      "eval_steps_per_second": 6.108,
      "step": 36000
    },
    {
      "epoch": 3.8800990205575285,
      "grad_norm": 0.6187303066253662,
      "learning_rate": 0.0003676802068742592,
      "loss": 3.4567,
      "step": 36050
    },
    {
      "epoch": 3.885480572597137,
      "grad_norm": 0.6245608925819397,
      "learning_rate": 0.0003673569658441978,
      "loss": 3.4581,
      "step": 36100
    },
    {
      "epoch": 3.890862124636745,
      "grad_norm": 0.6241065263748169,
      "learning_rate": 0.0003670401896347376,
      "loss": 3.4457,
      "step": 36150
    },
    {
      "epoch": 3.8962436766763533,
      "grad_norm": 0.6071236729621887,
      "learning_rate": 0.0003667169486046762,
      "loss": 3.4711,
      "step": 36200
    },
    {
      "epoch": 3.901625228715962,
      "grad_norm": 0.6289729475975037,
      "learning_rate": 0.00036639370757461475,
      "loss": 3.4563,
      "step": 36250
    },
    {
      "epoch": 3.90700678075557,
      "grad_norm": 0.7036654949188232,
      "learning_rate": 0.00036607046654455334,
      "loss": 3.4891,
      "step": 36300
    },
    {
      "epoch": 3.912388332795178,
      "grad_norm": 0.6079036593437195,
      "learning_rate": 0.00036574722551449193,
      "loss": 3.4529,
      "step": 36350
    },
    {
      "epoch": 3.9177698848347866,
      "grad_norm": 0.6209616661071777,
      "learning_rate": 0.00036542398448443053,
      "loss": 3.4487,
      "step": 36400
    },
    {
      "epoch": 3.9231514368743947,
      "grad_norm": 0.6106263399124146,
      "learning_rate": 0.0003651007434543691,
      "loss": 3.4364,
      "step": 36450
    },
    {
      "epoch": 3.928532988914003,
      "grad_norm": 0.6158912777900696,
      "learning_rate": 0.0003647775024243077,
      "loss": 3.4647,
      "step": 36500
    },
    {
      "epoch": 3.933914540953611,
      "grad_norm": 0.6436479687690735,
      "learning_rate": 0.00036445426139424626,
      "loss": 3.4579,
      "step": 36550
    },
    {
      "epoch": 3.939296092993219,
      "grad_norm": 0.6153359413146973,
      "learning_rate": 0.00036413102036418485,
      "loss": 3.4624,
      "step": 36600
    },
    {
      "epoch": 3.9446776450328276,
      "grad_norm": 0.6433219313621521,
      "learning_rate": 0.0003638077793341234,
      "loss": 3.4537,
      "step": 36650
    },
    {
      "epoch": 3.9500591970724357,
      "grad_norm": 0.5967046022415161,
      "learning_rate": 0.00036348453830406204,
      "loss": 3.4637,
      "step": 36700
    },
    {
      "epoch": 3.955440749112044,
      "grad_norm": 0.6453063488006592,
      "learning_rate": 0.00036316129727400064,
      "loss": 3.4488,
      "step": 36750
    },
    {
      "epoch": 3.9608223011516523,
      "grad_norm": 0.6288018226623535,
      "learning_rate": 0.0003628380562439392,
      "loss": 3.4439,
      "step": 36800
    },
    {
      "epoch": 3.9662038531912605,
      "grad_norm": 0.6240979433059692,
      "learning_rate": 0.00036251481521387777,
      "loss": 3.4585,
      "step": 36850
    },
    {
      "epoch": 3.9715854052308686,
      "grad_norm": 0.6409454345703125,
      "learning_rate": 0.00036219157418381637,
      "loss": 3.4469,
      "step": 36900
    },
    {
      "epoch": 3.9769669572704767,
      "grad_norm": 0.6430519819259644,
      "learning_rate": 0.0003618683331537549,
      "loss": 3.4697,
      "step": 36950
    },
    {
      "epoch": 3.9823485093100848,
      "grad_norm": 0.6589070558547974,
      "learning_rate": 0.00036154509212369356,
      "loss": 3.4524,
      "step": 37000
    },
    {
      "epoch": 3.9823485093100848,
      "eval_accuracy": 0.3757677548869081,
      "eval_loss": 3.450047016143799,
      "eval_runtime": 184.6029,
      "eval_samples_per_second": 97.566,
      "eval_steps_per_second": 6.1,
      "step": 37000
    },
    {
      "epoch": 3.9877300613496933,
      "grad_norm": 0.6687272191047668,
      "learning_rate": 0.00036122185109363215,
      "loss": 3.4402,
      "step": 37050
    },
    {
      "epoch": 3.9931116133893014,
      "grad_norm": 0.677852988243103,
      "learning_rate": 0.0003608986100635707,
      "loss": 3.4562,
      "step": 37100
    },
    {
      "epoch": 3.9984931654289095,
      "grad_norm": 0.6325944662094116,
      "learning_rate": 0.0003605753690335093,
      "loss": 3.4701,
      "step": 37150
    },
    {
      "epoch": 4.003874717468518,
      "grad_norm": 0.6537460088729858,
      "learning_rate": 0.0003602521280034478,
      "loss": 3.3768,
      "step": 37200
    },
    {
      "epoch": 4.009256269508126,
      "grad_norm": 0.647741973400116,
      "learning_rate": 0.0003599288869733865,
      "loss": 3.3647,
      "step": 37250
    },
    {
      "epoch": 4.014637821547734,
      "grad_norm": 0.6205103993415833,
      "learning_rate": 0.00035960564594332507,
      "loss": 3.3697,
      "step": 37300
    },
    {
      "epoch": 4.020019373587343,
      "grad_norm": 0.6998143196105957,
      "learning_rate": 0.0003592824049132636,
      "loss": 3.3571,
      "step": 37350
    },
    {
      "epoch": 4.0254009256269505,
      "grad_norm": 0.6265931129455566,
      "learning_rate": 0.0003589591638832022,
      "loss": 3.3488,
      "step": 37400
    },
    {
      "epoch": 4.030782477666559,
      "grad_norm": 0.6777806878089905,
      "learning_rate": 0.0003586359228531408,
      "loss": 3.359,
      "step": 37450
    },
    {
      "epoch": 4.036164029706168,
      "grad_norm": 0.6596590876579285,
      "learning_rate": 0.00035831268182307934,
      "loss": 3.3777,
      "step": 37500
    },
    {
      "epoch": 4.041545581745775,
      "grad_norm": 0.6730514764785767,
      "learning_rate": 0.000357989440793018,
      "loss": 3.3527,
      "step": 37550
    },
    {
      "epoch": 4.046927133785384,
      "grad_norm": 0.6630343794822693,
      "learning_rate": 0.0003576661997629566,
      "loss": 3.3547,
      "step": 37600
    },
    {
      "epoch": 4.0523086858249915,
      "grad_norm": 0.660653829574585,
      "learning_rate": 0.0003573429587328951,
      "loss": 3.3949,
      "step": 37650
    },
    {
      "epoch": 4.0576902378646,
      "grad_norm": 0.6169551014900208,
      "learning_rate": 0.0003570197177028337,
      "loss": 3.379,
      "step": 37700
    },
    {
      "epoch": 4.063071789904209,
      "grad_norm": 0.6500664949417114,
      "learning_rate": 0.00035669647667277226,
      "loss": 3.3714,
      "step": 37750
    },
    {
      "epoch": 4.068453341943816,
      "grad_norm": 0.6486415266990662,
      "learning_rate": 0.00035637323564271085,
      "loss": 3.3735,
      "step": 37800
    },
    {
      "epoch": 4.073834893983425,
      "grad_norm": 0.6548293232917786,
      "learning_rate": 0.0003560499946126495,
      "loss": 3.3695,
      "step": 37850
    },
    {
      "epoch": 4.079216446023033,
      "grad_norm": 0.6383534669876099,
      "learning_rate": 0.00035572675358258804,
      "loss": 3.3741,
      "step": 37900
    },
    {
      "epoch": 4.084597998062641,
      "grad_norm": 0.6793497204780579,
      "learning_rate": 0.00035540351255252664,
      "loss": 3.3779,
      "step": 37950
    },
    {
      "epoch": 4.08997955010225,
      "grad_norm": 0.6776780486106873,
      "learning_rate": 0.00035508027152246523,
      "loss": 3.3825,
      "step": 38000
    },
    {
      "epoch": 4.08997955010225,
      "eval_accuracy": 0.3756581241231033,
      "eval_loss": 3.4540176391601562,
      "eval_runtime": 184.3996,
      "eval_samples_per_second": 97.674,
      "eval_steps_per_second": 6.106,
      "step": 38000
    },
    {
      "epoch": 4.095361102141858,
      "grad_norm": 0.6693257689476013,
      "learning_rate": 0.0003547570304924038,
      "loss": 3.3932,
      "step": 38050
    },
    {
      "epoch": 4.100742654181466,
      "grad_norm": 0.6108912229537964,
      "learning_rate": 0.0003544337894623424,
      "loss": 3.3812,
      "step": 38100
    },
    {
      "epoch": 4.106124206221074,
      "grad_norm": 0.6528989672660828,
      "learning_rate": 0.000354110548432281,
      "loss": 3.383,
      "step": 38150
    },
    {
      "epoch": 4.111505758260682,
      "grad_norm": 0.6330206990242004,
      "learning_rate": 0.0003537937722228208,
      "loss": 3.3866,
      "step": 38200
    },
    {
      "epoch": 4.1168873103002905,
      "grad_norm": 0.700124204158783,
      "learning_rate": 0.00035347053119275936,
      "loss": 3.3848,
      "step": 38250
    },
    {
      "epoch": 4.122268862339899,
      "grad_norm": 0.6745419502258301,
      "learning_rate": 0.00035314729016269796,
      "loss": 3.383,
      "step": 38300
    },
    {
      "epoch": 4.127650414379507,
      "grad_norm": 0.6428250670433044,
      "learning_rate": 0.00035282404913263655,
      "loss": 3.3616,
      "step": 38350
    },
    {
      "epoch": 4.133031966419115,
      "grad_norm": 0.6580246686935425,
      "learning_rate": 0.0003525008081025751,
      "loss": 3.3808,
      "step": 38400
    },
    {
      "epoch": 4.138413518458724,
      "grad_norm": 0.6682559251785278,
      "learning_rate": 0.0003521775670725137,
      "loss": 3.3839,
      "step": 38450
    },
    {
      "epoch": 4.1437950704983315,
      "grad_norm": 0.6492109298706055,
      "learning_rate": 0.00035185432604245234,
      "loss": 3.3793,
      "step": 38500
    },
    {
      "epoch": 4.14917662253794,
      "grad_norm": 0.6978576183319092,
      "learning_rate": 0.0003515310850123909,
      "loss": 3.3641,
      "step": 38550
    },
    {
      "epoch": 4.154558174577549,
      "grad_norm": 0.6698665618896484,
      "learning_rate": 0.00035120784398232947,
      "loss": 3.3818,
      "step": 38600
    },
    {
      "epoch": 4.159939726617156,
      "grad_norm": 0.6258059740066528,
      "learning_rate": 0.000350884602952268,
      "loss": 3.3702,
      "step": 38650
    },
    {
      "epoch": 4.165321278656765,
      "grad_norm": 0.6838365793228149,
      "learning_rate": 0.0003505613619222066,
      "loss": 3.3946,
      "step": 38700
    },
    {
      "epoch": 4.1707028306963725,
      "grad_norm": 0.704094648361206,
      "learning_rate": 0.0003502381208921452,
      "loss": 3.3846,
      "step": 38750
    },
    {
      "epoch": 4.176084382735981,
      "grad_norm": 0.625027060508728,
      "learning_rate": 0.0003499148798620838,
      "loss": 3.3808,
      "step": 38800
    },
    {
      "epoch": 4.18146593477559,
      "grad_norm": 0.700301468372345,
      "learning_rate": 0.0003495916388320224,
      "loss": 3.3911,
      "step": 38850
    },
    {
      "epoch": 4.186847486815197,
      "grad_norm": 0.7061277031898499,
      "learning_rate": 0.000349268397801961,
      "loss": 3.3985,
      "step": 38900
    },
    {
      "epoch": 4.192229038854806,
      "grad_norm": 0.6783473491668701,
      "learning_rate": 0.0003489451567718995,
      "loss": 3.3627,
      "step": 38950
    },
    {
      "epoch": 4.197610590894414,
      "grad_norm": 0.6807799935340881,
      "learning_rate": 0.0003486219157418381,
      "loss": 3.3814,
      "step": 39000
    },
    {
      "epoch": 4.197610590894414,
      "eval_accuracy": 0.3761339151188429,
      "eval_loss": 3.4509873390197754,
      "eval_runtime": 184.3645,
      "eval_samples_per_second": 97.692,
      "eval_steps_per_second": 6.107,
      "step": 39000
    },
    {
      "epoch": 4.202992142934022,
      "grad_norm": 0.6863417029380798,
      "learning_rate": 0.00034829867471177677,
      "loss": 3.3808,
      "step": 39050
    },
    {
      "epoch": 4.208373694973631,
      "grad_norm": 0.6438162326812744,
      "learning_rate": 0.0003479754336817153,
      "loss": 3.3973,
      "step": 39100
    },
    {
      "epoch": 4.213755247013238,
      "grad_norm": 0.6513394713401794,
      "learning_rate": 0.0003476521926516539,
      "loss": 3.3859,
      "step": 39150
    },
    {
      "epoch": 4.219136799052847,
      "grad_norm": 0.6030909419059753,
      "learning_rate": 0.00034732895162159245,
      "loss": 3.3753,
      "step": 39200
    },
    {
      "epoch": 4.224518351092455,
      "grad_norm": 0.6581124067306519,
      "learning_rate": 0.00034700571059153104,
      "loss": 3.3855,
      "step": 39250
    },
    {
      "epoch": 4.229899903132063,
      "grad_norm": 0.6777631044387817,
      "learning_rate": 0.00034668246956146963,
      "loss": 3.3831,
      "step": 39300
    },
    {
      "epoch": 4.2352814551716715,
      "grad_norm": 0.6349054574966431,
      "learning_rate": 0.00034635922853140823,
      "loss": 3.3897,
      "step": 39350
    },
    {
      "epoch": 4.24066300721128,
      "grad_norm": 0.6555729508399963,
      "learning_rate": 0.0003460359875013468,
      "loss": 3.386,
      "step": 39400
    },
    {
      "epoch": 4.246044559250888,
      "grad_norm": 0.6674342751502991,
      "learning_rate": 0.0003457127464712854,
      "loss": 3.385,
      "step": 39450
    },
    {
      "epoch": 4.251426111290496,
      "grad_norm": 0.7295530438423157,
      "learning_rate": 0.00034538950544122396,
      "loss": 3.4145,
      "step": 39500
    },
    {
      "epoch": 4.256807663330104,
      "grad_norm": 0.6763948798179626,
      "learning_rate": 0.00034506626441116255,
      "loss": 3.3947,
      "step": 39550
    },
    {
      "epoch": 4.2621892153697125,
      "grad_norm": 0.7632237076759338,
      "learning_rate": 0.0003447430233811011,
      "loss": 3.39,
      "step": 39600
    },
    {
      "epoch": 4.267570767409321,
      "grad_norm": 0.6812435984611511,
      "learning_rate": 0.00034441978235103974,
      "loss": 3.3988,
      "step": 39650
    },
    {
      "epoch": 4.272952319448929,
      "grad_norm": 0.6726835370063782,
      "learning_rate": 0.00034409654132097834,
      "loss": 3.3989,
      "step": 39700
    },
    {
      "epoch": 4.278333871488537,
      "grad_norm": 0.6316977143287659,
      "learning_rate": 0.0003437733002909169,
      "loss": 3.4094,
      "step": 39750
    },
    {
      "epoch": 4.283715423528146,
      "grad_norm": 0.7431002855300903,
      "learning_rate": 0.00034345005926085547,
      "loss": 3.386,
      "step": 39800
    },
    {
      "epoch": 4.2890969755677535,
      "grad_norm": 0.6456528902053833,
      "learning_rate": 0.00034312681823079407,
      "loss": 3.3864,
      "step": 39850
    },
    {
      "epoch": 4.294478527607362,
      "grad_norm": 0.6434244513511658,
      "learning_rate": 0.00034280357720073266,
      "loss": 3.3952,
      "step": 39900
    },
    {
      "epoch": 4.299860079646971,
      "grad_norm": 0.6549796462059021,
      "learning_rate": 0.00034248033617067126,
      "loss": 3.3948,
      "step": 39950
    },
    {
      "epoch": 4.305241631686578,
      "grad_norm": 0.6274550557136536,
      "learning_rate": 0.00034215709514060985,
      "loss": 3.3748,
      "step": 40000
    },
    {
      "epoch": 4.305241631686578,
      "eval_accuracy": 0.3765880841899074,
      "eval_loss": 3.445739269256592,
      "eval_runtime": 184.3137,
      "eval_samples_per_second": 97.719,
      "eval_steps_per_second": 6.109,
      "step": 40000
    },
    {
      "epoch": 4.310623183726187,
      "grad_norm": 0.6608153581619263,
      "learning_rate": 0.0003418338541105484,
      "loss": 3.4074,
      "step": 40050
    },
    {
      "epoch": 4.3160047357657945,
      "grad_norm": 0.641989529132843,
      "learning_rate": 0.000341510613080487,
      "loss": 3.3848,
      "step": 40100
    },
    {
      "epoch": 4.321386287805403,
      "grad_norm": 0.731931209564209,
      "learning_rate": 0.0003411873720504255,
      "loss": 3.3996,
      "step": 40150
    },
    {
      "epoch": 4.326767839845012,
      "grad_norm": 0.672824501991272,
      "learning_rate": 0.0003408705958409654,
      "loss": 3.4023,
      "step": 40200
    },
    {
      "epoch": 4.332149391884619,
      "grad_norm": 0.694875180721283,
      "learning_rate": 0.00034054735481090393,
      "loss": 3.3845,
      "step": 40250
    },
    {
      "epoch": 4.337530943924228,
      "grad_norm": 0.693645179271698,
      "learning_rate": 0.0003402241137808426,
      "loss": 3.3843,
      "step": 40300
    },
    {
      "epoch": 4.342912495963836,
      "grad_norm": 0.6279447078704834,
      "learning_rate": 0.00033990087275078117,
      "loss": 3.4033,
      "step": 40350
    },
    {
      "epoch": 4.348294048003444,
      "grad_norm": 0.6362826228141785,
      "learning_rate": 0.0003395776317207197,
      "loss": 3.4085,
      "step": 40400
    },
    {
      "epoch": 4.3536756000430525,
      "grad_norm": 0.6189711093902588,
      "learning_rate": 0.0003392543906906583,
      "loss": 3.391,
      "step": 40450
    },
    {
      "epoch": 4.359057152082661,
      "grad_norm": 0.6788389682769775,
      "learning_rate": 0.00033893114966059685,
      "loss": 3.3907,
      "step": 40500
    },
    {
      "epoch": 4.364438704122269,
      "grad_norm": 0.6872463226318359,
      "learning_rate": 0.00033860790863053544,
      "loss": 3.3888,
      "step": 40550
    },
    {
      "epoch": 4.369820256161877,
      "grad_norm": 0.7346479296684265,
      "learning_rate": 0.0003382846676004741,
      "loss": 3.3799,
      "step": 40600
    },
    {
      "epoch": 4.375201808201485,
      "grad_norm": 0.6281903982162476,
      "learning_rate": 0.00033796142657041263,
      "loss": 3.4074,
      "step": 40650
    },
    {
      "epoch": 4.3805833602410935,
      "grad_norm": 0.7750927805900574,
      "learning_rate": 0.0003376381855403512,
      "loss": 3.3981,
      "step": 40700
    },
    {
      "epoch": 4.385964912280702,
      "grad_norm": 0.6273497343063354,
      "learning_rate": 0.0003373149445102898,
      "loss": 3.4079,
      "step": 40750
    },
    {
      "epoch": 4.39134646432031,
      "grad_norm": 0.6399444341659546,
      "learning_rate": 0.00033699170348022836,
      "loss": 3.3924,
      "step": 40800
    },
    {
      "epoch": 4.396728016359918,
      "grad_norm": 0.6856110095977783,
      "learning_rate": 0.000336668462450167,
      "loss": 3.4077,
      "step": 40850
    },
    {
      "epoch": 4.402109568399527,
      "grad_norm": 0.6612463593482971,
      "learning_rate": 0.0003363452214201056,
      "loss": 3.3948,
      "step": 40900
    },
    {
      "epoch": 4.4074911204391345,
      "grad_norm": 0.6539753079414368,
      "learning_rate": 0.00033602198039004415,
      "loss": 3.3819,
      "step": 40950
    },
    {
      "epoch": 4.412872672478743,
      "grad_norm": 0.6288073658943176,
      "learning_rate": 0.00033569873935998274,
      "loss": 3.4025,
      "step": 41000
    },
    {
      "epoch": 4.412872672478743,
      "eval_accuracy": 0.3771389543311269,
      "eval_loss": 3.4400978088378906,
      "eval_runtime": 184.5839,
      "eval_samples_per_second": 97.576,
      "eval_steps_per_second": 6.1,
      "step": 41000
    },
    {
      "epoch": 4.418254224518351,
      "grad_norm": 0.6586681008338928,
      "learning_rate": 0.0003353754983299213,
      "loss": 3.3982,
      "step": 41050
    },
    {
      "epoch": 4.423635776557959,
      "grad_norm": 0.652645468711853,
      "learning_rate": 0.0003350522572998599,
      "loss": 3.3838,
      "step": 41100
    },
    {
      "epoch": 4.429017328597568,
      "grad_norm": 0.6442344188690186,
      "learning_rate": 0.0003347290162697985,
      "loss": 3.3652,
      "step": 41150
    },
    {
      "epoch": 4.4343988806371755,
      "grad_norm": 0.692879855632782,
      "learning_rate": 0.00033440577523973706,
      "loss": 3.3928,
      "step": 41200
    },
    {
      "epoch": 4.439780432676784,
      "grad_norm": 0.6954250931739807,
      "learning_rate": 0.00033408253420967566,
      "loss": 3.4042,
      "step": 41250
    },
    {
      "epoch": 4.445161984716393,
      "grad_norm": 0.6553548574447632,
      "learning_rate": 0.00033375929317961425,
      "loss": 3.412,
      "step": 41300
    },
    {
      "epoch": 4.450543536756,
      "grad_norm": 0.6722890138626099,
      "learning_rate": 0.0003334360521495528,
      "loss": 3.3936,
      "step": 41350
    },
    {
      "epoch": 4.455925088795609,
      "grad_norm": 0.6921664476394653,
      "learning_rate": 0.0003331128111194914,
      "loss": 3.393,
      "step": 41400
    },
    {
      "epoch": 4.461306640835216,
      "grad_norm": 0.657521665096283,
      "learning_rate": 0.00033278957008943004,
      "loss": 3.3996,
      "step": 41450
    },
    {
      "epoch": 4.466688192874825,
      "grad_norm": 0.6841620206832886,
      "learning_rate": 0.0003324663290593686,
      "loss": 3.4023,
      "step": 41500
    },
    {
      "epoch": 4.4720697449144335,
      "grad_norm": 0.7023841142654419,
      "learning_rate": 0.00033214308802930717,
      "loss": 3.3836,
      "step": 41550
    },
    {
      "epoch": 4.477451296954041,
      "grad_norm": 0.6275233626365662,
      "learning_rate": 0.0003318198469992457,
      "loss": 3.3897,
      "step": 41600
    },
    {
      "epoch": 4.48283284899365,
      "grad_norm": 0.6907360553741455,
      "learning_rate": 0.0003314966059691843,
      "loss": 3.3832,
      "step": 41650
    },
    {
      "epoch": 4.488214401033258,
      "grad_norm": 0.7085883617401123,
      "learning_rate": 0.00033117336493912296,
      "loss": 3.3936,
      "step": 41700
    },
    {
      "epoch": 4.493595953072866,
      "grad_norm": 0.6712943911552429,
      "learning_rate": 0.0003308501239090615,
      "loss": 3.4087,
      "step": 41750
    },
    {
      "epoch": 4.4989775051124745,
      "grad_norm": 0.6680316925048828,
      "learning_rate": 0.0003305268828790001,
      "loss": 3.3956,
      "step": 41800
    },
    {
      "epoch": 4.504359057152083,
      "grad_norm": 0.6576794981956482,
      "learning_rate": 0.0003302036418489387,
      "loss": 3.4089,
      "step": 41850
    },
    {
      "epoch": 4.509740609191691,
      "grad_norm": 0.6819412708282471,
      "learning_rate": 0.0003298804008188772,
      "loss": 3.3925,
      "step": 41900
    },
    {
      "epoch": 4.515122161231299,
      "grad_norm": 0.6794874668121338,
      "learning_rate": 0.0003295571597888158,
      "loss": 3.3908,
      "step": 41950
    },
    {
      "epoch": 4.520503713270907,
      "grad_norm": 0.6656818389892578,
      "learning_rate": 0.00032923391875875447,
      "loss": 3.4143,
      "step": 42000
    },
    {
      "epoch": 4.520503713270907,
      "eval_accuracy": 0.3778304213091782,
      "eval_loss": 3.4336068630218506,
      "eval_runtime": 184.7162,
      "eval_samples_per_second": 97.506,
      "eval_steps_per_second": 6.096,
      "step": 42000
    },
    {
      "epoch": 4.5258852653105155,
      "grad_norm": 0.672918438911438,
      "learning_rate": 0.000328910677728693,
      "loss": 3.3995,
      "step": 42050
    },
    {
      "epoch": 4.531266817350124,
      "grad_norm": 0.7011841535568237,
      "learning_rate": 0.0003285874366986316,
      "loss": 3.3871,
      "step": 42100
    },
    {
      "epoch": 4.536648369389732,
      "grad_norm": 0.7325242757797241,
      "learning_rate": 0.00032826419566857015,
      "loss": 3.3944,
      "step": 42150
    },
    {
      "epoch": 4.54202992142934,
      "grad_norm": 0.6681964993476868,
      "learning_rate": 0.00032794741945911,
      "loss": 3.4031,
      "step": 42200
    },
    {
      "epoch": 4.547411473468949,
      "grad_norm": 0.6524448990821838,
      "learning_rate": 0.00032762417842904855,
      "loss": 3.3932,
      "step": 42250
    },
    {
      "epoch": 4.5527930255085565,
      "grad_norm": 0.6727701425552368,
      "learning_rate": 0.00032730093739898714,
      "loss": 3.4137,
      "step": 42300
    },
    {
      "epoch": 4.558174577548165,
      "grad_norm": 0.7065811157226562,
      "learning_rate": 0.0003269776963689257,
      "loss": 3.3942,
      "step": 42350
    },
    {
      "epoch": 4.563556129587774,
      "grad_norm": 0.6715471148490906,
      "learning_rate": 0.00032665445533886433,
      "loss": 3.4131,
      "step": 42400
    },
    {
      "epoch": 4.568937681627381,
      "grad_norm": 0.6726318597793579,
      "learning_rate": 0.0003263312143088029,
      "loss": 3.3932,
      "step": 42450
    },
    {
      "epoch": 4.57431923366699,
      "grad_norm": 0.6428508758544922,
      "learning_rate": 0.00032600797327874147,
      "loss": 3.3975,
      "step": 42500
    },
    {
      "epoch": 4.579700785706597,
      "grad_norm": 0.6645674705505371,
      "learning_rate": 0.00032568473224868006,
      "loss": 3.4091,
      "step": 42550
    },
    {
      "epoch": 4.585082337746206,
      "grad_norm": 0.6850250959396362,
      "learning_rate": 0.00032536149121861866,
      "loss": 3.3983,
      "step": 42600
    },
    {
      "epoch": 4.5904638897858145,
      "grad_norm": 0.6411243081092834,
      "learning_rate": 0.00032503825018855725,
      "loss": 3.4032,
      "step": 42650
    },
    {
      "epoch": 4.595845441825422,
      "grad_norm": 0.6782857179641724,
      "learning_rate": 0.00032471500915849584,
      "loss": 3.3989,
      "step": 42700
    },
    {
      "epoch": 4.601226993865031,
      "grad_norm": 0.7062470316886902,
      "learning_rate": 0.00032439176812843444,
      "loss": 3.3831,
      "step": 42750
    },
    {
      "epoch": 4.606608545904638,
      "grad_norm": 0.6140323877334595,
      "learning_rate": 0.000324068527098373,
      "loss": 3.3997,
      "step": 42800
    },
    {
      "epoch": 4.611990097944247,
      "grad_norm": 0.6945203542709351,
      "learning_rate": 0.0003237452860683116,
      "loss": 3.3963,
      "step": 42850
    },
    {
      "epoch": 4.6173716499838555,
      "grad_norm": 0.6892925500869751,
      "learning_rate": 0.0003234220450382501,
      "loss": 3.3862,
      "step": 42900
    },
    {
      "epoch": 4.622753202023463,
      "grad_norm": 0.6654855608940125,
      "learning_rate": 0.00032309880400818876,
      "loss": 3.3885,
      "step": 42950
    },
    {
      "epoch": 4.628134754063072,
      "grad_norm": 0.6750428080558777,
      "learning_rate": 0.00032277556297812736,
      "loss": 3.383,
      "step": 43000
    },
    {
      "epoch": 4.628134754063072,
      "eval_accuracy": 0.37829143551217503,
      "eval_loss": 3.430992364883423,
      "eval_runtime": 184.2236,
      "eval_samples_per_second": 97.767,
      "eval_steps_per_second": 6.112,
      "step": 43000
    },
    {
      "epoch": 4.63351630610268,
      "grad_norm": 0.6586692929267883,
      "learning_rate": 0.0003224523219480659,
      "loss": 3.3956,
      "step": 43050
    },
    {
      "epoch": 4.638897858142288,
      "grad_norm": 0.6755226254463196,
      "learning_rate": 0.0003221290809180045,
      "loss": 3.4156,
      "step": 43100
    },
    {
      "epoch": 4.6442794101818965,
      "grad_norm": 0.6715269088745117,
      "learning_rate": 0.0003218058398879431,
      "loss": 3.4049,
      "step": 43150
    },
    {
      "epoch": 4.649660962221505,
      "grad_norm": 0.6735200881958008,
      "learning_rate": 0.00032148259885788163,
      "loss": 3.4079,
      "step": 43200
    },
    {
      "epoch": 4.655042514261113,
      "grad_norm": 0.681214451789856,
      "learning_rate": 0.0003211593578278203,
      "loss": 3.3966,
      "step": 43250
    },
    {
      "epoch": 4.660424066300721,
      "grad_norm": 0.626481831073761,
      "learning_rate": 0.00032083611679775887,
      "loss": 3.3928,
      "step": 43300
    },
    {
      "epoch": 4.665805618340329,
      "grad_norm": 0.6378631591796875,
      "learning_rate": 0.0003205128757676974,
      "loss": 3.3803,
      "step": 43350
    },
    {
      "epoch": 4.6711871703799375,
      "grad_norm": 0.6602470874786377,
      "learning_rate": 0.000320189634737636,
      "loss": 3.3874,
      "step": 43400
    },
    {
      "epoch": 4.676568722419546,
      "grad_norm": 0.7285584807395935,
      "learning_rate": 0.00031986639370757455,
      "loss": 3.409,
      "step": 43450
    },
    {
      "epoch": 4.681950274459154,
      "grad_norm": 0.6813101768493652,
      "learning_rate": 0.0003195431526775132,
      "loss": 3.4014,
      "step": 43500
    },
    {
      "epoch": 4.687331826498762,
      "grad_norm": 0.6950844526290894,
      "learning_rate": 0.0003192199116474518,
      "loss": 3.4217,
      "step": 43550
    },
    {
      "epoch": 4.692713378538371,
      "grad_norm": 0.6679322123527527,
      "learning_rate": 0.00031889667061739033,
      "loss": 3.3896,
      "step": 43600
    },
    {
      "epoch": 4.6980949305779784,
      "grad_norm": 0.677979588508606,
      "learning_rate": 0.0003185734295873289,
      "loss": 3.3877,
      "step": 43650
    },
    {
      "epoch": 4.703476482617587,
      "grad_norm": 0.680182695388794,
      "learning_rate": 0.00031825018855726747,
      "loss": 3.4028,
      "step": 43700
    },
    {
      "epoch": 4.7088580346571955,
      "grad_norm": 0.6772196888923645,
      "learning_rate": 0.00031792694752720606,
      "loss": 3.3909,
      "step": 43750
    },
    {
      "epoch": 4.714239586696803,
      "grad_norm": 0.679763674736023,
      "learning_rate": 0.0003176037064971447,
      "loss": 3.4077,
      "step": 43800
    },
    {
      "epoch": 4.719621138736412,
      "grad_norm": 0.7105866074562073,
      "learning_rate": 0.0003172804654670833,
      "loss": 3.4145,
      "step": 43850
    },
    {
      "epoch": 4.725002690776019,
      "grad_norm": 0.6510501503944397,
      "learning_rate": 0.00031695722443702185,
      "loss": 3.4196,
      "step": 43900
    },
    {
      "epoch": 4.730384242815628,
      "grad_norm": 0.6640481948852539,
      "learning_rate": 0.00031663398340696044,
      "loss": 3.3934,
      "step": 43950
    },
    {
      "epoch": 4.7357657948552365,
      "grad_norm": 0.634267270565033,
      "learning_rate": 0.000316310742376899,
      "loss": 3.4011,
      "step": 44000
    },
    {
      "epoch": 4.7357657948552365,
      "eval_accuracy": 0.37864488335623553,
      "eval_loss": 3.4251606464385986,
      "eval_runtime": 191.3582,
      "eval_samples_per_second": 94.122,
      "eval_steps_per_second": 5.884,
      "step": 44000
    },
    {
      "epoch": 4.741147346894844,
      "grad_norm": 0.654681384563446,
      "learning_rate": 0.0003159875013468376,
      "loss": 3.4015,
      "step": 44050
    },
    {
      "epoch": 4.746528898934453,
      "grad_norm": 0.6699573993682861,
      "learning_rate": 0.0003156642603167762,
      "loss": 3.3924,
      "step": 44100
    },
    {
      "epoch": 4.751910450974061,
      "grad_norm": 0.6511507630348206,
      "learning_rate": 0.00031534101928671476,
      "loss": 3.3933,
      "step": 44150
    },
    {
      "epoch": 4.757292003013669,
      "grad_norm": 0.6849619746208191,
      "learning_rate": 0.0003150242430772546,
      "loss": 3.403,
      "step": 44200
    },
    {
      "epoch": 4.7626735550532775,
      "grad_norm": 0.6576505899429321,
      "learning_rate": 0.00031470100204719317,
      "loss": 3.4115,
      "step": 44250
    },
    {
      "epoch": 4.768055107092886,
      "grad_norm": 0.7254562377929688,
      "learning_rate": 0.00031437776101713176,
      "loss": 3.4159,
      "step": 44300
    },
    {
      "epoch": 4.773436659132494,
      "grad_norm": 0.6806635856628418,
      "learning_rate": 0.0003140545199870703,
      "loss": 3.4089,
      "step": 44350
    },
    {
      "epoch": 4.778818211172102,
      "grad_norm": 0.7005418539047241,
      "learning_rate": 0.0003137312789570089,
      "loss": 3.4039,
      "step": 44400
    },
    {
      "epoch": 4.78419976321171,
      "grad_norm": 0.7077776789665222,
      "learning_rate": 0.00031340803792694754,
      "loss": 3.4011,
      "step": 44450
    },
    {
      "epoch": 4.7895813152513185,
      "grad_norm": 0.7974015474319458,
      "learning_rate": 0.0003130847968968861,
      "loss": 3.4066,
      "step": 44500
    },
    {
      "epoch": 4.794962867290927,
      "grad_norm": 0.6721147894859314,
      "learning_rate": 0.0003127615558668247,
      "loss": 3.3979,
      "step": 44550
    },
    {
      "epoch": 4.800344419330535,
      "grad_norm": 0.7216663360595703,
      "learning_rate": 0.0003124383148367633,
      "loss": 3.4079,
      "step": 44600
    },
    {
      "epoch": 4.805725971370143,
      "grad_norm": 0.7155595421791077,
      "learning_rate": 0.0003121150738067018,
      "loss": 3.3757,
      "step": 44650
    },
    {
      "epoch": 4.811107523409751,
      "grad_norm": 0.68537437915802,
      "learning_rate": 0.0003117918327766404,
      "loss": 3.4227,
      "step": 44700
    },
    {
      "epoch": 4.8164890754493594,
      "grad_norm": 0.73895663022995,
      "learning_rate": 0.00031146859174657906,
      "loss": 3.3958,
      "step": 44750
    },
    {
      "epoch": 4.821870627488968,
      "grad_norm": 0.6579586267471313,
      "learning_rate": 0.0003111453507165176,
      "loss": 3.4147,
      "step": 44800
    },
    {
      "epoch": 4.827252179528576,
      "grad_norm": 0.7473381161689758,
      "learning_rate": 0.0003108221096864562,
      "loss": 3.4035,
      "step": 44850
    },
    {
      "epoch": 4.832633731568184,
      "grad_norm": 0.6541971564292908,
      "learning_rate": 0.00031049886865639473,
      "loss": 3.3993,
      "step": 44900
    },
    {
      "epoch": 4.838015283607793,
      "grad_norm": 0.6812194585800171,
      "learning_rate": 0.00031017562762633333,
      "loss": 3.4003,
      "step": 44950
    },
    {
      "epoch": 4.8433968356474,
      "grad_norm": 0.6767518520355225,
      "learning_rate": 0.00030985238659627187,
      "loss": 3.3797,
      "step": 45000
    },
    {
      "epoch": 4.8433968356474,
      "eval_accuracy": 0.3792688768909541,
      "eval_loss": 3.4181933403015137,
      "eval_runtime": 216.8812,
      "eval_samples_per_second": 83.045,
      "eval_steps_per_second": 5.192,
      "step": 45000
    },
    {
      "epoch": 4.848778387687009,
      "grad_norm": 0.6793429255485535,
      "learning_rate": 0.0003095291455662105,
      "loss": 3.3936,
      "step": 45050
    },
    {
      "epoch": 4.8541599397266175,
      "grad_norm": 0.6906301975250244,
      "learning_rate": 0.0003092059045361491,
      "loss": 3.3908,
      "step": 45100
    },
    {
      "epoch": 4.859541491766225,
      "grad_norm": 0.6933501958847046,
      "learning_rate": 0.00030888266350608765,
      "loss": 3.395,
      "step": 45150
    },
    {
      "epoch": 4.864923043805834,
      "grad_norm": 0.7470901012420654,
      "learning_rate": 0.00030855942247602625,
      "loss": 3.3882,
      "step": 45200
    },
    {
      "epoch": 4.870304595845441,
      "grad_norm": 0.6697605848312378,
      "learning_rate": 0.00030823618144596484,
      "loss": 3.405,
      "step": 45250
    },
    {
      "epoch": 4.87568614788505,
      "grad_norm": 0.6537520885467529,
      "learning_rate": 0.0003079129404159035,
      "loss": 3.4044,
      "step": 45300
    },
    {
      "epoch": 4.8810676999246585,
      "grad_norm": 0.6495329141616821,
      "learning_rate": 0.00030758969938584203,
      "loss": 3.3867,
      "step": 45350
    },
    {
      "epoch": 4.886449251964266,
      "grad_norm": 0.7292621731758118,
      "learning_rate": 0.0003072664583557806,
      "loss": 3.3989,
      "step": 45400
    },
    {
      "epoch": 4.891830804003875,
      "grad_norm": 0.6901919841766357,
      "learning_rate": 0.00030694321732571917,
      "loss": 3.3985,
      "step": 45450
    },
    {
      "epoch": 4.897212356043483,
      "grad_norm": 0.701548159122467,
      "learning_rate": 0.00030661997629565776,
      "loss": 3.3808,
      "step": 45500
    },
    {
      "epoch": 4.902593908083091,
      "grad_norm": 0.6636687517166138,
      "learning_rate": 0.0003062967352655963,
      "loss": 3.4042,
      "step": 45550
    },
    {
      "epoch": 4.9079754601226995,
      "grad_norm": 0.7266779541969299,
      "learning_rate": 0.00030597349423553495,
      "loss": 3.3858,
      "step": 45600
    },
    {
      "epoch": 4.913357012162308,
      "grad_norm": 0.7061110138893127,
      "learning_rate": 0.00030565025320547355,
      "loss": 3.392,
      "step": 45650
    },
    {
      "epoch": 4.918738564201916,
      "grad_norm": 0.622070848941803,
      "learning_rate": 0.0003053270121754121,
      "loss": 3.3994,
      "step": 45700
    },
    {
      "epoch": 4.924120116241524,
      "grad_norm": 0.6944285035133362,
      "learning_rate": 0.0003050037711453507,
      "loss": 3.3907,
      "step": 45750
    },
    {
      "epoch": 4.929501668281132,
      "grad_norm": 0.6846098899841309,
      "learning_rate": 0.0003046805301152893,
      "loss": 3.391,
      "step": 45800
    },
    {
      "epoch": 4.9348832203207404,
      "grad_norm": 0.7091798782348633,
      "learning_rate": 0.0003043572890852278,
      "loss": 3.4018,
      "step": 45850
    },
    {
      "epoch": 4.940264772360349,
      "grad_norm": 0.6524344682693481,
      "learning_rate": 0.00030403404805516646,
      "loss": 3.3985,
      "step": 45900
    },
    {
      "epoch": 4.945646324399957,
      "grad_norm": 0.6928407549858093,
      "learning_rate": 0.00030371080702510506,
      "loss": 3.3867,
      "step": 45950
    },
    {
      "epoch": 4.951027876439565,
      "grad_norm": 0.7544770836830139,
      "learning_rate": 0.0003033875659950436,
      "loss": 3.4029,
      "step": 46000
    },
    {
      "epoch": 4.951027876439565,
      "eval_accuracy": 0.37985810150157207,
      "eval_loss": 3.4149184226989746,
      "eval_runtime": 199.3295,
      "eval_samples_per_second": 90.358,
      "eval_steps_per_second": 5.649,
      "step": 46000
    },
    {
      "epoch": 4.956409428479174,
      "grad_norm": 0.6892170310020447,
      "learning_rate": 0.0003030643249649822,
      "loss": 3.3988,
      "step": 46050
    },
    {
      "epoch": 4.961790980518781,
      "grad_norm": 0.6995299458503723,
      "learning_rate": 0.00030274108393492073,
      "loss": 3.3849,
      "step": 46100
    },
    {
      "epoch": 4.96717253255839,
      "grad_norm": 0.6999881267547607,
      "learning_rate": 0.00030241784290485933,
      "loss": 3.3901,
      "step": 46150
    },
    {
      "epoch": 4.9725540845979985,
      "grad_norm": 0.7442740201950073,
      "learning_rate": 0.000302094601874798,
      "loss": 3.3897,
      "step": 46200
    },
    {
      "epoch": 4.977935636637606,
      "grad_norm": 0.6908767819404602,
      "learning_rate": 0.0003017778256653378,
      "loss": 3.3825,
      "step": 46250
    },
    {
      "epoch": 4.983317188677215,
      "grad_norm": 0.6903411746025085,
      "learning_rate": 0.0003014545846352764,
      "loss": 3.3901,
      "step": 46300
    },
    {
      "epoch": 4.988698740716822,
      "grad_norm": 0.6614586710929871,
      "learning_rate": 0.0003011313436052149,
      "loss": 3.3948,
      "step": 46350
    },
    {
      "epoch": 4.994080292756431,
      "grad_norm": 0.6683346629142761,
      "learning_rate": 0.0003008081025751535,
      "loss": 3.3888,
      "step": 46400
    },
    {
      "epoch": 4.9994618447960395,
      "grad_norm": 0.6821146011352539,
      "learning_rate": 0.00030048486154509205,
      "loss": 3.3948,
      "step": 46450
    },
    {
      "epoch": 5.004843396835647,
      "grad_norm": 0.6793750524520874,
      "learning_rate": 0.0003001680853356319,
      "loss": 3.3187,
      "step": 46500
    },
    {
      "epoch": 5.010224948875256,
      "grad_norm": 0.6967900395393372,
      "learning_rate": 0.0002998448443055705,
      "loss": 3.3129,
      "step": 46550
    },
    {
      "epoch": 5.015606500914864,
      "grad_norm": 0.703115701675415,
      "learning_rate": 0.0002995216032755091,
      "loss": 3.3005,
      "step": 46600
    },
    {
      "epoch": 5.020988052954472,
      "grad_norm": 0.6616016030311584,
      "learning_rate": 0.00029919836224544765,
      "loss": 3.293,
      "step": 46650
    },
    {
      "epoch": 5.0263696049940805,
      "grad_norm": 0.6906588673591614,
      "learning_rate": 0.00029887512121538624,
      "loss": 3.3068,
      "step": 46700
    },
    {
      "epoch": 5.031751157033688,
      "grad_norm": 0.678109347820282,
      "learning_rate": 0.00029855188018532483,
      "loss": 3.3165,
      "step": 46750
    },
    {
      "epoch": 5.037132709073297,
      "grad_norm": 0.7122964859008789,
      "learning_rate": 0.00029822863915526343,
      "loss": 3.3161,
      "step": 46800
    },
    {
      "epoch": 5.042514261112905,
      "grad_norm": 0.6973577737808228,
      "learning_rate": 0.000297905398125202,
      "loss": 3.3165,
      "step": 46850
    },
    {
      "epoch": 5.047895813152513,
      "grad_norm": 0.703662097454071,
      "learning_rate": 0.00029758215709514056,
      "loss": 3.2987,
      "step": 46900
    },
    {
      "epoch": 5.0532773651921215,
      "grad_norm": 0.6552888751029968,
      "learning_rate": 0.00029725891606507916,
      "loss": 3.3216,
      "step": 46950
    },
    {
      "epoch": 5.05865891723173,
      "grad_norm": 0.7037060260772705,
      "learning_rate": 0.00029693567503501775,
      "loss": 3.3309,
      "step": 47000
    },
    {
      "epoch": 5.05865891723173,
      "eval_accuracy": 0.3797071826403977,
      "eval_loss": 3.417891263961792,
      "eval_runtime": 218.609,
      "eval_samples_per_second": 82.389,
      "eval_steps_per_second": 5.151,
      "step": 47000
    },
    {
      "epoch": 5.064040469271338,
      "grad_norm": 0.6501442193984985,
      "learning_rate": 0.00029661243400495635,
      "loss": 3.3077,
      "step": 47050
    },
    {
      "epoch": 5.069422021310946,
      "grad_norm": 0.6868728399276733,
      "learning_rate": 0.0002962891929748949,
      "loss": 3.3201,
      "step": 47100
    },
    {
      "epoch": 5.074803573350554,
      "grad_norm": 0.6883155703544617,
      "learning_rate": 0.00029596595194483354,
      "loss": 3.3179,
      "step": 47150
    },
    {
      "epoch": 5.080185125390162,
      "grad_norm": 0.6880685091018677,
      "learning_rate": 0.0002956427109147721,
      "loss": 3.3224,
      "step": 47200
    },
    {
      "epoch": 5.085566677429771,
      "grad_norm": 0.6512710452079773,
      "learning_rate": 0.00029531946988471067,
      "loss": 3.312,
      "step": 47250
    },
    {
      "epoch": 5.090948229469379,
      "grad_norm": 0.6832526326179504,
      "learning_rate": 0.00029499622885464927,
      "loss": 3.3126,
      "step": 47300
    },
    {
      "epoch": 5.096329781508987,
      "grad_norm": 0.7213174700737,
      "learning_rate": 0.00029467298782458786,
      "loss": 3.3154,
      "step": 47350
    },
    {
      "epoch": 5.101711333548596,
      "grad_norm": 0.6840560436248779,
      "learning_rate": 0.00029434974679452646,
      "loss": 3.3295,
      "step": 47400
    },
    {
      "epoch": 5.107092885588203,
      "grad_norm": 0.6937680840492249,
      "learning_rate": 0.000294026505764465,
      "loss": 3.3114,
      "step": 47450
    },
    {
      "epoch": 5.112474437627812,
      "grad_norm": 0.6524426937103271,
      "learning_rate": 0.0002937032647344036,
      "loss": 3.3178,
      "step": 47500
    },
    {
      "epoch": 5.1178559896674205,
      "grad_norm": 0.671640157699585,
      "learning_rate": 0.0002933800237043422,
      "loss": 3.3263,
      "step": 47550
    },
    {
      "epoch": 5.123237541707028,
      "grad_norm": 0.673067033290863,
      "learning_rate": 0.0002930567826742808,
      "loss": 3.3323,
      "step": 47600
    },
    {
      "epoch": 5.128619093746637,
      "grad_norm": 0.7187327742576599,
      "learning_rate": 0.0002927335416442193,
      "loss": 3.3128,
      "step": 47650
    },
    {
      "epoch": 5.134000645786244,
      "grad_norm": 0.720647394657135,
      "learning_rate": 0.00029241030061415797,
      "loss": 3.3149,
      "step": 47700
    },
    {
      "epoch": 5.139382197825853,
      "grad_norm": 0.7243995666503906,
      "learning_rate": 0.0002920870595840965,
      "loss": 3.3226,
      "step": 47750
    },
    {
      "epoch": 5.1447637498654615,
      "grad_norm": 0.7178491950035095,
      "learning_rate": 0.0002917638185540351,
      "loss": 3.3327,
      "step": 47800
    },
    {
      "epoch": 5.150145301905069,
      "grad_norm": 0.7136359810829163,
      "learning_rate": 0.0002914405775239737,
      "loss": 3.338,
      "step": 47850
    },
    {
      "epoch": 5.155526853944678,
      "grad_norm": 0.7224206328392029,
      "learning_rate": 0.0002911173364939123,
      "loss": 3.3337,
      "step": 47900
    },
    {
      "epoch": 5.160908405984286,
      "grad_norm": 0.6842013597488403,
      "learning_rate": 0.00029079409546385084,
      "loss": 3.3345,
      "step": 47950
    },
    {
      "epoch": 5.166289958023894,
      "grad_norm": 0.7609216570854187,
      "learning_rate": 0.00029047085443378943,
      "loss": 3.33,
      "step": 48000
    },
    {
      "epoch": 5.166289958023894,
      "eval_accuracy": 0.3803921304451802,
      "eval_loss": 3.4171924591064453,
      "eval_runtime": 195.0712,
      "eval_samples_per_second": 92.33,
      "eval_steps_per_second": 5.772,
      "step": 48000
    },
    {
      "epoch": 5.1716715100635025,
      "grad_norm": 0.7493017315864563,
      "learning_rate": 0.000290147613403728,
      "loss": 3.3106,
      "step": 48050
    },
    {
      "epoch": 5.17705306210311,
      "grad_norm": 0.6611179113388062,
      "learning_rate": 0.00028982437237366656,
      "loss": 3.3251,
      "step": 48100
    },
    {
      "epoch": 5.182434614142719,
      "grad_norm": 0.7239701151847839,
      "learning_rate": 0.0002895011313436052,
      "loss": 3.3156,
      "step": 48150
    },
    {
      "epoch": 5.187816166182327,
      "grad_norm": 0.6911793947219849,
      "learning_rate": 0.00028917789031354375,
      "loss": 3.3182,
      "step": 48200
    },
    {
      "epoch": 5.193197718221935,
      "grad_norm": 0.6898624300956726,
      "learning_rate": 0.00028885464928348235,
      "loss": 3.3159,
      "step": 48250
    },
    {
      "epoch": 5.198579270261543,
      "grad_norm": 0.6699836850166321,
      "learning_rate": 0.00028853140825342094,
      "loss": 3.3186,
      "step": 48300
    },
    {
      "epoch": 5.203960822301152,
      "grad_norm": 0.6687467098236084,
      "learning_rate": 0.00028820816722335954,
      "loss": 3.3348,
      "step": 48350
    },
    {
      "epoch": 5.20934237434076,
      "grad_norm": 0.6951101422309875,
      "learning_rate": 0.00028788492619329813,
      "loss": 3.3244,
      "step": 48400
    },
    {
      "epoch": 5.214723926380368,
      "grad_norm": 0.6730702519416809,
      "learning_rate": 0.0002875616851632367,
      "loss": 3.3123,
      "step": 48450
    },
    {
      "epoch": 5.220105478419977,
      "grad_norm": 0.7009641528129578,
      "learning_rate": 0.00028723844413317527,
      "loss": 3.3212,
      "step": 48500
    },
    {
      "epoch": 5.225487030459584,
      "grad_norm": 0.6732349395751953,
      "learning_rate": 0.00028691520310311386,
      "loss": 3.339,
      "step": 48550
    },
    {
      "epoch": 5.230868582499193,
      "grad_norm": 0.6993374824523926,
      "learning_rate": 0.00028659196207305246,
      "loss": 3.3404,
      "step": 48600
    },
    {
      "epoch": 5.236250134538801,
      "grad_norm": 0.6773171424865723,
      "learning_rate": 0.000286268721042991,
      "loss": 3.3415,
      "step": 48650
    },
    {
      "epoch": 5.241631686578409,
      "grad_norm": 0.6922721862792969,
      "learning_rate": 0.00028594548001292965,
      "loss": 3.3299,
      "step": 48700
    },
    {
      "epoch": 5.247013238618018,
      "grad_norm": 0.6626165509223938,
      "learning_rate": 0.0002856222389828682,
      "loss": 3.3284,
      "step": 48750
    },
    {
      "epoch": 5.252394790657625,
      "grad_norm": 0.727383553981781,
      "learning_rate": 0.0002852989979528068,
      "loss": 3.3394,
      "step": 48800
    },
    {
      "epoch": 5.257776342697234,
      "grad_norm": 0.7016910314559937,
      "learning_rate": 0.0002849757569227454,
      "loss": 3.3314,
      "step": 48850
    },
    {
      "epoch": 5.2631578947368425,
      "grad_norm": 0.6999107599258423,
      "learning_rate": 0.00028465251589268397,
      "loss": 3.3173,
      "step": 48900
    },
    {
      "epoch": 5.26853944677645,
      "grad_norm": 0.7047147750854492,
      "learning_rate": 0.0002843292748626225,
      "loss": 3.3477,
      "step": 48950
    },
    {
      "epoch": 5.273920998816059,
      "grad_norm": 0.6883307695388794,
      "learning_rate": 0.0002840060338325611,
      "loss": 3.334,
      "step": 49000
    },
    {
      "epoch": 5.273920998816059,
      "eval_accuracy": 0.38084564759891787,
      "eval_loss": 3.412295341491699,
      "eval_runtime": 201.8894,
      "eval_samples_per_second": 89.212,
      "eval_steps_per_second": 5.577,
      "step": 49000
    },
    {
      "epoch": 5.279302550855666,
      "grad_norm": 0.7405427098274231,
      "learning_rate": 0.0002836827928024997,
      "loss": 3.3207,
      "step": 49050
    },
    {
      "epoch": 5.284684102895275,
      "grad_norm": 0.7168774604797363,
      "learning_rate": 0.0002833595517724383,
      "loss": 3.3402,
      "step": 49100
    },
    {
      "epoch": 5.2900656549348835,
      "grad_norm": 0.7123434543609619,
      "learning_rate": 0.0002830363107423769,
      "loss": 3.3437,
      "step": 49150
    },
    {
      "epoch": 5.295447206974491,
      "grad_norm": 0.7152898907661438,
      "learning_rate": 0.00028271306971231543,
      "loss": 3.344,
      "step": 49200
    },
    {
      "epoch": 5.3008287590141,
      "grad_norm": 0.6967481374740601,
      "learning_rate": 0.000282389828682254,
      "loss": 3.3317,
      "step": 49250
    },
    {
      "epoch": 5.306210311053708,
      "grad_norm": 0.7272992134094238,
      "learning_rate": 0.0002820665876521926,
      "loss": 3.341,
      "step": 49300
    },
    {
      "epoch": 5.311591863093316,
      "grad_norm": 0.6962252259254456,
      "learning_rate": 0.0002817433466221312,
      "loss": 3.338,
      "step": 49350
    },
    {
      "epoch": 5.316973415132924,
      "grad_norm": 0.6720114350318909,
      "learning_rate": 0.0002814201055920698,
      "loss": 3.3422,
      "step": 49400
    },
    {
      "epoch": 5.322354967172533,
      "grad_norm": 0.6960915327072144,
      "learning_rate": 0.0002810968645620084,
      "loss": 3.3492,
      "step": 49450
    },
    {
      "epoch": 5.327736519212141,
      "grad_norm": 0.7320829629898071,
      "learning_rate": 0.00028077362353194694,
      "loss": 3.3495,
      "step": 49500
    },
    {
      "epoch": 5.333118071251749,
      "grad_norm": 0.6984890699386597,
      "learning_rate": 0.00028045038250188554,
      "loss": 3.3374,
      "step": 49550
    },
    {
      "epoch": 5.338499623291357,
      "grad_norm": 0.7055871486663818,
      "learning_rate": 0.00028012714147182413,
      "loss": 3.3458,
      "step": 49600
    },
    {
      "epoch": 5.343881175330965,
      "grad_norm": 0.766610324382782,
      "learning_rate": 0.00027980390044176273,
      "loss": 3.3486,
      "step": 49650
    },
    {
      "epoch": 5.349262727370574,
      "grad_norm": 0.736114501953125,
      "learning_rate": 0.00027948712423230253,
      "loss": 3.3562,
      "step": 49700
    },
    {
      "epoch": 5.354644279410182,
      "grad_norm": 0.681599497795105,
      "learning_rate": 0.0002791638832022411,
      "loss": 3.3344,
      "step": 49750
    },
    {
      "epoch": 5.36002583144979,
      "grad_norm": 0.7354723811149597,
      "learning_rate": 0.0002788406421721797,
      "loss": 3.3276,
      "step": 49800
    },
    {
      "epoch": 5.365407383489399,
      "grad_norm": 0.6823558211326599,
      "learning_rate": 0.00027851740114211826,
      "loss": 3.3348,
      "step": 49850
    },
    {
      "epoch": 5.370788935529006,
      "grad_norm": 0.6879205703735352,
      "learning_rate": 0.00027819416011205686,
      "loss": 3.3235,
      "step": 49900
    },
    {
      "epoch": 5.376170487568615,
      "grad_norm": 0.7181028723716736,
      "learning_rate": 0.00027787091908199545,
      "loss": 3.3485,
      "step": 49950
    },
    {
      "epoch": 5.3815520396082235,
      "grad_norm": 0.7358317971229553,
      "learning_rate": 0.00027754767805193405,
      "loss": 3.3359,
      "step": 50000
    },
    {
      "epoch": 5.3815520396082235,
      "eval_accuracy": 0.3811460728337249,
      "eval_loss": 3.407122850418091,
      "eval_runtime": 189.3116,
      "eval_samples_per_second": 95.139,
      "eval_steps_per_second": 5.948,
      "step": 50000
    },
    {
      "epoch": 5.386933591647831,
      "grad_norm": 0.6749621033668518,
      "learning_rate": 0.0002772244370218726,
      "loss": 3.3595,
      "step": 50050
    },
    {
      "epoch": 5.39231514368744,
      "grad_norm": 0.6878547072410583,
      "learning_rate": 0.0002769011959918112,
      "loss": 3.34,
      "step": 50100
    },
    {
      "epoch": 5.397696695727047,
      "grad_norm": 0.6463252902030945,
      "learning_rate": 0.0002765779549617498,
      "loss": 3.3551,
      "step": 50150
    },
    {
      "epoch": 5.403078247766656,
      "grad_norm": 0.6702800393104553,
      "learning_rate": 0.0002762547139316884,
      "loss": 3.327,
      "step": 50200
    },
    {
      "epoch": 5.4084597998062645,
      "grad_norm": 0.694614052772522,
      "learning_rate": 0.00027593147290162697,
      "loss": 3.3572,
      "step": 50250
    },
    {
      "epoch": 5.413841351845872,
      "grad_norm": 0.6810118556022644,
      "learning_rate": 0.0002756082318715655,
      "loss": 3.3297,
      "step": 50300
    },
    {
      "epoch": 5.419222903885481,
      "grad_norm": 0.673312783241272,
      "learning_rate": 0.00027528499084150416,
      "loss": 3.3311,
      "step": 50350
    },
    {
      "epoch": 5.424604455925088,
      "grad_norm": 0.7154831290245056,
      "learning_rate": 0.0002749617498114427,
      "loss": 3.3256,
      "step": 50400
    },
    {
      "epoch": 5.429986007964697,
      "grad_norm": 0.6521028280258179,
      "learning_rate": 0.0002746385087813813,
      "loss": 3.3506,
      "step": 50450
    },
    {
      "epoch": 5.435367560004305,
      "grad_norm": 0.7348833680152893,
      "learning_rate": 0.0002743152677513199,
      "loss": 3.338,
      "step": 50500
    },
    {
      "epoch": 5.440749112043913,
      "grad_norm": 0.6742364764213562,
      "learning_rate": 0.0002739920267212585,
      "loss": 3.3583,
      "step": 50550
    },
    {
      "epoch": 5.446130664083522,
      "grad_norm": 0.6520490646362305,
      "learning_rate": 0.000273668785691197,
      "loss": 3.3304,
      "step": 50600
    },
    {
      "epoch": 5.45151221612313,
      "grad_norm": 0.7477188110351562,
      "learning_rate": 0.0002733455446611356,
      "loss": 3.343,
      "step": 50650
    },
    {
      "epoch": 5.456893768162738,
      "grad_norm": 0.6800089478492737,
      "learning_rate": 0.0002730223036310742,
      "loss": 3.3289,
      "step": 50700
    },
    {
      "epoch": 5.462275320202346,
      "grad_norm": 0.6529966592788696,
      "learning_rate": 0.0002726990626010128,
      "loss": 3.3394,
      "step": 50750
    },
    {
      "epoch": 5.467656872241955,
      "grad_norm": 0.7075954079627991,
      "learning_rate": 0.0002723758215709514,
      "loss": 3.3414,
      "step": 50800
    },
    {
      "epoch": 5.473038424281563,
      "grad_norm": 0.6758164763450623,
      "learning_rate": 0.00027205258054088994,
      "loss": 3.3334,
      "step": 50850
    },
    {
      "epoch": 5.478419976321171,
      "grad_norm": 0.7001152634620667,
      "learning_rate": 0.00027172933951082854,
      "loss": 3.3304,
      "step": 50900
    },
    {
      "epoch": 5.483801528360779,
      "grad_norm": 0.6786561608314514,
      "learning_rate": 0.00027140609848076713,
      "loss": 3.3415,
      "step": 50950
    },
    {
      "epoch": 5.489183080400387,
      "grad_norm": 0.6682949066162109,
      "learning_rate": 0.0002710828574507057,
      "loss": 3.3324,
      "step": 51000
    },
    {
      "epoch": 5.489183080400387,
      "eval_accuracy": 0.38178429989674717,
      "eval_loss": 3.4022624492645264,
      "eval_runtime": 190.4353,
      "eval_samples_per_second": 94.578,
      "eval_steps_per_second": 5.913,
      "step": 51000
    },
    {
      "epoch": 5.494564632439996,
      "grad_norm": 0.6904833912849426,
      "learning_rate": 0.00027075961642064427,
      "loss": 3.3337,
      "step": 51050
    },
    {
      "epoch": 5.499946184479604,
      "grad_norm": 0.6703086495399475,
      "learning_rate": 0.0002704363753905829,
      "loss": 3.3603,
      "step": 51100
    },
    {
      "epoch": 5.505327736519212,
      "grad_norm": 0.7118019461631775,
      "learning_rate": 0.00027011313436052145,
      "loss": 3.3503,
      "step": 51150
    },
    {
      "epoch": 5.510709288558821,
      "grad_norm": 0.6811361908912659,
      "learning_rate": 0.00026978989333046005,
      "loss": 3.3468,
      "step": 51200
    },
    {
      "epoch": 5.516090840598428,
      "grad_norm": 1.386173963546753,
      "learning_rate": 0.00026946665230039864,
      "loss": 3.3466,
      "step": 51250
    },
    {
      "epoch": 5.521472392638037,
      "grad_norm": 0.6784133911132812,
      "learning_rate": 0.00026914341127033724,
      "loss": 3.3445,
      "step": 51300
    },
    {
      "epoch": 5.5268539446776455,
      "grad_norm": 0.7018135190010071,
      "learning_rate": 0.00026882017024027583,
      "loss": 3.3294,
      "step": 51350
    },
    {
      "epoch": 5.532235496717253,
      "grad_norm": 0.6830986738204956,
      "learning_rate": 0.0002684969292102144,
      "loss": 3.3307,
      "step": 51400
    },
    {
      "epoch": 5.537617048756862,
      "grad_norm": 0.7251037359237671,
      "learning_rate": 0.00026817368818015297,
      "loss": 3.3338,
      "step": 51450
    },
    {
      "epoch": 5.542998600796469,
      "grad_norm": 0.7147461175918579,
      "learning_rate": 0.00026785044715009156,
      "loss": 3.3375,
      "step": 51500
    },
    {
      "epoch": 5.548380152836078,
      "grad_norm": 0.7019665837287903,
      "learning_rate": 0.00026752720612003016,
      "loss": 3.3413,
      "step": 51550
    },
    {
      "epoch": 5.553761704875686,
      "grad_norm": 0.7293707728385925,
      "learning_rate": 0.0002672039650899687,
      "loss": 3.3278,
      "step": 51600
    },
    {
      "epoch": 5.559143256915294,
      "grad_norm": 0.6743173003196716,
      "learning_rate": 0.00026688072405990735,
      "loss": 3.3232,
      "step": 51650
    },
    {
      "epoch": 5.564524808954903,
      "grad_norm": 0.7335731387138367,
      "learning_rate": 0.0002665574830298459,
      "loss": 3.3468,
      "step": 51700
    },
    {
      "epoch": 5.569906360994511,
      "grad_norm": 0.7518755197525024,
      "learning_rate": 0.0002662342419997845,
      "loss": 3.3277,
      "step": 51750
    },
    {
      "epoch": 5.575287913034119,
      "grad_norm": 0.7123223543167114,
      "learning_rate": 0.0002659174657903243,
      "loss": 3.3582,
      "step": 51800
    },
    {
      "epoch": 5.580669465073727,
      "grad_norm": 0.6899611949920654,
      "learning_rate": 0.0002655942247602629,
      "loss": 3.3433,
      "step": 51850
    },
    {
      "epoch": 5.586051017113336,
      "grad_norm": 0.7059025764465332,
      "learning_rate": 0.0002652709837302015,
      "loss": 3.349,
      "step": 51900
    },
    {
      "epoch": 5.591432569152944,
      "grad_norm": 0.7855048179626465,
      "learning_rate": 0.00026494774270014,
      "loss": 3.3309,
      "step": 51950
    },
    {
      "epoch": 5.596814121192552,
      "grad_norm": 0.692093551158905,
      "learning_rate": 0.00026462450167007867,
      "loss": 3.3505,
      "step": 52000
    },
    {
      "epoch": 5.596814121192552,
      "eval_accuracy": 0.3822000058455254,
      "eval_loss": 3.399010419845581,
      "eval_runtime": 212.0432,
      "eval_samples_per_second": 84.94,
      "eval_steps_per_second": 5.31,
      "step": 52000
    },
    {
      "epoch": 5.60219567323216,
      "grad_norm": 0.7186334729194641,
      "learning_rate": 0.0002643012606400172,
      "loss": 3.3551,
      "step": 52050
    },
    {
      "epoch": 5.607577225271768,
      "grad_norm": 0.7129858732223511,
      "learning_rate": 0.0002639780196099558,
      "loss": 3.3346,
      "step": 52100
    },
    {
      "epoch": 5.612958777311377,
      "grad_norm": 0.6894886493682861,
      "learning_rate": 0.0002636547785798944,
      "loss": 3.3547,
      "step": 52150
    },
    {
      "epoch": 5.618340329350985,
      "grad_norm": 0.7041341066360474,
      "learning_rate": 0.000263331537549833,
      "loss": 3.3471,
      "step": 52200
    },
    {
      "epoch": 5.623721881390593,
      "grad_norm": 0.7379045486450195,
      "learning_rate": 0.00026300829651977153,
      "loss": 3.3354,
      "step": 52250
    },
    {
      "epoch": 5.629103433430201,
      "grad_norm": 0.7196996212005615,
      "learning_rate": 0.0002626850554897101,
      "loss": 3.3316,
      "step": 52300
    },
    {
      "epoch": 5.634484985469809,
      "grad_norm": 0.7082961797714233,
      "learning_rate": 0.0002623618144596487,
      "loss": 3.3379,
      "step": 52350
    },
    {
      "epoch": 5.639866537509418,
      "grad_norm": 0.7027478218078613,
      "learning_rate": 0.0002620385734295873,
      "loss": 3.3535,
      "step": 52400
    },
    {
      "epoch": 5.645248089549026,
      "grad_norm": 0.70321124792099,
      "learning_rate": 0.0002617153323995259,
      "loss": 3.3234,
      "step": 52450
    },
    {
      "epoch": 5.650629641588634,
      "grad_norm": 0.6914188265800476,
      "learning_rate": 0.00026139209136946445,
      "loss": 3.3589,
      "step": 52500
    },
    {
      "epoch": 5.656011193628243,
      "grad_norm": 0.7002469897270203,
      "learning_rate": 0.00026106885033940305,
      "loss": 3.3486,
      "step": 52550
    },
    {
      "epoch": 5.66139274566785,
      "grad_norm": 0.6917188763618469,
      "learning_rate": 0.00026074560930934164,
      "loss": 3.327,
      "step": 52600
    },
    {
      "epoch": 5.666774297707459,
      "grad_norm": 0.7423471212387085,
      "learning_rate": 0.00026042236827928024,
      "loss": 3.3578,
      "step": 52650
    },
    {
      "epoch": 5.672155849747067,
      "grad_norm": 0.6731488704681396,
      "learning_rate": 0.0002600991272492188,
      "loss": 3.3521,
      "step": 52700
    },
    {
      "epoch": 5.677537401786675,
      "grad_norm": 0.7297781705856323,
      "learning_rate": 0.0002597758862191574,
      "loss": 3.3338,
      "step": 52750
    },
    {
      "epoch": 5.682918953826284,
      "grad_norm": 0.7086893320083618,
      "learning_rate": 0.00025945264518909596,
      "loss": 3.3346,
      "step": 52800
    },
    {
      "epoch": 5.688300505865891,
      "grad_norm": 0.7108243107795715,
      "learning_rate": 0.00025912940415903456,
      "loss": 3.3342,
      "step": 52850
    },
    {
      "epoch": 5.6936820579055,
      "grad_norm": 0.7097220420837402,
      "learning_rate": 0.00025880616312897315,
      "loss": 3.3497,
      "step": 52900
    },
    {
      "epoch": 5.699063609945108,
      "grad_norm": 0.7394960522651672,
      "learning_rate": 0.00025848292209891175,
      "loss": 3.3371,
      "step": 52950
    },
    {
      "epoch": 5.704445161984716,
      "grad_norm": 0.6763904094696045,
      "learning_rate": 0.00025815968106885034,
      "loss": 3.3439,
      "step": 53000
    },
    {
      "epoch": 5.704445161984716,
      "eval_accuracy": 0.38251705497216804,
      "eval_loss": 3.3959052562713623,
      "eval_runtime": 200.1646,
      "eval_samples_per_second": 89.981,
      "eval_steps_per_second": 5.625,
      "step": 53000
    },
    {
      "epoch": 5.709826714024325,
      "grad_norm": 0.7205132246017456,
      "learning_rate": 0.0002578364400387889,
      "loss": 3.3496,
      "step": 53050
    },
    {
      "epoch": 5.715208266063933,
      "grad_norm": 0.7770679593086243,
      "learning_rate": 0.0002575131990087275,
      "loss": 3.3384,
      "step": 53100
    },
    {
      "epoch": 5.720589818103541,
      "grad_norm": 0.7161393761634827,
      "learning_rate": 0.0002571899579786661,
      "loss": 3.3391,
      "step": 53150
    },
    {
      "epoch": 5.725971370143149,
      "grad_norm": 0.7258470058441162,
      "learning_rate": 0.00025686671694860467,
      "loss": 3.3439,
      "step": 53200
    },
    {
      "epoch": 5.731352922182758,
      "grad_norm": 0.7805595993995667,
      "learning_rate": 0.0002565434759185432,
      "loss": 3.3468,
      "step": 53250
    },
    {
      "epoch": 5.736734474222366,
      "grad_norm": 0.6672064661979675,
      "learning_rate": 0.00025622023488848186,
      "loss": 3.352,
      "step": 53300
    },
    {
      "epoch": 5.742116026261974,
      "grad_norm": 0.701865017414093,
      "learning_rate": 0.0002558969938584204,
      "loss": 3.3378,
      "step": 53350
    },
    {
      "epoch": 5.747497578301582,
      "grad_norm": 0.7268505692481995,
      "learning_rate": 0.000255573752828359,
      "loss": 3.3361,
      "step": 53400
    },
    {
      "epoch": 5.75287913034119,
      "grad_norm": 0.731370747089386,
      "learning_rate": 0.0002552505117982976,
      "loss": 3.3658,
      "step": 53450
    },
    {
      "epoch": 5.758260682380799,
      "grad_norm": 0.7171875238418579,
      "learning_rate": 0.0002549272707682362,
      "loss": 3.3372,
      "step": 53500
    },
    {
      "epoch": 5.763642234420407,
      "grad_norm": 0.6759555339813232,
      "learning_rate": 0.0002546040297381747,
      "loss": 3.3422,
      "step": 53550
    },
    {
      "epoch": 5.769023786460015,
      "grad_norm": 0.7288383841514587,
      "learning_rate": 0.0002542807887081133,
      "loss": 3.3552,
      "step": 53600
    },
    {
      "epoch": 5.774405338499624,
      "grad_norm": 0.7491167783737183,
      "learning_rate": 0.0002539575476780519,
      "loss": 3.3513,
      "step": 53650
    },
    {
      "epoch": 5.779786890539231,
      "grad_norm": 0.7122647762298584,
      "learning_rate": 0.0002536343066479905,
      "loss": 3.344,
      "step": 53700
    },
    {
      "epoch": 5.78516844257884,
      "grad_norm": 0.7158402800559998,
      "learning_rate": 0.0002533110656179291,
      "loss": 3.3364,
      "step": 53750
    },
    {
      "epoch": 5.790549994618448,
      "grad_norm": 0.7261003255844116,
      "learning_rate": 0.00025298782458786764,
      "loss": 3.3494,
      "step": 53800
    },
    {
      "epoch": 5.795931546658056,
      "grad_norm": 0.7443234920501709,
      "learning_rate": 0.00025266458355780624,
      "loss": 3.3425,
      "step": 53850
    },
    {
      "epoch": 5.801313098697665,
      "grad_norm": 0.7296776175498962,
      "learning_rate": 0.00025234134252774483,
      "loss": 3.3552,
      "step": 53900
    },
    {
      "epoch": 5.806694650737272,
      "grad_norm": 0.7506049275398254,
      "learning_rate": 0.0002520181014976834,
      "loss": 3.3403,
      "step": 53950
    },
    {
      "epoch": 5.812076202776881,
      "grad_norm": 0.6987310647964478,
      "learning_rate": 0.000251694860467622,
      "loss": 3.3471,
      "step": 54000
    },
    {
      "epoch": 5.812076202776881,
      "eval_accuracy": 0.3829692682912519,
      "eval_loss": 3.3907604217529297,
      "eval_runtime": 186.7939,
      "eval_samples_per_second": 96.422,
      "eval_steps_per_second": 6.028,
      "step": 54000
    },
    {
      "epoch": 5.817457754816489,
      "grad_norm": 0.6977421045303345,
      "learning_rate": 0.0002513716194375606,
      "loss": 3.3415,
      "step": 54050
    },
    {
      "epoch": 5.822839306856097,
      "grad_norm": 0.6712203621864319,
      "learning_rate": 0.00025104837840749916,
      "loss": 3.3354,
      "step": 54100
    },
    {
      "epoch": 5.828220858895706,
      "grad_norm": 0.6803415417671204,
      "learning_rate": 0.00025072513737743775,
      "loss": 3.3362,
      "step": 54150
    },
    {
      "epoch": 5.833602410935313,
      "grad_norm": 0.682086169719696,
      "learning_rate": 0.00025040189634737634,
      "loss": 3.3425,
      "step": 54200
    },
    {
      "epoch": 5.838983962974922,
      "grad_norm": 0.6937381625175476,
      "learning_rate": 0.00025007865531731494,
      "loss": 3.3318,
      "step": 54250
    },
    {
      "epoch": 5.84436551501453,
      "grad_norm": 0.7256398797035217,
      "learning_rate": 0.00024975541428725353,
      "loss": 3.351,
      "step": 54300
    },
    {
      "epoch": 5.849747067054138,
      "grad_norm": 0.7517046332359314,
      "learning_rate": 0.0002494386380777933,
      "loss": 3.3315,
      "step": 54350
    },
    {
      "epoch": 5.855128619093747,
      "grad_norm": 0.83865886926651,
      "learning_rate": 0.00024911539704773193,
      "loss": 3.335,
      "step": 54400
    },
    {
      "epoch": 5.860510171133355,
      "grad_norm": 0.7318394184112549,
      "learning_rate": 0.0002487921560176705,
      "loss": 3.3414,
      "step": 54450
    },
    {
      "epoch": 5.865891723172963,
      "grad_norm": 0.7068241238594055,
      "learning_rate": 0.00024846891498760907,
      "loss": 3.3566,
      "step": 54500
    },
    {
      "epoch": 5.871273275212571,
      "grad_norm": 0.705195426940918,
      "learning_rate": 0.00024814567395754766,
      "loss": 3.3389,
      "step": 54550
    },
    {
      "epoch": 5.87665482725218,
      "grad_norm": 0.6855528354644775,
      "learning_rate": 0.00024782243292748626,
      "loss": 3.3314,
      "step": 54600
    },
    {
      "epoch": 5.882036379291788,
      "grad_norm": 0.6834759712219238,
      "learning_rate": 0.0002474991918974248,
      "loss": 3.3692,
      "step": 54650
    },
    {
      "epoch": 5.887417931331396,
      "grad_norm": 0.7025681734085083,
      "learning_rate": 0.0002471759508673634,
      "loss": 3.3292,
      "step": 54700
    },
    {
      "epoch": 5.892799483371004,
      "grad_norm": 0.7142512202262878,
      "learning_rate": 0.000246852709837302,
      "loss": 3.3287,
      "step": 54750
    },
    {
      "epoch": 5.898181035410612,
      "grad_norm": 0.7166208028793335,
      "learning_rate": 0.0002465294688072406,
      "loss": 3.3301,
      "step": 54800
    },
    {
      "epoch": 5.903562587450221,
      "grad_norm": 0.8234772682189941,
      "learning_rate": 0.0002462062277771792,
      "loss": 3.3377,
      "step": 54850
    },
    {
      "epoch": 5.9089441394898286,
      "grad_norm": 0.7680366039276123,
      "learning_rate": 0.0002458829867471177,
      "loss": 3.3289,
      "step": 54900
    },
    {
      "epoch": 5.914325691529437,
      "grad_norm": 0.747733473777771,
      "learning_rate": 0.00024555974571705637,
      "loss": 3.3494,
      "step": 54950
    },
    {
      "epoch": 5.919707243569046,
      "grad_norm": 0.6918068528175354,
      "learning_rate": 0.0002452365046869949,
      "loss": 3.3394,
      "step": 55000
    },
    {
      "epoch": 5.919707243569046,
      "eval_accuracy": 0.3834948223096105,
      "eval_loss": 3.385463237762451,
      "eval_runtime": 216.4848,
      "eval_samples_per_second": 83.198,
      "eval_steps_per_second": 5.201,
      "step": 55000
    },
    {
      "epoch": 5.925088795608653,
      "grad_norm": 0.705939769744873,
      "learning_rate": 0.0002449132636569335,
      "loss": 3.3274,
      "step": 55050
    },
    {
      "epoch": 5.930470347648262,
      "grad_norm": 0.6875008940696716,
      "learning_rate": 0.0002445900226268721,
      "loss": 3.3251,
      "step": 55100
    },
    {
      "epoch": 5.93585189968787,
      "grad_norm": 0.7483250498771667,
      "learning_rate": 0.0002442667815968107,
      "loss": 3.3437,
      "step": 55150
    },
    {
      "epoch": 5.941233451727478,
      "grad_norm": 0.7351670861244202,
      "learning_rate": 0.00024394354056674923,
      "loss": 3.3321,
      "step": 55200
    },
    {
      "epoch": 5.946615003767087,
      "grad_norm": 0.7016918659210205,
      "learning_rate": 0.00024362029953668785,
      "loss": 3.347,
      "step": 55250
    },
    {
      "epoch": 5.951996555806694,
      "grad_norm": 0.742716908454895,
      "learning_rate": 0.00024329705850662642,
      "loss": 3.3351,
      "step": 55300
    },
    {
      "epoch": 5.957378107846303,
      "grad_norm": 0.7227046489715576,
      "learning_rate": 0.000242973817476565,
      "loss": 3.3484,
      "step": 55350
    },
    {
      "epoch": 5.962759659885911,
      "grad_norm": 0.6967488527297974,
      "learning_rate": 0.00024265057644650358,
      "loss": 3.3453,
      "step": 55400
    },
    {
      "epoch": 5.968141211925519,
      "grad_norm": 0.7211737036705017,
      "learning_rate": 0.00024232733541644218,
      "loss": 3.3431,
      "step": 55450
    },
    {
      "epoch": 5.973522763965128,
      "grad_norm": 0.7636567950248718,
      "learning_rate": 0.00024200409438638075,
      "loss": 3.3417,
      "step": 55500
    },
    {
      "epoch": 5.978904316004736,
      "grad_norm": 0.7380962371826172,
      "learning_rate": 0.00024168085335631934,
      "loss": 3.3467,
      "step": 55550
    },
    {
      "epoch": 5.984285868044344,
      "grad_norm": 0.7630906701087952,
      "learning_rate": 0.0002413576123262579,
      "loss": 3.336,
      "step": 55600
    },
    {
      "epoch": 5.989667420083952,
      "grad_norm": 0.7228858470916748,
      "learning_rate": 0.0002410343712961965,
      "loss": 3.325,
      "step": 55650
    },
    {
      "epoch": 5.995048972123561,
      "grad_norm": 0.7207238078117371,
      "learning_rate": 0.0002407111302661351,
      "loss": 3.3317,
      "step": 55700
    },
    {
      "epoch": 6.000430524163169,
      "grad_norm": 0.6902097463607788,
      "learning_rate": 0.00024038788923607367,
      "loss": 3.3281,
      "step": 55750
    },
    {
      "epoch": 6.005812076202777,
      "grad_norm": 0.6920804977416992,
      "learning_rate": 0.0002400646482060123,
      "loss": 3.2285,
      "step": 55800
    },
    {
      "epoch": 6.011193628242385,
      "grad_norm": 0.7115839123725891,
      "learning_rate": 0.00023974140717595085,
      "loss": 3.2658,
      "step": 55850
    },
    {
      "epoch": 6.016575180281993,
      "grad_norm": 0.7384124398231506,
      "learning_rate": 0.00023941816614588942,
      "loss": 3.2385,
      "step": 55900
    },
    {
      "epoch": 6.021956732321602,
      "grad_norm": 0.741339921951294,
      "learning_rate": 0.00023909492511582802,
      "loss": 3.2545,
      "step": 55950
    },
    {
      "epoch": 6.0273382843612096,
      "grad_norm": 0.6965171694755554,
      "learning_rate": 0.0002387716840857666,
      "loss": 3.2478,
      "step": 56000
    },
    {
      "epoch": 6.0273382843612096,
      "eval_accuracy": 0.3833505312745929,
      "eval_loss": 3.388998508453369,
      "eval_runtime": 219.5095,
      "eval_samples_per_second": 82.051,
      "eval_steps_per_second": 5.13,
      "step": 56000
    },
    {
      "epoch": 6.032719836400818,
      "grad_norm": 0.7261655926704407,
      "learning_rate": 0.00023844844305570518,
      "loss": 3.263,
      "step": 56050
    },
    {
      "epoch": 6.038101388440427,
      "grad_norm": 0.7194091081619263,
      "learning_rate": 0.00023812520202564377,
      "loss": 3.2396,
      "step": 56100
    },
    {
      "epoch": 6.043482940480034,
      "grad_norm": 0.7391347885131836,
      "learning_rate": 0.00023780196099558234,
      "loss": 3.247,
      "step": 56150
    },
    {
      "epoch": 6.048864492519643,
      "grad_norm": 0.7657032608985901,
      "learning_rate": 0.00023747871996552094,
      "loss": 3.2663,
      "step": 56200
    },
    {
      "epoch": 6.0542460445592505,
      "grad_norm": 0.7573240995407104,
      "learning_rate": 0.00023715547893545953,
      "loss": 3.2659,
      "step": 56250
    },
    {
      "epoch": 6.059627596598859,
      "grad_norm": 0.720176100730896,
      "learning_rate": 0.0002368322379053981,
      "loss": 3.2576,
      "step": 56300
    },
    {
      "epoch": 6.065009148638468,
      "grad_norm": 0.8777809143066406,
      "learning_rate": 0.00023650899687533667,
      "loss": 3.2292,
      "step": 56350
    },
    {
      "epoch": 6.070390700678075,
      "grad_norm": 0.7772096395492554,
      "learning_rate": 0.0002361857558452753,
      "loss": 3.2749,
      "step": 56400
    },
    {
      "epoch": 6.075772252717684,
      "grad_norm": 0.7186575531959534,
      "learning_rate": 0.00023586251481521386,
      "loss": 3.2681,
      "step": 56450
    },
    {
      "epoch": 6.081153804757292,
      "grad_norm": 0.7018833756446838,
      "learning_rate": 0.00023553927378515242,
      "loss": 3.2653,
      "step": 56500
    },
    {
      "epoch": 6.0865353567969,
      "grad_norm": 0.7191780805587769,
      "learning_rate": 0.00023521603275509104,
      "loss": 3.2707,
      "step": 56550
    },
    {
      "epoch": 6.091916908836509,
      "grad_norm": 0.74022376537323,
      "learning_rate": 0.0002348927917250296,
      "loss": 3.2687,
      "step": 56600
    },
    {
      "epoch": 6.097298460876116,
      "grad_norm": 0.7290888428688049,
      "learning_rate": 0.00023456955069496818,
      "loss": 3.2621,
      "step": 56650
    },
    {
      "epoch": 6.102680012915725,
      "grad_norm": 0.6947687864303589,
      "learning_rate": 0.00023424630966490677,
      "loss": 3.2706,
      "step": 56700
    },
    {
      "epoch": 6.108061564955333,
      "grad_norm": 0.7154489755630493,
      "learning_rate": 0.00023392306863484537,
      "loss": 3.2658,
      "step": 56750
    },
    {
      "epoch": 6.113443116994941,
      "grad_norm": 0.7476992607116699,
      "learning_rate": 0.00023359982760478396,
      "loss": 3.2881,
      "step": 56800
    },
    {
      "epoch": 6.11882466903455,
      "grad_norm": 0.6735149025917053,
      "learning_rate": 0.00023327658657472253,
      "loss": 3.2614,
      "step": 56850
    },
    {
      "epoch": 6.124206221074158,
      "grad_norm": 0.7266242504119873,
      "learning_rate": 0.0002329533455446611,
      "loss": 3.2754,
      "step": 56900
    },
    {
      "epoch": 6.129587773113766,
      "grad_norm": 0.7465396523475647,
      "learning_rate": 0.00023263010451459972,
      "loss": 3.273,
      "step": 56950
    },
    {
      "epoch": 6.134969325153374,
      "grad_norm": 0.7545667290687561,
      "learning_rate": 0.0002323068634845383,
      "loss": 3.2706,
      "step": 57000
    },
    {
      "epoch": 6.134969325153374,
      "eval_accuracy": 0.38392758676177563,
      "eval_loss": 3.3891730308532715,
      "eval_runtime": 198.3881,
      "eval_samples_per_second": 90.787,
      "eval_steps_per_second": 5.676,
      "step": 57000
    },
    {
      "epoch": 6.140350877192983,
      "grad_norm": 0.773764967918396,
      "learning_rate": 0.00023198362245447686,
      "loss": 3.286,
      "step": 57050
    },
    {
      "epoch": 6.1457324292325906,
      "grad_norm": 0.7388854026794434,
      "learning_rate": 0.00023166038142441548,
      "loss": 3.2861,
      "step": 57100
    },
    {
      "epoch": 6.151113981272199,
      "grad_norm": 0.7281585931777954,
      "learning_rate": 0.00023133714039435404,
      "loss": 3.2778,
      "step": 57150
    },
    {
      "epoch": 6.156495533311807,
      "grad_norm": 0.7503560781478882,
      "learning_rate": 0.0002310138993642926,
      "loss": 3.2712,
      "step": 57200
    },
    {
      "epoch": 6.161877085351415,
      "grad_norm": 0.7558592557907104,
      "learning_rate": 0.0002306906583342312,
      "loss": 3.2747,
      "step": 57250
    },
    {
      "epoch": 6.167258637391024,
      "grad_norm": 0.7067055106163025,
      "learning_rate": 0.000230373882124771,
      "loss": 3.2749,
      "step": 57300
    },
    {
      "epoch": 6.1726401894306315,
      "grad_norm": 0.737853467464447,
      "learning_rate": 0.0002300506410947096,
      "loss": 3.2717,
      "step": 57350
    },
    {
      "epoch": 6.17802174147024,
      "grad_norm": 0.7693945169448853,
      "learning_rate": 0.00022972740006464818,
      "loss": 3.2744,
      "step": 57400
    },
    {
      "epoch": 6.183403293509849,
      "grad_norm": 0.7323774695396423,
      "learning_rate": 0.00022940415903458674,
      "loss": 3.2491,
      "step": 57450
    },
    {
      "epoch": 6.188784845549456,
      "grad_norm": 0.7609051465988159,
      "learning_rate": 0.00022908091800452537,
      "loss": 3.2792,
      "step": 57500
    },
    {
      "epoch": 6.194166397589065,
      "grad_norm": 0.7921963334083557,
      "learning_rate": 0.00022875767697446393,
      "loss": 3.2684,
      "step": 57550
    },
    {
      "epoch": 6.1995479496286725,
      "grad_norm": 0.750731885433197,
      "learning_rate": 0.00022843443594440253,
      "loss": 3.2869,
      "step": 57600
    },
    {
      "epoch": 6.204929501668281,
      "grad_norm": 0.7311643362045288,
      "learning_rate": 0.00022811119491434112,
      "loss": 3.2666,
      "step": 57650
    },
    {
      "epoch": 6.21031105370789,
      "grad_norm": 0.7532967329025269,
      "learning_rate": 0.0002277879538842797,
      "loss": 3.2728,
      "step": 57700
    },
    {
      "epoch": 6.215692605747497,
      "grad_norm": 0.7333381175994873,
      "learning_rate": 0.00022746471285421828,
      "loss": 3.2788,
      "step": 57750
    },
    {
      "epoch": 6.221074157787106,
      "grad_norm": 0.7154975533485413,
      "learning_rate": 0.00022714147182415685,
      "loss": 3.2873,
      "step": 57800
    },
    {
      "epoch": 6.226455709826714,
      "grad_norm": 0.7327954173088074,
      "learning_rate": 0.00022681823079409545,
      "loss": 3.2677,
      "step": 57850
    },
    {
      "epoch": 6.231837261866322,
      "grad_norm": 0.715451180934906,
      "learning_rate": 0.00022649498976403404,
      "loss": 3.2811,
      "step": 57900
    },
    {
      "epoch": 6.237218813905931,
      "grad_norm": 0.7642385363578796,
      "learning_rate": 0.0002261717487339726,
      "loss": 3.2857,
      "step": 57950
    },
    {
      "epoch": 6.242600365945538,
      "grad_norm": 0.7612628936767578,
      "learning_rate": 0.00022584850770391118,
      "loss": 3.2703,
      "step": 58000
    },
    {
      "epoch": 6.242600365945538,
      "eval_accuracy": 0.38411848988566566,
      "eval_loss": 3.3868725299835205,
      "eval_runtime": 203.836,
      "eval_samples_per_second": 88.36,
      "eval_steps_per_second": 5.524,
      "step": 58000
    },
    {
      "epoch": 6.247981917985147,
      "grad_norm": 0.7095159292221069,
      "learning_rate": 0.0002255252666738498,
      "loss": 3.2841,
      "step": 58050
    },
    {
      "epoch": 6.253363470024755,
      "grad_norm": 0.7202694416046143,
      "learning_rate": 0.00022520202564378837,
      "loss": 3.2742,
      "step": 58100
    },
    {
      "epoch": 6.258745022064363,
      "grad_norm": 0.7857133150100708,
      "learning_rate": 0.00022487878461372693,
      "loss": 3.2833,
      "step": 58150
    },
    {
      "epoch": 6.264126574103972,
      "grad_norm": 0.7608604431152344,
      "learning_rate": 0.00022455554358366555,
      "loss": 3.2725,
      "step": 58200
    },
    {
      "epoch": 6.26950812614358,
      "grad_norm": 0.7771002650260925,
      "learning_rate": 0.00022423230255360412,
      "loss": 3.289,
      "step": 58250
    },
    {
      "epoch": 6.274889678183188,
      "grad_norm": 0.6962699890136719,
      "learning_rate": 0.0002239090615235427,
      "loss": 3.2659,
      "step": 58300
    },
    {
      "epoch": 6.280271230222796,
      "grad_norm": 0.7796829342842102,
      "learning_rate": 0.00022358582049348128,
      "loss": 3.2875,
      "step": 58350
    },
    {
      "epoch": 6.285652782262405,
      "grad_norm": 0.6904585957527161,
      "learning_rate": 0.00022326257946341988,
      "loss": 3.2837,
      "step": 58400
    },
    {
      "epoch": 6.2910343343020125,
      "grad_norm": 0.8047196269035339,
      "learning_rate": 0.00022293933843335845,
      "loss": 3.2669,
      "step": 58450
    },
    {
      "epoch": 6.296415886341621,
      "grad_norm": 0.7967615723609924,
      "learning_rate": 0.00022261609740329704,
      "loss": 3.2932,
      "step": 58500
    },
    {
      "epoch": 6.301797438381229,
      "grad_norm": 0.7712148427963257,
      "learning_rate": 0.0002222928563732356,
      "loss": 3.2687,
      "step": 58550
    },
    {
      "epoch": 6.307178990420837,
      "grad_norm": 0.6820370554924011,
      "learning_rate": 0.00022196961534317423,
      "loss": 3.2871,
      "step": 58600
    },
    {
      "epoch": 6.312560542460446,
      "grad_norm": 0.7562189698219299,
      "learning_rate": 0.0002216463743131128,
      "loss": 3.2954,
      "step": 58650
    },
    {
      "epoch": 6.3179420945000535,
      "grad_norm": 0.7117153406143188,
      "learning_rate": 0.00022132313328305137,
      "loss": 3.2862,
      "step": 58700
    },
    {
      "epoch": 6.323323646539662,
      "grad_norm": 0.7891095876693726,
      "learning_rate": 0.00022099989225299,
      "loss": 3.2947,
      "step": 58750
    },
    {
      "epoch": 6.328705198579271,
      "grad_norm": 0.7475429177284241,
      "learning_rate": 0.00022067665122292856,
      "loss": 3.291,
      "step": 58800
    },
    {
      "epoch": 6.334086750618878,
      "grad_norm": 0.7194887399673462,
      "learning_rate": 0.00022035341019286712,
      "loss": 3.2864,
      "step": 58850
    },
    {
      "epoch": 6.339468302658487,
      "grad_norm": 0.7489674687385559,
      "learning_rate": 0.00022003016916280572,
      "loss": 3.2849,
      "step": 58900
    },
    {
      "epoch": 6.344849854698095,
      "grad_norm": 0.8098034262657166,
      "learning_rate": 0.00021970692813274428,
      "loss": 3.2698,
      "step": 58950
    },
    {
      "epoch": 6.350231406737703,
      "grad_norm": 0.7568627595901489,
      "learning_rate": 0.00021938368710268288,
      "loss": 3.2757,
      "step": 59000
    },
    {
      "epoch": 6.350231406737703,
      "eval_accuracy": 0.38441011423655974,
      "eval_loss": 3.381363868713379,
      "eval_runtime": 206.599,
      "eval_samples_per_second": 87.179,
      "eval_steps_per_second": 5.45,
      "step": 59000
    },
    {
      "epoch": 6.355612958777312,
      "grad_norm": 0.7968950271606445,
      "learning_rate": 0.00021906044607262147,
      "loss": 3.2854,
      "step": 59050
    },
    {
      "epoch": 6.360994510816919,
      "grad_norm": 0.670035183429718,
      "learning_rate": 0.00021873720504256004,
      "loss": 3.2679,
      "step": 59100
    },
    {
      "epoch": 6.366376062856528,
      "grad_norm": 0.7569753527641296,
      "learning_rate": 0.0002184139640124986,
      "loss": 3.2798,
      "step": 59150
    },
    {
      "epoch": 6.371757614896136,
      "grad_norm": 0.7587354183197021,
      "learning_rate": 0.00021809072298243723,
      "loss": 3.2955,
      "step": 59200
    },
    {
      "epoch": 6.377139166935744,
      "grad_norm": 0.7599956393241882,
      "learning_rate": 0.0002177674819523758,
      "loss": 3.2811,
      "step": 59250
    },
    {
      "epoch": 6.382520718975353,
      "grad_norm": 0.7595645785331726,
      "learning_rate": 0.00021744424092231437,
      "loss": 3.2948,
      "step": 59300
    },
    {
      "epoch": 6.387902271014961,
      "grad_norm": 0.7290259599685669,
      "learning_rate": 0.000217120999892253,
      "loss": 3.2771,
      "step": 59350
    },
    {
      "epoch": 6.393283823054569,
      "grad_norm": 0.7082527279853821,
      "learning_rate": 0.00021679775886219156,
      "loss": 3.2888,
      "step": 59400
    },
    {
      "epoch": 6.398665375094177,
      "grad_norm": 0.7418312430381775,
      "learning_rate": 0.00021647451783213012,
      "loss": 3.2992,
      "step": 59450
    },
    {
      "epoch": 6.404046927133785,
      "grad_norm": 0.7465775012969971,
      "learning_rate": 0.00021615127680206872,
      "loss": 3.2879,
      "step": 59500
    },
    {
      "epoch": 6.4094284791733935,
      "grad_norm": 0.7534443140029907,
      "learning_rate": 0.0002158280357720073,
      "loss": 3.2954,
      "step": 59550
    },
    {
      "epoch": 6.414810031213002,
      "grad_norm": 0.7314438223838806,
      "learning_rate": 0.00021551125956254712,
      "loss": 3.301,
      "step": 59600
    },
    {
      "epoch": 6.42019158325261,
      "grad_norm": 0.7473573088645935,
      "learning_rate": 0.00021518801853248569,
      "loss": 3.2845,
      "step": 59650
    },
    {
      "epoch": 6.425573135292218,
      "grad_norm": 0.7379694581031799,
      "learning_rate": 0.0002148647775024243,
      "loss": 3.2957,
      "step": 59700
    },
    {
      "epoch": 6.430954687331827,
      "grad_norm": 0.7331494092941284,
      "learning_rate": 0.00021454153647236288,
      "loss": 3.2984,
      "step": 59750
    },
    {
      "epoch": 6.4363362393714345,
      "grad_norm": 0.7132901549339294,
      "learning_rate": 0.00021421829544230144,
      "loss": 3.2986,
      "step": 59800
    },
    {
      "epoch": 6.441717791411043,
      "grad_norm": 0.7398157119750977,
      "learning_rate": 0.00021389505441224007,
      "loss": 3.3109,
      "step": 59850
    },
    {
      "epoch": 6.447099343450651,
      "grad_norm": 0.7737865447998047,
      "learning_rate": 0.00021357181338217863,
      "loss": 3.2962,
      "step": 59900
    },
    {
      "epoch": 6.452480895490259,
      "grad_norm": 0.7534198760986328,
      "learning_rate": 0.0002132485723521172,
      "loss": 3.2808,
      "step": 59950
    },
    {
      "epoch": 6.457862447529868,
      "grad_norm": 0.748082160949707,
      "learning_rate": 0.0002129253313220558,
      "loss": 3.2957,
      "step": 60000
    },
    {
      "epoch": 6.457862447529868,
      "eval_accuracy": 0.38493925380021626,
      "eval_loss": 3.3789491653442383,
      "eval_runtime": 194.3757,
      "eval_samples_per_second": 92.661,
      "eval_steps_per_second": 5.793,
      "step": 60000
    },
    {
      "epoch": 6.4632439995694755,
      "grad_norm": 0.8238241076469421,
      "learning_rate": 0.00021260209029199436,
      "loss": 3.2791,
      "step": 60050
    },
    {
      "epoch": 6.468625551609084,
      "grad_norm": 0.7775644063949585,
      "learning_rate": 0.00021227884926193296,
      "loss": 3.2915,
      "step": 60100
    },
    {
      "epoch": 6.474007103648693,
      "grad_norm": 0.7641563415527344,
      "learning_rate": 0.00021195560823187155,
      "loss": 3.2991,
      "step": 60150
    },
    {
      "epoch": 6.4793886556883,
      "grad_norm": 0.7693634629249573,
      "learning_rate": 0.00021163236720181012,
      "loss": 3.2802,
      "step": 60200
    },
    {
      "epoch": 6.484770207727909,
      "grad_norm": 0.7530620098114014,
      "learning_rate": 0.0002113091261717487,
      "loss": 3.2906,
      "step": 60250
    },
    {
      "epoch": 6.490151759767517,
      "grad_norm": 0.7825722098350525,
      "learning_rate": 0.0002109858851416873,
      "loss": 3.3047,
      "step": 60300
    },
    {
      "epoch": 6.495533311807125,
      "grad_norm": 0.7141666412353516,
      "learning_rate": 0.00021066264411162588,
      "loss": 3.2873,
      "step": 60350
    },
    {
      "epoch": 6.500914863846734,
      "grad_norm": 0.7286450266838074,
      "learning_rate": 0.00021033940308156447,
      "loss": 3.2928,
      "step": 60400
    },
    {
      "epoch": 6.506296415886341,
      "grad_norm": 0.7147504687309265,
      "learning_rate": 0.00021001616205150307,
      "loss": 3.2976,
      "step": 60450
    },
    {
      "epoch": 6.51167796792595,
      "grad_norm": 0.7341115474700928,
      "learning_rate": 0.00020969292102144163,
      "loss": 3.2853,
      "step": 60500
    },
    {
      "epoch": 6.517059519965558,
      "grad_norm": 0.7411824464797974,
      "learning_rate": 0.00020936967999138023,
      "loss": 3.2844,
      "step": 60550
    },
    {
      "epoch": 6.522441072005166,
      "grad_norm": 0.7643809914588928,
      "learning_rate": 0.0002090464389613188,
      "loss": 3.2994,
      "step": 60600
    },
    {
      "epoch": 6.5278226240447745,
      "grad_norm": 0.7889879941940308,
      "learning_rate": 0.0002087231979312574,
      "loss": 3.3019,
      "step": 60650
    },
    {
      "epoch": 6.533204176084383,
      "grad_norm": 0.716715395450592,
      "learning_rate": 0.00020839995690119598,
      "loss": 3.2929,
      "step": 60700
    },
    {
      "epoch": 6.538585728123991,
      "grad_norm": 0.7614070177078247,
      "learning_rate": 0.00020807671587113455,
      "loss": 3.2964,
      "step": 60750
    },
    {
      "epoch": 6.543967280163599,
      "grad_norm": 0.7336043119430542,
      "learning_rate": 0.00020775347484107312,
      "loss": 3.2943,
      "step": 60800
    },
    {
      "epoch": 6.549348832203208,
      "grad_norm": 0.7477988004684448,
      "learning_rate": 0.00020743023381101174,
      "loss": 3.2923,
      "step": 60850
    },
    {
      "epoch": 6.5547303842428155,
      "grad_norm": 0.720605731010437,
      "learning_rate": 0.0002071069927809503,
      "loss": 3.2836,
      "step": 60900
    },
    {
      "epoch": 6.560111936282424,
      "grad_norm": 0.7655743956565857,
      "learning_rate": 0.00020678375175088888,
      "loss": 3.2932,
      "step": 60950
    },
    {
      "epoch": 6.565493488322032,
      "grad_norm": 0.711134672164917,
      "learning_rate": 0.0002064605107208275,
      "loss": 3.2904,
      "step": 61000
    },
    {
      "epoch": 6.565493488322032,
      "eval_accuracy": 0.38553206395613204,
      "eval_loss": 3.3737478256225586,
      "eval_runtime": 204.527,
      "eval_samples_per_second": 88.062,
      "eval_steps_per_second": 5.505,
      "step": 61000
    },
    {
      "epoch": 6.57087504036164,
      "grad_norm": 0.7845826745033264,
      "learning_rate": 0.00020613726969076607,
      "loss": 3.2885,
      "step": 61050
    },
    {
      "epoch": 6.576256592401249,
      "grad_norm": 0.7667222023010254,
      "learning_rate": 0.00020581402866070463,
      "loss": 3.2775,
      "step": 61100
    },
    {
      "epoch": 6.5816381444408565,
      "grad_norm": 0.7512591481208801,
      "learning_rate": 0.00020549078763064323,
      "loss": 3.2908,
      "step": 61150
    },
    {
      "epoch": 6.587019696480465,
      "grad_norm": 0.822928249835968,
      "learning_rate": 0.00020516754660058182,
      "loss": 3.2994,
      "step": 61200
    },
    {
      "epoch": 6.592401248520073,
      "grad_norm": 0.7494385242462158,
      "learning_rate": 0.0002048443055705204,
      "loss": 3.2906,
      "step": 61250
    },
    {
      "epoch": 6.597782800559681,
      "grad_norm": 0.7336588501930237,
      "learning_rate": 0.00020452106454045898,
      "loss": 3.2906,
      "step": 61300
    },
    {
      "epoch": 6.60316435259929,
      "grad_norm": 0.7292616963386536,
      "learning_rate": 0.00020419782351039755,
      "loss": 3.2969,
      "step": 61350
    },
    {
      "epoch": 6.608545904638898,
      "grad_norm": 0.799895703792572,
      "learning_rate": 0.00020387458248033617,
      "loss": 3.2937,
      "step": 61400
    },
    {
      "epoch": 6.613927456678506,
      "grad_norm": 0.7496054172515869,
      "learning_rate": 0.00020355134145027474,
      "loss": 3.2896,
      "step": 61450
    },
    {
      "epoch": 6.619309008718115,
      "grad_norm": 0.7322660684585571,
      "learning_rate": 0.0002032281004202133,
      "loss": 3.286,
      "step": 61500
    },
    {
      "epoch": 6.624690560757722,
      "grad_norm": 0.7610208988189697,
      "learning_rate": 0.00020290485939015193,
      "loss": 3.2949,
      "step": 61550
    },
    {
      "epoch": 6.630072112797331,
      "grad_norm": 0.816765546798706,
      "learning_rate": 0.0002025816183600905,
      "loss": 3.2693,
      "step": 61600
    },
    {
      "epoch": 6.635453664836939,
      "grad_norm": 0.7446416020393372,
      "learning_rate": 0.00020225837733002907,
      "loss": 3.2851,
      "step": 61650
    },
    {
      "epoch": 6.640835216876547,
      "grad_norm": 0.7142714262008667,
      "learning_rate": 0.00020193513629996766,
      "loss": 3.2994,
      "step": 61700
    },
    {
      "epoch": 6.6462167689161555,
      "grad_norm": 0.7345561981201172,
      "learning_rate": 0.00020161189526990626,
      "loss": 3.2922,
      "step": 61750
    },
    {
      "epoch": 6.651598320955763,
      "grad_norm": 0.7665655016899109,
      "learning_rate": 0.00020128865423984482,
      "loss": 3.2936,
      "step": 61800
    },
    {
      "epoch": 6.656979872995372,
      "grad_norm": 0.7898482084274292,
      "learning_rate": 0.00020096541320978342,
      "loss": 3.2682,
      "step": 61850
    },
    {
      "epoch": 6.66236142503498,
      "grad_norm": 0.7559719085693359,
      "learning_rate": 0.00020064217217972199,
      "loss": 3.288,
      "step": 61900
    },
    {
      "epoch": 6.667742977074588,
      "grad_norm": 0.7487570643424988,
      "learning_rate": 0.00020031893114966058,
      "loss": 3.288,
      "step": 61950
    },
    {
      "epoch": 6.6731245291141965,
      "grad_norm": 0.8141021132469177,
      "learning_rate": 0.00019999569011959917,
      "loss": 3.2969,
      "step": 62000
    },
    {
      "epoch": 6.6731245291141965,
      "eval_accuracy": 0.3857010191966836,
      "eval_loss": 3.369795560836792,
      "eval_runtime": 202.5752,
      "eval_samples_per_second": 88.91,
      "eval_steps_per_second": 5.558,
      "step": 62000
    },
    {
      "epoch": 6.678506081153805,
      "grad_norm": 0.7196623086929321,
      "learning_rate": 0.00019967244908953774,
      "loss": 3.2763,
      "step": 62050
    },
    {
      "epoch": 6.683887633193413,
      "grad_norm": 0.726294994354248,
      "learning_rate": 0.0001993492080594763,
      "loss": 3.2938,
      "step": 62100
    },
    {
      "epoch": 6.689269185233021,
      "grad_norm": 0.7594084739685059,
      "learning_rate": 0.00019902596702941493,
      "loss": 3.2751,
      "step": 62150
    },
    {
      "epoch": 6.69465073727263,
      "grad_norm": 0.7860019207000732,
      "learning_rate": 0.0001987027259993535,
      "loss": 3.2702,
      "step": 62200
    },
    {
      "epoch": 6.7000322893122375,
      "grad_norm": 0.745943546295166,
      "learning_rate": 0.00019837948496929207,
      "loss": 3.2905,
      "step": 62250
    },
    {
      "epoch": 6.705413841351846,
      "grad_norm": 0.7444291710853577,
      "learning_rate": 0.00019805624393923066,
      "loss": 3.3002,
      "step": 62300
    },
    {
      "epoch": 6.710795393391454,
      "grad_norm": 0.7500001192092896,
      "learning_rate": 0.00019773300290916926,
      "loss": 3.2895,
      "step": 62350
    },
    {
      "epoch": 6.716176945431062,
      "grad_norm": 0.7602120637893677,
      "learning_rate": 0.00019740976187910785,
      "loss": 3.2994,
      "step": 62400
    },
    {
      "epoch": 6.721558497470671,
      "grad_norm": 0.7956641912460327,
      "learning_rate": 0.00019708652084904642,
      "loss": 3.3066,
      "step": 62450
    },
    {
      "epoch": 6.7269400495102785,
      "grad_norm": 0.751025915145874,
      "learning_rate": 0.00019676327981898499,
      "loss": 3.2885,
      "step": 62500
    },
    {
      "epoch": 6.732321601549887,
      "grad_norm": 0.8175959587097168,
      "learning_rate": 0.0001964400387889236,
      "loss": 3.2902,
      "step": 62550
    },
    {
      "epoch": 6.737703153589496,
      "grad_norm": 0.7570839524269104,
      "learning_rate": 0.00019611679775886217,
      "loss": 3.3131,
      "step": 62600
    },
    {
      "epoch": 6.743084705629103,
      "grad_norm": 0.7345150113105774,
      "learning_rate": 0.00019579355672880074,
      "loss": 3.2913,
      "step": 62650
    },
    {
      "epoch": 6.748466257668712,
      "grad_norm": 0.7458459734916687,
      "learning_rate": 0.00019547031569873936,
      "loss": 3.2859,
      "step": 62700
    },
    {
      "epoch": 6.75384780970832,
      "grad_norm": 0.713222324848175,
      "learning_rate": 0.00019514707466867793,
      "loss": 3.2901,
      "step": 62750
    },
    {
      "epoch": 6.759229361747928,
      "grad_norm": 0.7876093983650208,
      "learning_rate": 0.0001948238336386165,
      "loss": 3.2818,
      "step": 62800
    },
    {
      "epoch": 6.7646109137875365,
      "grad_norm": 0.73554527759552,
      "learning_rate": 0.0001945005926085551,
      "loss": 3.3021,
      "step": 62850
    },
    {
      "epoch": 6.769992465827144,
      "grad_norm": 0.7704194784164429,
      "learning_rate": 0.0001941773515784937,
      "loss": 3.2882,
      "step": 62900
    },
    {
      "epoch": 6.775374017866753,
      "grad_norm": 0.8291070461273193,
      "learning_rate": 0.00019385411054843226,
      "loss": 3.2966,
      "step": 62950
    },
    {
      "epoch": 6.780755569906361,
      "grad_norm": 0.794005811214447,
      "learning_rate": 0.00019353086951837085,
      "loss": 3.3017,
      "step": 63000
    },
    {
      "epoch": 6.780755569906361,
      "eval_accuracy": 0.3859023529978038,
      "eval_loss": 3.367936849594116,
      "eval_runtime": 192.8573,
      "eval_samples_per_second": 93.39,
      "eval_steps_per_second": 5.839,
      "step": 63000
    },
    {
      "epoch": 6.786137121945969,
      "grad_norm": 0.7725952863693237,
      "learning_rate": 0.00019320762848830942,
      "loss": 3.2969,
      "step": 63050
    },
    {
      "epoch": 6.7915186739855775,
      "grad_norm": 0.7649474143981934,
      "learning_rate": 0.000192884387458248,
      "loss": 3.2864,
      "step": 63100
    },
    {
      "epoch": 6.796900226025185,
      "grad_norm": 0.7265825867652893,
      "learning_rate": 0.0001925611464281866,
      "loss": 3.2795,
      "step": 63150
    },
    {
      "epoch": 6.802281778064794,
      "grad_norm": 0.7690824270248413,
      "learning_rate": 0.00019223790539812518,
      "loss": 3.279,
      "step": 63200
    },
    {
      "epoch": 6.807663330104402,
      "grad_norm": 0.7158901691436768,
      "learning_rate": 0.00019191466436806374,
      "loss": 3.2853,
      "step": 63250
    },
    {
      "epoch": 6.813044882144011,
      "grad_norm": 0.7686758041381836,
      "learning_rate": 0.00019159142333800236,
      "loss": 3.284,
      "step": 63300
    },
    {
      "epoch": 6.8184264341836185,
      "grad_norm": 0.7819889187812805,
      "learning_rate": 0.00019126818230794093,
      "loss": 3.3023,
      "step": 63350
    },
    {
      "epoch": 6.823807986223227,
      "grad_norm": 0.7525413036346436,
      "learning_rate": 0.00019094494127787953,
      "loss": 3.2892,
      "step": 63400
    },
    {
      "epoch": 6.829189538262835,
      "grad_norm": 0.7915401458740234,
      "learning_rate": 0.00019062170024781812,
      "loss": 3.2932,
      "step": 63450
    },
    {
      "epoch": 6.834571090302443,
      "grad_norm": 0.760724663734436,
      "learning_rate": 0.0001902984592177567,
      "loss": 3.2899,
      "step": 63500
    },
    {
      "epoch": 6.839952642342052,
      "grad_norm": 0.7784046530723572,
      "learning_rate": 0.00018997521818769528,
      "loss": 3.3147,
      "step": 63550
    },
    {
      "epoch": 6.8453341943816595,
      "grad_norm": 0.7777194380760193,
      "learning_rate": 0.00018965197715763385,
      "loss": 3.2875,
      "step": 63600
    },
    {
      "epoch": 6.850715746421268,
      "grad_norm": 0.8105765581130981,
      "learning_rate": 0.00018933520094817368,
      "loss": 3.2991,
      "step": 63650
    },
    {
      "epoch": 6.856097298460876,
      "grad_norm": 0.7386813759803772,
      "learning_rate": 0.00018901195991811225,
      "loss": 3.3049,
      "step": 63700
    },
    {
      "epoch": 6.861478850500484,
      "grad_norm": 0.7848280668258667,
      "learning_rate": 0.00018868871888805082,
      "loss": 3.3024,
      "step": 63750
    },
    {
      "epoch": 6.866860402540093,
      "grad_norm": 0.7499856948852539,
      "learning_rate": 0.00018836547785798944,
      "loss": 3.2906,
      "step": 63800
    },
    {
      "epoch": 6.8722419545797,
      "grad_norm": 0.8316891193389893,
      "learning_rate": 0.000188042236827928,
      "loss": 3.3057,
      "step": 63850
    },
    {
      "epoch": 6.877623506619309,
      "grad_norm": 0.7566866278648376,
      "learning_rate": 0.00018771899579786658,
      "loss": 3.2958,
      "step": 63900
    },
    {
      "epoch": 6.8830050586589175,
      "grad_norm": 0.7290393114089966,
      "learning_rate": 0.00018739575476780517,
      "loss": 3.3017,
      "step": 63950
    },
    {
      "epoch": 6.888386610698525,
      "grad_norm": 0.753031313419342,
      "learning_rate": 0.00018707251373774377,
      "loss": 3.2973,
      "step": 64000
    },
    {
      "epoch": 6.888386610698525,
      "eval_accuracy": 0.3864932074017389,
      "eval_loss": 3.3619954586029053,
      "eval_runtime": 184.6401,
      "eval_samples_per_second": 97.547,
      "eval_steps_per_second": 6.098,
      "step": 64000
    },
    {
      "epoch": 6.893768162738134,
      "grad_norm": 0.7685363292694092,
      "learning_rate": 0.00018674927270768233,
      "loss": 3.2798,
      "step": 64050
    },
    {
      "epoch": 6.899149714777742,
      "grad_norm": 0.7665897011756897,
      "learning_rate": 0.00018642603167762093,
      "loss": 3.2936,
      "step": 64100
    },
    {
      "epoch": 6.90453126681735,
      "grad_norm": 0.7315976023674011,
      "learning_rate": 0.0001861027906475595,
      "loss": 3.3051,
      "step": 64150
    },
    {
      "epoch": 6.9099128188569585,
      "grad_norm": 0.8117026686668396,
      "learning_rate": 0.00018577954961749812,
      "loss": 3.2972,
      "step": 64200
    },
    {
      "epoch": 6.915294370896566,
      "grad_norm": 0.7579342126846313,
      "learning_rate": 0.00018545630858743669,
      "loss": 3.2764,
      "step": 64250
    },
    {
      "epoch": 6.920675922936175,
      "grad_norm": 0.8170520067214966,
      "learning_rate": 0.00018513306755737525,
      "loss": 3.2872,
      "step": 64300
    },
    {
      "epoch": 6.926057474975783,
      "grad_norm": 0.7994571924209595,
      "learning_rate": 0.00018480982652731387,
      "loss": 3.298,
      "step": 64350
    },
    {
      "epoch": 6.931439027015391,
      "grad_norm": 0.7614535689353943,
      "learning_rate": 0.00018448658549725244,
      "loss": 3.2851,
      "step": 64400
    },
    {
      "epoch": 6.9368205790549995,
      "grad_norm": 0.7701461315155029,
      "learning_rate": 0.000184163344467191,
      "loss": 3.2827,
      "step": 64450
    },
    {
      "epoch": 6.942202131094608,
      "grad_norm": 0.8451718091964722,
      "learning_rate": 0.0001838401034371296,
      "loss": 3.3129,
      "step": 64500
    },
    {
      "epoch": 6.947583683134216,
      "grad_norm": 0.8217495679855347,
      "learning_rate": 0.0001835168624070682,
      "loss": 3.2794,
      "step": 64550
    },
    {
      "epoch": 6.952965235173824,
      "grad_norm": 0.7501987218856812,
      "learning_rate": 0.00018319362137700677,
      "loss": 3.2931,
      "step": 64600
    },
    {
      "epoch": 6.958346787213433,
      "grad_norm": 0.7633636593818665,
      "learning_rate": 0.00018287038034694536,
      "loss": 3.2956,
      "step": 64650
    },
    {
      "epoch": 6.9637283392530405,
      "grad_norm": 0.7722277045249939,
      "learning_rate": 0.00018254713931688393,
      "loss": 3.2941,
      "step": 64700
    },
    {
      "epoch": 6.969109891292649,
      "grad_norm": 0.7616075277328491,
      "learning_rate": 0.00018222389828682252,
      "loss": 3.309,
      "step": 64750
    },
    {
      "epoch": 6.974491443332257,
      "grad_norm": 0.8005858063697815,
      "learning_rate": 0.00018190065725676112,
      "loss": 3.2864,
      "step": 64800
    },
    {
      "epoch": 6.979872995371865,
      "grad_norm": 0.7860175371170044,
      "learning_rate": 0.00018157741622669969,
      "loss": 3.2855,
      "step": 64850
    },
    {
      "epoch": 6.985254547411474,
      "grad_norm": 0.7549825310707092,
      "learning_rate": 0.00018125417519663825,
      "loss": 3.3049,
      "step": 64900
    },
    {
      "epoch": 6.990636099451081,
      "grad_norm": 0.7355542778968811,
      "learning_rate": 0.00018093093416657687,
      "loss": 3.2792,
      "step": 64950
    },
    {
      "epoch": 6.99601765149069,
      "grad_norm": 0.7631430625915527,
      "learning_rate": 0.00018060769313651544,
      "loss": 3.3034,
      "step": 65000
    },
    {
      "epoch": 6.99601765149069,
      "eval_accuracy": 0.38705222650954446,
      "eval_loss": 3.357828140258789,
      "eval_runtime": 184.5064,
      "eval_samples_per_second": 97.617,
      "eval_steps_per_second": 6.103,
      "step": 65000
    },
    {
      "epoch": 7.0013992035302985,
      "grad_norm": 0.7348278760910034,
      "learning_rate": 0.000180284452106454,
      "loss": 3.2708,
      "step": 65050
    },
    {
      "epoch": 7.006780755569906,
      "grad_norm": 0.789198637008667,
      "learning_rate": 0.00017996121107639263,
      "loss": 3.2092,
      "step": 65100
    },
    {
      "epoch": 7.012162307609515,
      "grad_norm": 0.823444128036499,
      "learning_rate": 0.0001796379700463312,
      "loss": 3.2018,
      "step": 65150
    },
    {
      "epoch": 7.017543859649122,
      "grad_norm": 0.7726554870605469,
      "learning_rate": 0.0001793147290162698,
      "loss": 3.2273,
      "step": 65200
    },
    {
      "epoch": 7.022925411688731,
      "grad_norm": 0.7580903172492981,
      "learning_rate": 0.00017899148798620836,
      "loss": 3.2011,
      "step": 65250
    },
    {
      "epoch": 7.0283069637283395,
      "grad_norm": 0.7943834662437439,
      "learning_rate": 0.00017866824695614696,
      "loss": 3.2019,
      "step": 65300
    },
    {
      "epoch": 7.033688515767947,
      "grad_norm": 0.7941718697547913,
      "learning_rate": 0.00017834500592608555,
      "loss": 3.2026,
      "step": 65350
    },
    {
      "epoch": 7.039070067807556,
      "grad_norm": 0.7380354404449463,
      "learning_rate": 0.00017802176489602412,
      "loss": 3.207,
      "step": 65400
    },
    {
      "epoch": 7.044451619847164,
      "grad_norm": 0.8549360036849976,
      "learning_rate": 0.00017769852386596269,
      "loss": 3.2173,
      "step": 65450
    },
    {
      "epoch": 7.049833171886772,
      "grad_norm": 0.772775411605835,
      "learning_rate": 0.0001773752828359013,
      "loss": 3.2096,
      "step": 65500
    },
    {
      "epoch": 7.0552147239263805,
      "grad_norm": 0.782207727432251,
      "learning_rate": 0.00017705204180583988,
      "loss": 3.2139,
      "step": 65550
    },
    {
      "epoch": 7.060596275965988,
      "grad_norm": 0.8233615159988403,
      "learning_rate": 0.00017672880077577844,
      "loss": 3.2345,
      "step": 65600
    },
    {
      "epoch": 7.065977828005597,
      "grad_norm": 0.7934371829032898,
      "learning_rate": 0.00017641202456631828,
      "loss": 3.2212,
      "step": 65650
    },
    {
      "epoch": 7.071359380045205,
      "grad_norm": 0.7492845058441162,
      "learning_rate": 0.00017608878353625684,
      "loss": 3.2217,
      "step": 65700
    },
    {
      "epoch": 7.076740932084813,
      "grad_norm": 0.723509669303894,
      "learning_rate": 0.00017576554250619544,
      "loss": 3.2049,
      "step": 65750
    },
    {
      "epoch": 7.0821224841244215,
      "grad_norm": 0.769831657409668,
      "learning_rate": 0.000175442301476134,
      "loss": 3.2285,
      "step": 65800
    },
    {
      "epoch": 7.08750403616403,
      "grad_norm": 0.7707816362380981,
      "learning_rate": 0.0001751190604460726,
      "loss": 3.2182,
      "step": 65850
    },
    {
      "epoch": 7.092885588203638,
      "grad_norm": 0.8252584934234619,
      "learning_rate": 0.0001747958194160112,
      "loss": 3.2275,
      "step": 65900
    },
    {
      "epoch": 7.098267140243246,
      "grad_norm": 0.8173145055770874,
      "learning_rate": 0.00017447257838594976,
      "loss": 3.2254,
      "step": 65950
    },
    {
      "epoch": 7.103648692282855,
      "grad_norm": 0.7946733236312866,
      "learning_rate": 0.00017414933735588838,
      "loss": 3.2471,
      "step": 66000
    },
    {
      "epoch": 7.103648692282855,
      "eval_accuracy": 0.3867682078607974,
      "eval_loss": 3.3641140460968018,
      "eval_runtime": 184.4963,
      "eval_samples_per_second": 97.623,
      "eval_steps_per_second": 6.103,
      "step": 66000
    },
    {
      "epoch": 7.109030244322462,
      "grad_norm": 0.760890781879425,
      "learning_rate": 0.00017382609632582695,
      "loss": 3.2201,
      "step": 66050
    },
    {
      "epoch": 7.114411796362071,
      "grad_norm": 0.8106924891471863,
      "learning_rate": 0.00017350285529576552,
      "loss": 3.2331,
      "step": 66100
    },
    {
      "epoch": 7.119793348401679,
      "grad_norm": 0.7920600771903992,
      "learning_rate": 0.00017317961426570411,
      "loss": 3.2145,
      "step": 66150
    },
    {
      "epoch": 7.125174900441287,
      "grad_norm": 0.7734333872795105,
      "learning_rate": 0.0001728563732356427,
      "loss": 3.2091,
      "step": 66200
    },
    {
      "epoch": 7.130556452480896,
      "grad_norm": 0.844710111618042,
      "learning_rate": 0.00017253313220558128,
      "loss": 3.2344,
      "step": 66250
    },
    {
      "epoch": 7.135938004520503,
      "grad_norm": 0.8509883284568787,
      "learning_rate": 0.00017220989117551987,
      "loss": 3.2165,
      "step": 66300
    },
    {
      "epoch": 7.141319556560112,
      "grad_norm": 0.7604499459266663,
      "learning_rate": 0.00017188665014545844,
      "loss": 3.2143,
      "step": 66350
    },
    {
      "epoch": 7.1467011085997205,
      "grad_norm": 0.7459873557090759,
      "learning_rate": 0.00017156340911539703,
      "loss": 3.227,
      "step": 66400
    },
    {
      "epoch": 7.152082660639328,
      "grad_norm": 0.7575955390930176,
      "learning_rate": 0.00017124016808533563,
      "loss": 3.2454,
      "step": 66450
    },
    {
      "epoch": 7.157464212678937,
      "grad_norm": 0.8272690773010254,
      "learning_rate": 0.0001709169270552742,
      "loss": 3.2342,
      "step": 66500
    },
    {
      "epoch": 7.162845764718545,
      "grad_norm": 0.7852942943572998,
      "learning_rate": 0.00017059368602521276,
      "loss": 3.2215,
      "step": 66550
    },
    {
      "epoch": 7.168227316758153,
      "grad_norm": 0.7997828125953674,
      "learning_rate": 0.00017027044499515139,
      "loss": 3.2295,
      "step": 66600
    },
    {
      "epoch": 7.1736088687977615,
      "grad_norm": 0.7966010570526123,
      "learning_rate": 0.00016994720396508995,
      "loss": 3.2398,
      "step": 66650
    },
    {
      "epoch": 7.178990420837369,
      "grad_norm": 0.7404817938804626,
      "learning_rate": 0.00016962396293502852,
      "loss": 3.2189,
      "step": 66700
    },
    {
      "epoch": 7.184371972876978,
      "grad_norm": 0.7725746035575867,
      "learning_rate": 0.00016930072190496714,
      "loss": 3.2365,
      "step": 66750
    },
    {
      "epoch": 7.189753524916586,
      "grad_norm": 0.7487515807151794,
      "learning_rate": 0.0001689774808749057,
      "loss": 3.2284,
      "step": 66800
    },
    {
      "epoch": 7.195135076956194,
      "grad_norm": 0.7499515414237976,
      "learning_rate": 0.00016866070466544552,
      "loss": 3.2288,
      "step": 66850
    },
    {
      "epoch": 7.2005166289958025,
      "grad_norm": 0.7855218648910522,
      "learning_rate": 0.00016833746363538408,
      "loss": 3.2048,
      "step": 66900
    },
    {
      "epoch": 7.205898181035411,
      "grad_norm": 0.7659774422645569,
      "learning_rate": 0.0001680142226053227,
      "loss": 3.2279,
      "step": 66950
    },
    {
      "epoch": 7.211279733075019,
      "grad_norm": 0.7899845242500305,
      "learning_rate": 0.00016769098157526127,
      "loss": 3.2302,
      "step": 67000
    },
    {
      "epoch": 7.211279733075019,
      "eval_accuracy": 0.3871081827467689,
      "eval_loss": 3.362567901611328,
      "eval_runtime": 184.7128,
      "eval_samples_per_second": 97.508,
      "eval_steps_per_second": 6.096,
      "step": 67000
    },
    {
      "epoch": 7.216661285114627,
      "grad_norm": 0.8087143898010254,
      "learning_rate": 0.00016736774054519984,
      "loss": 3.2136,
      "step": 67050
    },
    {
      "epoch": 7.222042837154235,
      "grad_norm": 0.7657251358032227,
      "learning_rate": 0.00016704449951513846,
      "loss": 3.2277,
      "step": 67100
    },
    {
      "epoch": 7.2274243891938434,
      "grad_norm": 0.7933040857315063,
      "learning_rate": 0.00016672125848507703,
      "loss": 3.239,
      "step": 67150
    },
    {
      "epoch": 7.232805941233452,
      "grad_norm": 0.771693229675293,
      "learning_rate": 0.0001663980174550156,
      "loss": 3.2236,
      "step": 67200
    },
    {
      "epoch": 7.23818749327306,
      "grad_norm": 0.7936459183692932,
      "learning_rate": 0.0001660747764249542,
      "loss": 3.2432,
      "step": 67250
    },
    {
      "epoch": 7.243569045312668,
      "grad_norm": 0.8471320867538452,
      "learning_rate": 0.0001657515353948928,
      "loss": 3.2486,
      "step": 67300
    },
    {
      "epoch": 7.248950597352277,
      "grad_norm": 0.8385496735572815,
      "learning_rate": 0.00016542829436483135,
      "loss": 3.2108,
      "step": 67350
    },
    {
      "epoch": 7.254332149391884,
      "grad_norm": 0.7460642457008362,
      "learning_rate": 0.00016510505333476995,
      "loss": 3.2311,
      "step": 67400
    },
    {
      "epoch": 7.259713701431493,
      "grad_norm": 0.7928326725959778,
      "learning_rate": 0.00016478181230470852,
      "loss": 3.2501,
      "step": 67450
    },
    {
      "epoch": 7.265095253471101,
      "grad_norm": 0.7807007431983948,
      "learning_rate": 0.0001644585712746471,
      "loss": 3.225,
      "step": 67500
    },
    {
      "epoch": 7.270476805510709,
      "grad_norm": 0.7632108330726624,
      "learning_rate": 0.0001641353302445857,
      "loss": 3.2298,
      "step": 67550
    },
    {
      "epoch": 7.275858357550318,
      "grad_norm": 0.7463772892951965,
      "learning_rate": 0.00016381208921452427,
      "loss": 3.2322,
      "step": 67600
    },
    {
      "epoch": 7.281239909589925,
      "grad_norm": 0.7410020232200623,
      "learning_rate": 0.00016348884818446284,
      "loss": 3.2272,
      "step": 67650
    },
    {
      "epoch": 7.286621461629534,
      "grad_norm": 0.8022161722183228,
      "learning_rate": 0.00016316560715440146,
      "loss": 3.24,
      "step": 67700
    },
    {
      "epoch": 7.2920030136691425,
      "grad_norm": 0.848987340927124,
      "learning_rate": 0.00016284236612434003,
      "loss": 3.2284,
      "step": 67750
    },
    {
      "epoch": 7.29738456570875,
      "grad_norm": 0.7826640605926514,
      "learning_rate": 0.00016251912509427863,
      "loss": 3.2286,
      "step": 67800
    },
    {
      "epoch": 7.302766117748359,
      "grad_norm": 0.780402660369873,
      "learning_rate": 0.00016219588406421722,
      "loss": 3.2427,
      "step": 67850
    },
    {
      "epoch": 7.308147669787967,
      "grad_norm": 0.7835871577262878,
      "learning_rate": 0.0001618726430341558,
      "loss": 3.2101,
      "step": 67900
    },
    {
      "epoch": 7.313529221827575,
      "grad_norm": 0.805471658706665,
      "learning_rate": 0.00016154940200409438,
      "loss": 3.232,
      "step": 67950
    },
    {
      "epoch": 7.3189107738671835,
      "grad_norm": 0.8375717401504517,
      "learning_rate": 0.00016122616097403295,
      "loss": 3.2508,
      "step": 68000
    },
    {
      "epoch": 7.3189107738671835,
      "eval_accuracy": 0.3872799629624036,
      "eval_loss": 3.358309030532837,
      "eval_runtime": 184.5402,
      "eval_samples_per_second": 97.599,
      "eval_steps_per_second": 6.102,
      "step": 68000
    },
    {
      "epoch": 7.324292325906791,
      "grad_norm": 0.8032613396644592,
      "learning_rate": 0.00016090291994397154,
      "loss": 3.2425,
      "step": 68050
    },
    {
      "epoch": 7.3296738779464,
      "grad_norm": 0.7996606230735779,
      "learning_rate": 0.00016057967891391014,
      "loss": 3.2371,
      "step": 68100
    },
    {
      "epoch": 7.335055429986008,
      "grad_norm": 0.7681496143341064,
      "learning_rate": 0.0001602564378838487,
      "loss": 3.2337,
      "step": 68150
    },
    {
      "epoch": 7.340436982025616,
      "grad_norm": 0.7919473648071289,
      "learning_rate": 0.00015993319685378727,
      "loss": 3.223,
      "step": 68200
    },
    {
      "epoch": 7.3458185340652244,
      "grad_norm": 0.8312409520149231,
      "learning_rate": 0.0001596099558237259,
      "loss": 3.2302,
      "step": 68250
    },
    {
      "epoch": 7.351200086104833,
      "grad_norm": 0.7892899513244629,
      "learning_rate": 0.00015928671479366446,
      "loss": 3.2249,
      "step": 68300
    },
    {
      "epoch": 7.356581638144441,
      "grad_norm": 0.7596795558929443,
      "learning_rate": 0.00015896347376360303,
      "loss": 3.2266,
      "step": 68350
    },
    {
      "epoch": 7.361963190184049,
      "grad_norm": 0.7913126349449158,
      "learning_rate": 0.00015864023273354165,
      "loss": 3.2346,
      "step": 68400
    },
    {
      "epoch": 7.367344742223658,
      "grad_norm": 0.8068732023239136,
      "learning_rate": 0.00015831699170348022,
      "loss": 3.2505,
      "step": 68450
    },
    {
      "epoch": 7.372726294263265,
      "grad_norm": 0.7833318114280701,
      "learning_rate": 0.0001579937506734188,
      "loss": 3.2298,
      "step": 68500
    },
    {
      "epoch": 7.378107846302874,
      "grad_norm": 0.7812859416007996,
      "learning_rate": 0.00015767050964335738,
      "loss": 3.2359,
      "step": 68550
    },
    {
      "epoch": 7.383489398342482,
      "grad_norm": 0.8268133401870728,
      "learning_rate": 0.00015734726861329595,
      "loss": 3.2345,
      "step": 68600
    },
    {
      "epoch": 7.38887095038209,
      "grad_norm": 0.8205984234809875,
      "learning_rate": 0.00015702402758323454,
      "loss": 3.2384,
      "step": 68650
    },
    {
      "epoch": 7.394252502421699,
      "grad_norm": 0.8627930283546448,
      "learning_rate": 0.00015670078655317314,
      "loss": 3.2454,
      "step": 68700
    },
    {
      "epoch": 7.399634054461306,
      "grad_norm": 0.816469669342041,
      "learning_rate": 0.0001563775455231117,
      "loss": 3.2425,
      "step": 68750
    },
    {
      "epoch": 7.405015606500915,
      "grad_norm": 0.8038166761398315,
      "learning_rate": 0.00015605430449305033,
      "loss": 3.2319,
      "step": 68800
    },
    {
      "epoch": 7.4103971585405235,
      "grad_norm": 0.8335810899734497,
      "learning_rate": 0.0001557310634629889,
      "loss": 3.2355,
      "step": 68850
    },
    {
      "epoch": 7.415778710580131,
      "grad_norm": 0.8542603850364685,
      "learning_rate": 0.00015540782243292746,
      "loss": 3.2325,
      "step": 68900
    },
    {
      "epoch": 7.42116026261974,
      "grad_norm": 0.7642379999160767,
      "learning_rate": 0.0001550910462234673,
      "loss": 3.2339,
      "step": 68950
    },
    {
      "epoch": 7.426541814659347,
      "grad_norm": 0.8091346025466919,
      "learning_rate": 0.00015476780519340586,
      "loss": 3.2397,
      "step": 69000
    },
    {
      "epoch": 7.426541814659347,
      "eval_accuracy": 0.38774771364444494,
      "eval_loss": 3.353181838989258,
      "eval_runtime": 184.4989,
      "eval_samples_per_second": 97.621,
      "eval_steps_per_second": 6.103,
      "step": 69000
    },
    {
      "epoch": 7.431923366698956,
      "grad_norm": 0.8532742857933044,
      "learning_rate": 0.00015444456416334446,
      "loss": 3.2444,
      "step": 69050
    },
    {
      "epoch": 7.4373049187385645,
      "grad_norm": 0.8678274154663086,
      "learning_rate": 0.00015412132313328303,
      "loss": 3.2448,
      "step": 69100
    },
    {
      "epoch": 7.442686470778172,
      "grad_norm": 0.8867831230163574,
      "learning_rate": 0.00015379808210322162,
      "loss": 3.2337,
      "step": 69150
    },
    {
      "epoch": 7.448068022817781,
      "grad_norm": 0.7986330389976501,
      "learning_rate": 0.00015347484107316022,
      "loss": 3.2685,
      "step": 69200
    },
    {
      "epoch": 7.453449574857389,
      "grad_norm": 0.7819139361381531,
      "learning_rate": 0.00015315160004309878,
      "loss": 3.225,
      "step": 69250
    },
    {
      "epoch": 7.458831126896997,
      "grad_norm": 0.8228380680084229,
      "learning_rate": 0.00015282835901303735,
      "loss": 3.2518,
      "step": 69300
    },
    {
      "epoch": 7.4642126789366054,
      "grad_norm": 0.7952637672424316,
      "learning_rate": 0.00015250511798297597,
      "loss": 3.233,
      "step": 69350
    },
    {
      "epoch": 7.469594230976213,
      "grad_norm": 0.7979111671447754,
      "learning_rate": 0.00015218187695291454,
      "loss": 3.2217,
      "step": 69400
    },
    {
      "epoch": 7.474975783015822,
      "grad_norm": 0.7970550656318665,
      "learning_rate": 0.0001518586359228531,
      "loss": 3.2353,
      "step": 69450
    },
    {
      "epoch": 7.48035733505543,
      "grad_norm": 0.8525084257125854,
      "learning_rate": 0.00015153539489279173,
      "loss": 3.2416,
      "step": 69500
    },
    {
      "epoch": 7.485738887095038,
      "grad_norm": 0.8426193594932556,
      "learning_rate": 0.0001512121538627303,
      "loss": 3.2402,
      "step": 69550
    },
    {
      "epoch": 7.491120439134646,
      "grad_norm": 0.8252077698707581,
      "learning_rate": 0.0001508889128326689,
      "loss": 3.2444,
      "step": 69600
    },
    {
      "epoch": 7.496501991174255,
      "grad_norm": 0.781542956829071,
      "learning_rate": 0.00015056567180260746,
      "loss": 3.2308,
      "step": 69650
    },
    {
      "epoch": 7.501883543213863,
      "grad_norm": 0.7765420079231262,
      "learning_rate": 0.00015024243077254603,
      "loss": 3.2335,
      "step": 69700
    },
    {
      "epoch": 7.507265095253471,
      "grad_norm": 0.8080489039421082,
      "learning_rate": 0.00014991918974248462,
      "loss": 3.2369,
      "step": 69750
    },
    {
      "epoch": 7.51264664729308,
      "grad_norm": 0.7995004057884216,
      "learning_rate": 0.00014959594871242322,
      "loss": 3.2453,
      "step": 69800
    },
    {
      "epoch": 7.518028199332687,
      "grad_norm": 0.7954206466674805,
      "learning_rate": 0.0001492727076823618,
      "loss": 3.2709,
      "step": 69850
    },
    {
      "epoch": 7.523409751372296,
      "grad_norm": 0.7697217464447021,
      "learning_rate": 0.00014894946665230038,
      "loss": 3.24,
      "step": 69900
    },
    {
      "epoch": 7.528791303411904,
      "grad_norm": 0.7994235754013062,
      "learning_rate": 0.00014862622562223897,
      "loss": 3.2267,
      "step": 69950
    },
    {
      "epoch": 7.534172855451512,
      "grad_norm": 0.7568729519844055,
      "learning_rate": 0.00014830298459217757,
      "loss": 3.245,
      "step": 70000
    },
    {
      "epoch": 7.534172855451512,
      "eval_accuracy": 0.3881374515530355,
      "eval_loss": 3.3511228561401367,
      "eval_runtime": 184.6187,
      "eval_samples_per_second": 97.558,
      "eval_steps_per_second": 6.099,
      "step": 70000
    },
    {
      "epoch": 7.539554407491121,
      "grad_norm": 0.8094336986541748,
      "learning_rate": 0.00014797974356211614,
      "loss": 3.2538,
      "step": 70050
    },
    {
      "epoch": 7.544935959530728,
      "grad_norm": 0.8163277506828308,
      "learning_rate": 0.00014765650253205473,
      "loss": 3.222,
      "step": 70100
    },
    {
      "epoch": 7.550317511570337,
      "grad_norm": 0.7944953441619873,
      "learning_rate": 0.0001473332615019933,
      "loss": 3.2391,
      "step": 70150
    },
    {
      "epoch": 7.5556990636099455,
      "grad_norm": 0.763783872127533,
      "learning_rate": 0.0001470100204719319,
      "loss": 3.234,
      "step": 70200
    },
    {
      "epoch": 7.561080615649553,
      "grad_norm": 0.7599479556083679,
      "learning_rate": 0.00014668677944187046,
      "loss": 3.2295,
      "step": 70250
    },
    {
      "epoch": 7.566462167689162,
      "grad_norm": 0.8006331324577332,
      "learning_rate": 0.00014636353841180905,
      "loss": 3.235,
      "step": 70300
    },
    {
      "epoch": 7.57184371972877,
      "grad_norm": 0.7456254363059998,
      "learning_rate": 0.00014604029738174765,
      "loss": 3.2438,
      "step": 70350
    },
    {
      "epoch": 7.577225271768378,
      "grad_norm": 0.7523251175880432,
      "learning_rate": 0.00014571705635168622,
      "loss": 3.2374,
      "step": 70400
    },
    {
      "epoch": 7.5826068238079865,
      "grad_norm": 0.8125523924827576,
      "learning_rate": 0.0001453938153216248,
      "loss": 3.2464,
      "step": 70450
    },
    {
      "epoch": 7.587988375847594,
      "grad_norm": 0.7764886021614075,
      "learning_rate": 0.0001450705742915634,
      "loss": 3.2444,
      "step": 70500
    },
    {
      "epoch": 7.593369927887203,
      "grad_norm": 0.7603163719177246,
      "learning_rate": 0.00014474733326150197,
      "loss": 3.2265,
      "step": 70550
    },
    {
      "epoch": 7.598751479926811,
      "grad_norm": 0.7764590978622437,
      "learning_rate": 0.00014442409223144057,
      "loss": 3.239,
      "step": 70600
    },
    {
      "epoch": 7.604133031966419,
      "grad_norm": 0.8062946200370789,
      "learning_rate": 0.00014410085120137916,
      "loss": 3.235,
      "step": 70650
    },
    {
      "epoch": 7.609514584006027,
      "grad_norm": 0.8161317110061646,
      "learning_rate": 0.00014377761017131773,
      "loss": 3.2576,
      "step": 70700
    },
    {
      "epoch": 7.614896136045635,
      "grad_norm": 0.8046597242355347,
      "learning_rate": 0.00014345436914125633,
      "loss": 3.2422,
      "step": 70750
    },
    {
      "epoch": 7.620277688085244,
      "grad_norm": 0.78916996717453,
      "learning_rate": 0.0001431311281111949,
      "loss": 3.2467,
      "step": 70800
    },
    {
      "epoch": 7.625659240124852,
      "grad_norm": 0.8157759308815002,
      "learning_rate": 0.0001428078870811335,
      "loss": 3.235,
      "step": 70850
    },
    {
      "epoch": 7.63104079216446,
      "grad_norm": 0.884231448173523,
      "learning_rate": 0.00014248464605107206,
      "loss": 3.2428,
      "step": 70900
    },
    {
      "epoch": 7.636422344204068,
      "grad_norm": 0.7843464612960815,
      "learning_rate": 0.00014216140502101065,
      "loss": 3.2351,
      "step": 70950
    },
    {
      "epoch": 7.641803896243677,
      "grad_norm": 0.7665687203407288,
      "learning_rate": 0.00014183816399094924,
      "loss": 3.2489,
      "step": 71000
    },
    {
      "epoch": 7.641803896243677,
      "eval_accuracy": 0.38861280793722386,
      "eval_loss": 3.3472373485565186,
      "eval_runtime": 184.5662,
      "eval_samples_per_second": 97.586,
      "eval_steps_per_second": 6.101,
      "step": 71000
    },
    {
      "epoch": 7.647185448283285,
      "grad_norm": 0.8148714303970337,
      "learning_rate": 0.0001415149229608878,
      "loss": 3.2657,
      "step": 71050
    },
    {
      "epoch": 7.652567000322893,
      "grad_norm": 0.7591752409934998,
      "learning_rate": 0.0001411916819308264,
      "loss": 3.2508,
      "step": 71100
    },
    {
      "epoch": 7.657948552362502,
      "grad_norm": 0.852451741695404,
      "learning_rate": 0.000140868440900765,
      "loss": 3.2269,
      "step": 71150
    },
    {
      "epoch": 7.663330104402109,
      "grad_norm": 0.8520247340202332,
      "learning_rate": 0.00014054519987070357,
      "loss": 3.23,
      "step": 71200
    },
    {
      "epoch": 7.668711656441718,
      "grad_norm": 0.772911012172699,
      "learning_rate": 0.00014022842366124338,
      "loss": 3.2266,
      "step": 71250
    },
    {
      "epoch": 7.674093208481326,
      "grad_norm": 0.8288286328315735,
      "learning_rate": 0.00013990518263118197,
      "loss": 3.2666,
      "step": 71300
    },
    {
      "epoch": 7.679474760520934,
      "grad_norm": 0.8074334263801575,
      "learning_rate": 0.00013958194160112054,
      "loss": 3.2463,
      "step": 71350
    },
    {
      "epoch": 7.684856312560543,
      "grad_norm": 0.7582359313964844,
      "learning_rate": 0.00013925870057105913,
      "loss": 3.2309,
      "step": 71400
    },
    {
      "epoch": 7.69023786460015,
      "grad_norm": 0.826146125793457,
      "learning_rate": 0.00013893545954099773,
      "loss": 3.2457,
      "step": 71450
    },
    {
      "epoch": 7.695619416639759,
      "grad_norm": 0.7911391258239746,
      "learning_rate": 0.0001386122185109363,
      "loss": 3.2356,
      "step": 71500
    },
    {
      "epoch": 7.7010009686793675,
      "grad_norm": 0.7555123567581177,
      "learning_rate": 0.0001382889774808749,
      "loss": 3.2348,
      "step": 71550
    },
    {
      "epoch": 7.706382520718975,
      "grad_norm": 0.8311915993690491,
      "learning_rate": 0.00013796573645081348,
      "loss": 3.2508,
      "step": 71600
    },
    {
      "epoch": 7.711764072758584,
      "grad_norm": 0.8325998783111572,
      "learning_rate": 0.00013764249542075208,
      "loss": 3.2425,
      "step": 71650
    },
    {
      "epoch": 7.717145624798192,
      "grad_norm": 0.7908602952957153,
      "learning_rate": 0.00013731925439069065,
      "loss": 3.2459,
      "step": 71700
    },
    {
      "epoch": 7.7225271768378,
      "grad_norm": 0.8664662837982178,
      "learning_rate": 0.00013699601336062924,
      "loss": 3.2396,
      "step": 71750
    },
    {
      "epoch": 7.727908728877408,
      "grad_norm": 0.8115105628967285,
      "learning_rate": 0.0001366727723305678,
      "loss": 3.2422,
      "step": 71800
    },
    {
      "epoch": 7.733290280917016,
      "grad_norm": 0.856195330619812,
      "learning_rate": 0.0001363495313005064,
      "loss": 3.2422,
      "step": 71850
    },
    {
      "epoch": 7.738671832956625,
      "grad_norm": 0.8175910711288452,
      "learning_rate": 0.00013602629027044497,
      "loss": 3.2374,
      "step": 71900
    },
    {
      "epoch": 7.744053384996233,
      "grad_norm": 0.8474863767623901,
      "learning_rate": 0.00013570304924038357,
      "loss": 3.242,
      "step": 71950
    },
    {
      "epoch": 7.749434937035841,
      "grad_norm": 0.7545700073242188,
      "learning_rate": 0.00013537980821032213,
      "loss": 3.2365,
      "step": 72000
    },
    {
      "epoch": 7.749434937035841,
      "eval_accuracy": 0.3889078005276402,
      "eval_loss": 3.343463659286499,
      "eval_runtime": 184.5257,
      "eval_samples_per_second": 97.607,
      "eval_steps_per_second": 6.102,
      "step": 72000
    },
    {
      "epoch": 7.754816489075449,
      "grad_norm": 0.8176170587539673,
      "learning_rate": 0.00013505656718026073,
      "loss": 3.2485,
      "step": 72050
    },
    {
      "epoch": 7.760198041115058,
      "grad_norm": 0.8031014800071716,
      "learning_rate": 0.00013473332615019932,
      "loss": 3.2443,
      "step": 72100
    },
    {
      "epoch": 7.765579593154666,
      "grad_norm": 0.833129346370697,
      "learning_rate": 0.00013441008512013792,
      "loss": 3.2327,
      "step": 72150
    },
    {
      "epoch": 7.770961145194274,
      "grad_norm": 0.8670672178268433,
      "learning_rate": 0.00013408684409007648,
      "loss": 3.2409,
      "step": 72200
    },
    {
      "epoch": 7.776342697233883,
      "grad_norm": 0.7718544602394104,
      "learning_rate": 0.00013376360306001508,
      "loss": 3.2471,
      "step": 72250
    },
    {
      "epoch": 7.78172424927349,
      "grad_norm": 0.7775774002075195,
      "learning_rate": 0.00013344036202995367,
      "loss": 3.233,
      "step": 72300
    },
    {
      "epoch": 7.787105801313099,
      "grad_norm": 0.8672437071800232,
      "learning_rate": 0.00013311712099989224,
      "loss": 3.2524,
      "step": 72350
    },
    {
      "epoch": 7.792487353352707,
      "grad_norm": 0.8262726664543152,
      "learning_rate": 0.00013279387996983084,
      "loss": 3.2372,
      "step": 72400
    },
    {
      "epoch": 7.797868905392315,
      "grad_norm": 0.8260082006454468,
      "learning_rate": 0.0001324706389397694,
      "loss": 3.2373,
      "step": 72450
    },
    {
      "epoch": 7.803250457431924,
      "grad_norm": 0.8064572215080261,
      "learning_rate": 0.000132147397909708,
      "loss": 3.2338,
      "step": 72500
    },
    {
      "epoch": 7.808632009471531,
      "grad_norm": 0.7858029007911682,
      "learning_rate": 0.00013182415687964657,
      "loss": 3.2383,
      "step": 72550
    },
    {
      "epoch": 7.81401356151114,
      "grad_norm": 0.8511091470718384,
      "learning_rate": 0.00013150091584958516,
      "loss": 3.2425,
      "step": 72600
    },
    {
      "epoch": 7.819395113550748,
      "grad_norm": 0.7787911295890808,
      "learning_rate": 0.00013117767481952375,
      "loss": 3.2536,
      "step": 72650
    },
    {
      "epoch": 7.824776665590356,
      "grad_norm": 0.8475034236907959,
      "learning_rate": 0.00013085443378946232,
      "loss": 3.2282,
      "step": 72700
    },
    {
      "epoch": 7.830158217629965,
      "grad_norm": 0.7796977162361145,
      "learning_rate": 0.00013053119275940092,
      "loss": 3.2345,
      "step": 72750
    },
    {
      "epoch": 7.835539769669572,
      "grad_norm": 0.8430459499359131,
      "learning_rate": 0.0001302079517293395,
      "loss": 3.2574,
      "step": 72800
    },
    {
      "epoch": 7.840921321709181,
      "grad_norm": 0.7826714515686035,
      "learning_rate": 0.00012988471069927808,
      "loss": 3.2304,
      "step": 72850
    },
    {
      "epoch": 7.846302873748789,
      "grad_norm": 0.7714261412620544,
      "learning_rate": 0.00012956146966921667,
      "loss": 3.2285,
      "step": 72900
    },
    {
      "epoch": 7.851684425788397,
      "grad_norm": 0.8180266618728638,
      "learning_rate": 0.00012923822863915527,
      "loss": 3.2535,
      "step": 72950
    },
    {
      "epoch": 7.857065977828006,
      "grad_norm": 0.7998464107513428,
      "learning_rate": 0.00012891498760909384,
      "loss": 3.257,
      "step": 73000
    },
    {
      "epoch": 7.857065977828006,
      "eval_accuracy": 0.38929525672558657,
      "eval_loss": 3.3379297256469727,
      "eval_runtime": 184.5134,
      "eval_samples_per_second": 97.614,
      "eval_steps_per_second": 6.103,
      "step": 73000
    },
    {
      "epoch": 7.862447529867614,
      "grad_norm": 0.7833364009857178,
      "learning_rate": 0.00012859174657903243,
      "loss": 3.2533,
      "step": 73050
    },
    {
      "epoch": 7.867829081907222,
      "grad_norm": 0.7800090312957764,
      "learning_rate": 0.000128268505548971,
      "loss": 3.2317,
      "step": 73100
    },
    {
      "epoch": 7.87321063394683,
      "grad_norm": 0.8276484608650208,
      "learning_rate": 0.0001279452645189096,
      "loss": 3.2524,
      "step": 73150
    },
    {
      "epoch": 7.878592185986438,
      "grad_norm": 0.868791401386261,
      "learning_rate": 0.00012762202348884816,
      "loss": 3.2484,
      "step": 73200
    },
    {
      "epoch": 7.883973738026047,
      "grad_norm": 0.8134067058563232,
      "learning_rate": 0.00012729878245878676,
      "loss": 3.2393,
      "step": 73250
    },
    {
      "epoch": 7.889355290065655,
      "grad_norm": 0.8356450796127319,
      "learning_rate": 0.00012697554142872535,
      "loss": 3.2457,
      "step": 73300
    },
    {
      "epoch": 7.894736842105263,
      "grad_norm": 0.8295549154281616,
      "learning_rate": 0.00012665230039866392,
      "loss": 3.2525,
      "step": 73350
    },
    {
      "epoch": 7.900118394144871,
      "grad_norm": 0.8256616592407227,
      "learning_rate": 0.0001263290593686025,
      "loss": 3.2487,
      "step": 73400
    },
    {
      "epoch": 7.90549994618448,
      "grad_norm": 0.8164888024330139,
      "learning_rate": 0.0001260058183385411,
      "loss": 3.2414,
      "step": 73450
    },
    {
      "epoch": 7.910881498224088,
      "grad_norm": 0.8514302968978882,
      "learning_rate": 0.00012568257730847967,
      "loss": 3.2231,
      "step": 73500
    },
    {
      "epoch": 7.916263050263696,
      "grad_norm": 0.844474196434021,
      "learning_rate": 0.00012535933627841827,
      "loss": 3.2373,
      "step": 73550
    },
    {
      "epoch": 7.921644602303305,
      "grad_norm": 0.8269355297088623,
      "learning_rate": 0.00012503609524835684,
      "loss": 3.2418,
      "step": 73600
    },
    {
      "epoch": 7.927026154342912,
      "grad_norm": 0.8630267381668091,
      "learning_rate": 0.00012471285421829543,
      "loss": 3.2377,
      "step": 73650
    },
    {
      "epoch": 7.932407706382521,
      "grad_norm": 0.8255678415298462,
      "learning_rate": 0.000124389613188234,
      "loss": 3.2437,
      "step": 73700
    },
    {
      "epoch": 7.937789258422129,
      "grad_norm": 0.830707848072052,
      "learning_rate": 0.0001240663721581726,
      "loss": 3.2332,
      "step": 73750
    },
    {
      "epoch": 7.943170810461737,
      "grad_norm": 0.8320891857147217,
      "learning_rate": 0.0001237431311281112,
      "loss": 3.2331,
      "step": 73800
    },
    {
      "epoch": 7.948552362501346,
      "grad_norm": 0.808674156665802,
      "learning_rate": 0.00012341989009804976,
      "loss": 3.2446,
      "step": 73850
    },
    {
      "epoch": 7.953933914540953,
      "grad_norm": 0.8536055684089661,
      "learning_rate": 0.00012309664906798835,
      "loss": 3.243,
      "step": 73900
    },
    {
      "epoch": 7.959315466580562,
      "grad_norm": 0.7962849140167236,
      "learning_rate": 0.00012277340803792694,
      "loss": 3.2448,
      "step": 73950
    },
    {
      "epoch": 7.96469701862017,
      "grad_norm": 0.8137184977531433,
      "learning_rate": 0.0001224501670078655,
      "loss": 3.2408,
      "step": 74000
    },
    {
      "epoch": 7.96469701862017,
      "eval_accuracy": 0.3896508776274034,
      "eval_loss": 3.335071325302124,
      "eval_runtime": 184.5058,
      "eval_samples_per_second": 97.618,
      "eval_steps_per_second": 6.103,
      "step": 74000
    },
    {
      "epoch": 7.970078570659778,
      "grad_norm": 0.813854455947876,
      "learning_rate": 0.0001221269259778041,
      "loss": 3.2279,
      "step": 74050
    },
    {
      "epoch": 7.975460122699387,
      "grad_norm": 0.8088878393173218,
      "learning_rate": 0.00012180368494774269,
      "loss": 3.2451,
      "step": 74100
    },
    {
      "epoch": 7.980841674738995,
      "grad_norm": 0.8375864624977112,
      "learning_rate": 0.00012148044391768128,
      "loss": 3.2295,
      "step": 74150
    },
    {
      "epoch": 7.986223226778603,
      "grad_norm": 0.848151683807373,
      "learning_rate": 0.00012115720288761985,
      "loss": 3.2395,
      "step": 74200
    },
    {
      "epoch": 7.991604778818211,
      "grad_norm": 0.8360744714736938,
      "learning_rate": 0.00012083396185755845,
      "loss": 3.2383,
      "step": 74250
    },
    {
      "epoch": 7.996986330857819,
      "grad_norm": 0.8315327167510986,
      "learning_rate": 0.00012051072082749703,
      "loss": 3.2525,
      "step": 74300
    },
    {
      "epoch": 8.002367882897428,
      "grad_norm": 0.9633062481880188,
      "learning_rate": 0.00012018747979743561,
      "loss": 3.2006,
      "step": 74350
    },
    {
      "epoch": 8.007749434937036,
      "grad_norm": 0.8674287796020508,
      "learning_rate": 0.00011986423876737419,
      "loss": 3.1705,
      "step": 74400
    },
    {
      "epoch": 8.013130986976645,
      "grad_norm": 0.7837016582489014,
      "learning_rate": 0.00011954099773731278,
      "loss": 3.1522,
      "step": 74450
    },
    {
      "epoch": 8.018512539016251,
      "grad_norm": 0.8284465670585632,
      "learning_rate": 0.00011921775670725135,
      "loss": 3.1729,
      "step": 74500
    },
    {
      "epoch": 8.02389409105586,
      "grad_norm": 0.7582801580429077,
      "learning_rate": 0.00011889451567718995,
      "loss": 3.1641,
      "step": 74550
    },
    {
      "epoch": 8.029275643095469,
      "grad_norm": 0.7998579740524292,
      "learning_rate": 0.00011857127464712854,
      "loss": 3.1561,
      "step": 74600
    },
    {
      "epoch": 8.034657195135077,
      "grad_norm": 0.8288883566856384,
      "learning_rate": 0.00011824803361706712,
      "loss": 3.1783,
      "step": 74650
    },
    {
      "epoch": 8.040038747174686,
      "grad_norm": 0.8424205183982849,
      "learning_rate": 0.00011792479258700569,
      "loss": 3.1678,
      "step": 74700
    },
    {
      "epoch": 8.045420299214294,
      "grad_norm": 0.8138177990913391,
      "learning_rate": 0.00011760155155694428,
      "loss": 3.1687,
      "step": 74750
    },
    {
      "epoch": 8.050801851253901,
      "grad_norm": 0.8044741749763489,
      "learning_rate": 0.00011727831052688288,
      "loss": 3.1652,
      "step": 74800
    },
    {
      "epoch": 8.05618340329351,
      "grad_norm": 0.857967734336853,
      "learning_rate": 0.00011695506949682145,
      "loss": 3.1832,
      "step": 74850
    },
    {
      "epoch": 8.061564955333118,
      "grad_norm": 0.8203487396240234,
      "learning_rate": 0.00011663182846676004,
      "loss": 3.1799,
      "step": 74900
    },
    {
      "epoch": 8.066946507372727,
      "grad_norm": 0.793319046497345,
      "learning_rate": 0.00011630858743669862,
      "loss": 3.1667,
      "step": 74950
    },
    {
      "epoch": 8.072328059412335,
      "grad_norm": 0.8443703055381775,
      "learning_rate": 0.0001159853464066372,
      "loss": 3.1993,
      "step": 75000
    },
    {
      "epoch": 8.072328059412335,
      "eval_accuracy": 0.389680974477328,
      "eval_loss": 3.34047532081604,
      "eval_runtime": 184.5769,
      "eval_samples_per_second": 97.58,
      "eval_steps_per_second": 6.1,
      "step": 75000
    },
    {
      "epoch": 8.077709611451942,
      "grad_norm": 0.8433181047439575,
      "learning_rate": 0.00011566210537657578,
      "loss": 3.159,
      "step": 75050
    },
    {
      "epoch": 8.08309116349155,
      "grad_norm": 0.7938767671585083,
      "learning_rate": 0.00011533886434651438,
      "loss": 3.1652,
      "step": 75100
    },
    {
      "epoch": 8.088472715531159,
      "grad_norm": 0.7978476881980896,
      "learning_rate": 0.00011501562331645296,
      "loss": 3.1829,
      "step": 75150
    },
    {
      "epoch": 8.093854267570768,
      "grad_norm": 0.8397863507270813,
      "learning_rate": 0.00011469238228639154,
      "loss": 3.1829,
      "step": 75200
    },
    {
      "epoch": 8.099235819610376,
      "grad_norm": 0.8552531003952026,
      "learning_rate": 0.00011437560607693136,
      "loss": 3.1686,
      "step": 75250
    },
    {
      "epoch": 8.104617371649983,
      "grad_norm": 0.834924578666687,
      "learning_rate": 0.00011405236504686993,
      "loss": 3.1777,
      "step": 75300
    },
    {
      "epoch": 8.109998923689592,
      "grad_norm": 0.8314476013183594,
      "learning_rate": 0.00011372912401680852,
      "loss": 3.1849,
      "step": 75350
    },
    {
      "epoch": 8.1153804757292,
      "grad_norm": 0.8211588263511658,
      "learning_rate": 0.0001134058829867471,
      "loss": 3.183,
      "step": 75400
    },
    {
      "epoch": 8.120762027768809,
      "grad_norm": 0.8183987736701965,
      "learning_rate": 0.0001130826419566857,
      "loss": 3.1799,
      "step": 75450
    },
    {
      "epoch": 8.126143579808417,
      "grad_norm": 0.8008273243904114,
      "learning_rate": 0.00011275940092662427,
      "loss": 3.184,
      "step": 75500
    },
    {
      "epoch": 8.131525131848026,
      "grad_norm": 0.7898052930831909,
      "learning_rate": 0.00011243615989656286,
      "loss": 3.1602,
      "step": 75550
    },
    {
      "epoch": 8.136906683887632,
      "grad_norm": 0.8465813994407654,
      "learning_rate": 0.00011211291886650146,
      "loss": 3.1834,
      "step": 75600
    },
    {
      "epoch": 8.142288235927241,
      "grad_norm": 0.8261603116989136,
      "learning_rate": 0.00011178967783644002,
      "loss": 3.1726,
      "step": 75650
    },
    {
      "epoch": 8.14766978796685,
      "grad_norm": 0.8099297285079956,
      "learning_rate": 0.00011146643680637862,
      "loss": 3.1889,
      "step": 75700
    },
    {
      "epoch": 8.153051340006458,
      "grad_norm": 0.7958865165710449,
      "learning_rate": 0.0001111431957763172,
      "loss": 3.1791,
      "step": 75750
    },
    {
      "epoch": 8.158432892046067,
      "grad_norm": 0.8619468808174133,
      "learning_rate": 0.00011081995474625578,
      "loss": 3.1754,
      "step": 75800
    },
    {
      "epoch": 8.163814444085673,
      "grad_norm": 0.8686273694038391,
      "learning_rate": 0.00011049671371619436,
      "loss": 3.1681,
      "step": 75850
    },
    {
      "epoch": 8.169195996125282,
      "grad_norm": 0.8688459396362305,
      "learning_rate": 0.00011017347268613296,
      "loss": 3.1987,
      "step": 75900
    },
    {
      "epoch": 8.17457754816489,
      "grad_norm": 0.8701615333557129,
      "learning_rate": 0.00010985023165607154,
      "loss": 3.1818,
      "step": 75950
    },
    {
      "epoch": 8.1799591002045,
      "grad_norm": 0.8842565417289734,
      "learning_rate": 0.00010952699062601012,
      "loss": 3.1666,
      "step": 76000
    },
    {
      "epoch": 8.1799591002045,
      "eval_accuracy": 0.3899865063978623,
      "eval_loss": 3.339653253555298,
      "eval_runtime": 184.3667,
      "eval_samples_per_second": 97.691,
      "eval_steps_per_second": 6.107,
      "step": 76000
    },
    {
      "epoch": 8.185340652244108,
      "grad_norm": 0.813447117805481,
      "learning_rate": 0.0001092037495959487,
      "loss": 3.1859,
      "step": 76050
    },
    {
      "epoch": 8.190722204283716,
      "grad_norm": 0.8704348802566528,
      "learning_rate": 0.0001088805085658873,
      "loss": 3.1811,
      "step": 76100
    },
    {
      "epoch": 8.196103756323323,
      "grad_norm": 0.8403463363647461,
      "learning_rate": 0.00010855726753582586,
      "loss": 3.1763,
      "step": 76150
    },
    {
      "epoch": 8.201485308362932,
      "grad_norm": 0.8404145240783691,
      "learning_rate": 0.00010823402650576446,
      "loss": 3.1909,
      "step": 76200
    },
    {
      "epoch": 8.20686686040254,
      "grad_norm": 0.9068716168403625,
      "learning_rate": 0.00010791078547570304,
      "loss": 3.1827,
      "step": 76250
    },
    {
      "epoch": 8.212248412442149,
      "grad_norm": 0.8582590222358704,
      "learning_rate": 0.00010758754444564162,
      "loss": 3.1786,
      "step": 76300
    },
    {
      "epoch": 8.217629964481757,
      "grad_norm": 0.8304372429847717,
      "learning_rate": 0.0001072643034155802,
      "loss": 3.1769,
      "step": 76350
    },
    {
      "epoch": 8.223011516521364,
      "grad_norm": 0.8133452534675598,
      "learning_rate": 0.0001069410623855188,
      "loss": 3.1915,
      "step": 76400
    },
    {
      "epoch": 8.228393068560973,
      "grad_norm": 0.8031824827194214,
      "learning_rate": 0.00010661782135545739,
      "loss": 3.1815,
      "step": 76450
    },
    {
      "epoch": 8.233774620600581,
      "grad_norm": 0.8759872317314148,
      "learning_rate": 0.00010629458032539596,
      "loss": 3.1886,
      "step": 76500
    },
    {
      "epoch": 8.23915617264019,
      "grad_norm": 0.7917901277542114,
      "learning_rate": 0.00010597133929533455,
      "loss": 3.1773,
      "step": 76550
    },
    {
      "epoch": 8.244537724679798,
      "grad_norm": 0.8822852373123169,
      "learning_rate": 0.00010564809826527313,
      "loss": 3.1758,
      "step": 76600
    },
    {
      "epoch": 8.249919276719407,
      "grad_norm": 0.8366295695304871,
      "learning_rate": 0.00010532485723521171,
      "loss": 3.1796,
      "step": 76650
    },
    {
      "epoch": 8.255300828759013,
      "grad_norm": 0.8097661137580872,
      "learning_rate": 0.0001050016162051503,
      "loss": 3.1813,
      "step": 76700
    },
    {
      "epoch": 8.260682380798622,
      "grad_norm": 0.8562833070755005,
      "learning_rate": 0.00010467837517508889,
      "loss": 3.1943,
      "step": 76750
    },
    {
      "epoch": 8.26606393283823,
      "grad_norm": 0.8109206557273865,
      "learning_rate": 0.00010435513414502746,
      "loss": 3.1877,
      "step": 76800
    },
    {
      "epoch": 8.27144548487784,
      "grad_norm": 0.8438268899917603,
      "learning_rate": 0.00010403189311496605,
      "loss": 3.1943,
      "step": 76850
    },
    {
      "epoch": 8.276827036917448,
      "grad_norm": 0.8549637794494629,
      "learning_rate": 0.00010370865208490463,
      "loss": 3.1922,
      "step": 76900
    },
    {
      "epoch": 8.282208588957054,
      "grad_norm": 0.8351240158081055,
      "learning_rate": 0.00010338541105484323,
      "loss": 3.1827,
      "step": 76950
    },
    {
      "epoch": 8.287590140996663,
      "grad_norm": 0.868952751159668,
      "learning_rate": 0.0001030621700247818,
      "loss": 3.1768,
      "step": 77000
    },
    {
      "epoch": 8.287590140996663,
      "eval_accuracy": 0.3901297109040018,
      "eval_loss": 3.336033344268799,
      "eval_runtime": 184.6915,
      "eval_samples_per_second": 97.519,
      "eval_steps_per_second": 6.097,
      "step": 77000
    },
    {
      "epoch": 8.292971693036272,
      "grad_norm": 0.8799475431442261,
      "learning_rate": 0.00010273892899472039,
      "loss": 3.1844,
      "step": 77050
    },
    {
      "epoch": 8.29835324507588,
      "grad_norm": 0.8950610756874084,
      "learning_rate": 0.00010241568796465898,
      "loss": 3.1841,
      "step": 77100
    },
    {
      "epoch": 8.303734797115489,
      "grad_norm": 0.9824919700622559,
      "learning_rate": 0.00010209244693459755,
      "loss": 3.1847,
      "step": 77150
    },
    {
      "epoch": 8.309116349155097,
      "grad_norm": 0.8694815635681152,
      "learning_rate": 0.00010176920590453613,
      "loss": 3.1749,
      "step": 77200
    },
    {
      "epoch": 8.314497901194704,
      "grad_norm": 0.7845666408538818,
      "learning_rate": 0.00010145242969507597,
      "loss": 3.1894,
      "step": 77250
    },
    {
      "epoch": 8.319879453234313,
      "grad_norm": 0.856328547000885,
      "learning_rate": 0.00010112918866501453,
      "loss": 3.2064,
      "step": 77300
    },
    {
      "epoch": 8.325261005273921,
      "grad_norm": 0.8453772068023682,
      "learning_rate": 0.00010080594763495313,
      "loss": 3.2156,
      "step": 77350
    },
    {
      "epoch": 8.33064255731353,
      "grad_norm": 0.8454362154006958,
      "learning_rate": 0.00010048270660489171,
      "loss": 3.1828,
      "step": 77400
    },
    {
      "epoch": 8.336024109353138,
      "grad_norm": 0.8105871081352234,
      "learning_rate": 0.00010015946557483029,
      "loss": 3.1862,
      "step": 77450
    },
    {
      "epoch": 8.341405661392745,
      "grad_norm": 0.8413558602333069,
      "learning_rate": 9.983622454476887e-05,
      "loss": 3.1943,
      "step": 77500
    },
    {
      "epoch": 8.346787213432354,
      "grad_norm": 0.8694208860397339,
      "learning_rate": 9.951298351470747e-05,
      "loss": 3.1938,
      "step": 77550
    },
    {
      "epoch": 8.352168765471962,
      "grad_norm": 0.8308895230293274,
      "learning_rate": 9.918974248464603e-05,
      "loss": 3.1739,
      "step": 77600
    },
    {
      "epoch": 8.35755031751157,
      "grad_norm": 0.8393214344978333,
      "learning_rate": 9.886650145458463e-05,
      "loss": 3.1834,
      "step": 77650
    },
    {
      "epoch": 8.36293186955118,
      "grad_norm": 0.8121619820594788,
      "learning_rate": 9.854326042452321e-05,
      "loss": 3.1891,
      "step": 77700
    },
    {
      "epoch": 8.368313421590786,
      "grad_norm": 0.8386716842651367,
      "learning_rate": 9.82200193944618e-05,
      "loss": 3.1849,
      "step": 77750
    },
    {
      "epoch": 8.373694973630395,
      "grad_norm": 0.8557778000831604,
      "learning_rate": 9.789677836440037e-05,
      "loss": 3.1915,
      "step": 77800
    },
    {
      "epoch": 8.379076525670003,
      "grad_norm": 0.9088544845581055,
      "learning_rate": 9.757353733433897e-05,
      "loss": 3.189,
      "step": 77850
    },
    {
      "epoch": 8.384458077709612,
      "grad_norm": 0.8937007784843445,
      "learning_rate": 9.725029630427755e-05,
      "loss": 3.2055,
      "step": 77900
    },
    {
      "epoch": 8.38983962974922,
      "grad_norm": 0.888757050037384,
      "learning_rate": 9.692705527421613e-05,
      "loss": 3.1871,
      "step": 77950
    },
    {
      "epoch": 8.395221181788829,
      "grad_norm": 0.8638023734092712,
      "learning_rate": 9.660381424415471e-05,
      "loss": 3.1937,
      "step": 78000
    },
    {
      "epoch": 8.395221181788829,
      "eval_accuracy": 0.3904857664173699,
      "eval_loss": 3.3330600261688232,
      "eval_runtime": 184.6364,
      "eval_samples_per_second": 97.548,
      "eval_steps_per_second": 6.098,
      "step": 78000
    },
    {
      "epoch": 8.400602733828435,
      "grad_norm": 0.8812534213066101,
      "learning_rate": 9.62805732140933e-05,
      "loss": 3.1894,
      "step": 78050
    },
    {
      "epoch": 8.405984285868044,
      "grad_norm": 0.8491891026496887,
      "learning_rate": 9.595733218403187e-05,
      "loss": 3.1968,
      "step": 78100
    },
    {
      "epoch": 8.411365837907653,
      "grad_norm": 0.8361186385154724,
      "learning_rate": 9.563409115397047e-05,
      "loss": 3.1931,
      "step": 78150
    },
    {
      "epoch": 8.416747389947261,
      "grad_norm": 0.8706521987915039,
      "learning_rate": 9.531085012390906e-05,
      "loss": 3.1675,
      "step": 78200
    },
    {
      "epoch": 8.42212894198687,
      "grad_norm": 0.8534976243972778,
      "learning_rate": 9.498760909384764e-05,
      "loss": 3.1868,
      "step": 78250
    },
    {
      "epoch": 8.427510494026476,
      "grad_norm": 0.8684397339820862,
      "learning_rate": 9.466436806378622e-05,
      "loss": 3.1767,
      "step": 78300
    },
    {
      "epoch": 8.432892046066085,
      "grad_norm": 0.8318513631820679,
      "learning_rate": 9.43411270337248e-05,
      "loss": 3.1832,
      "step": 78350
    },
    {
      "epoch": 8.438273598105694,
      "grad_norm": 0.8551844358444214,
      "learning_rate": 9.40178860036634e-05,
      "loss": 3.1831,
      "step": 78400
    },
    {
      "epoch": 8.443655150145302,
      "grad_norm": 0.8822528719902039,
      "learning_rate": 9.369464497360197e-05,
      "loss": 3.1753,
      "step": 78450
    },
    {
      "epoch": 8.44903670218491,
      "grad_norm": 0.85441654920578,
      "learning_rate": 9.337140394354056e-05,
      "loss": 3.1875,
      "step": 78500
    },
    {
      "epoch": 8.45441825422452,
      "grad_norm": 0.839636504650116,
      "learning_rate": 9.304816291347914e-05,
      "loss": 3.1873,
      "step": 78550
    },
    {
      "epoch": 8.459799806264126,
      "grad_norm": 0.8585395812988281,
      "learning_rate": 9.272492188341772e-05,
      "loss": 3.189,
      "step": 78600
    },
    {
      "epoch": 8.465181358303735,
      "grad_norm": 0.8621167540550232,
      "learning_rate": 9.24016808533563e-05,
      "loss": 3.1981,
      "step": 78650
    },
    {
      "epoch": 8.470562910343343,
      "grad_norm": 0.8876769542694092,
      "learning_rate": 9.20784398232949e-05,
      "loss": 3.1977,
      "step": 78700
    },
    {
      "epoch": 8.475944462382952,
      "grad_norm": 0.8398529887199402,
      "learning_rate": 9.175519879323348e-05,
      "loss": 3.1939,
      "step": 78750
    },
    {
      "epoch": 8.48132601442256,
      "grad_norm": 0.845101535320282,
      "learning_rate": 9.143195776317206e-05,
      "loss": 3.1723,
      "step": 78800
    },
    {
      "epoch": 8.486707566462167,
      "grad_norm": 0.8674041628837585,
      "learning_rate": 9.110871673311064e-05,
      "loss": 3.1832,
      "step": 78850
    },
    {
      "epoch": 8.492089118501776,
      "grad_norm": 0.8334956169128418,
      "learning_rate": 9.078547570304924e-05,
      "loss": 3.2087,
      "step": 78900
    },
    {
      "epoch": 8.497470670541384,
      "grad_norm": 0.8290042877197266,
      "learning_rate": 9.04622346729878e-05,
      "loss": 3.203,
      "step": 78950
    },
    {
      "epoch": 8.502852222580993,
      "grad_norm": 0.8271029591560364,
      "learning_rate": 9.01389936429264e-05,
      "loss": 3.1784,
      "step": 79000
    },
    {
      "epoch": 8.502852222580993,
      "eval_accuracy": 0.3908978868208502,
      "eval_loss": 3.3294878005981445,
      "eval_runtime": 184.7902,
      "eval_samples_per_second": 97.467,
      "eval_steps_per_second": 6.093,
      "step": 79000
    },
    {
      "epoch": 8.508233774620601,
      "grad_norm": 0.8316333293914795,
      "learning_rate": 8.9815752612865e-05,
      "loss": 3.183,
      "step": 79050
    },
    {
      "epoch": 8.513615326660208,
      "grad_norm": 0.8319535255432129,
      "learning_rate": 8.949251158280356e-05,
      "loss": 3.1854,
      "step": 79100
    },
    {
      "epoch": 8.518996878699816,
      "grad_norm": 0.8060797452926636,
      "learning_rate": 8.916927055274216e-05,
      "loss": 3.1814,
      "step": 79150
    },
    {
      "epoch": 8.524378430739425,
      "grad_norm": 0.8370040655136108,
      "learning_rate": 8.884602952268074e-05,
      "loss": 3.1845,
      "step": 79200
    },
    {
      "epoch": 8.529759982779034,
      "grad_norm": 0.8340920209884644,
      "learning_rate": 8.852925331322054e-05,
      "loss": 3.1927,
      "step": 79250
    },
    {
      "epoch": 8.535141534818642,
      "grad_norm": 0.835109293460846,
      "learning_rate": 8.820601228315914e-05,
      "loss": 3.193,
      "step": 79300
    },
    {
      "epoch": 8.54052308685825,
      "grad_norm": 0.8529259562492371,
      "learning_rate": 8.788923607369896e-05,
      "loss": 3.1944,
      "step": 79350
    },
    {
      "epoch": 8.545904638897857,
      "grad_norm": 0.9165598750114441,
      "learning_rate": 8.756599504363753e-05,
      "loss": 3.1867,
      "step": 79400
    },
    {
      "epoch": 8.551286190937466,
      "grad_norm": 0.8577670454978943,
      "learning_rate": 8.724275401357612e-05,
      "loss": 3.1956,
      "step": 79450
    },
    {
      "epoch": 8.556667742977075,
      "grad_norm": 0.8654495477676392,
      "learning_rate": 8.69195129835147e-05,
      "loss": 3.2014,
      "step": 79500
    },
    {
      "epoch": 8.562049295016683,
      "grad_norm": 0.8619284629821777,
      "learning_rate": 8.659627195345328e-05,
      "loss": 3.1908,
      "step": 79550
    },
    {
      "epoch": 8.567430847056292,
      "grad_norm": 0.8602727651596069,
      "learning_rate": 8.627303092339186e-05,
      "loss": 3.1921,
      "step": 79600
    },
    {
      "epoch": 8.572812399095898,
      "grad_norm": 0.8377135992050171,
      "learning_rate": 8.594978989333046e-05,
      "loss": 3.1961,
      "step": 79650
    },
    {
      "epoch": 8.578193951135507,
      "grad_norm": 0.8515350222587585,
      "learning_rate": 8.562654886326903e-05,
      "loss": 3.1863,
      "step": 79700
    },
    {
      "epoch": 8.583575503175116,
      "grad_norm": 0.8238023519515991,
      "learning_rate": 8.530330783320762e-05,
      "loss": 3.1848,
      "step": 79750
    },
    {
      "epoch": 8.588957055214724,
      "grad_norm": 0.8730199337005615,
      "learning_rate": 8.49800668031462e-05,
      "loss": 3.1919,
      "step": 79800
    },
    {
      "epoch": 8.594338607254333,
      "grad_norm": 0.851908802986145,
      "learning_rate": 8.46568257730848e-05,
      "loss": 3.1963,
      "step": 79850
    },
    {
      "epoch": 8.599720159293941,
      "grad_norm": 0.7989451885223389,
      "learning_rate": 8.433358474302336e-05,
      "loss": 3.1989,
      "step": 79900
    },
    {
      "epoch": 8.605101711333548,
      "grad_norm": 0.8555760979652405,
      "learning_rate": 8.401034371296196e-05,
      "loss": 3.1831,
      "step": 79950
    },
    {
      "epoch": 8.610483263373157,
      "grad_norm": 0.9119009971618652,
      "learning_rate": 8.368710268290055e-05,
      "loss": 3.1774,
      "step": 80000
    },
    {
      "epoch": 8.610483263373157,
      "eval_accuracy": 0.3912303646575625,
      "eval_loss": 3.3266663551330566,
      "eval_runtime": 184.4493,
      "eval_samples_per_second": 97.647,
      "eval_steps_per_second": 6.105,
      "step": 80000
    },
    {
      "epoch": 8.615864815412765,
      "grad_norm": 0.8387112617492676,
      "learning_rate": 8.336386165283912e-05,
      "loss": 3.1756,
      "step": 80050
    },
    {
      "epoch": 8.621246367452374,
      "grad_norm": 0.8585613965988159,
      "learning_rate": 8.304062062277772e-05,
      "loss": 3.178,
      "step": 80100
    },
    {
      "epoch": 8.626627919491982,
      "grad_norm": 0.88638836145401,
      "learning_rate": 8.27173795927163e-05,
      "loss": 3.1947,
      "step": 80150
    },
    {
      "epoch": 8.632009471531589,
      "grad_norm": 0.8351496458053589,
      "learning_rate": 8.239413856265488e-05,
      "loss": 3.185,
      "step": 80200
    },
    {
      "epoch": 8.637391023571197,
      "grad_norm": 0.89128577709198,
      "learning_rate": 8.207089753259346e-05,
      "loss": 3.1869,
      "step": 80250
    },
    {
      "epoch": 8.642772575610806,
      "grad_norm": 0.8513466715812683,
      "learning_rate": 8.174765650253205e-05,
      "loss": 3.1875,
      "step": 80300
    },
    {
      "epoch": 8.648154127650415,
      "grad_norm": 0.8323937654495239,
      "learning_rate": 8.142441547247063e-05,
      "loss": 3.1854,
      "step": 80350
    },
    {
      "epoch": 8.653535679690023,
      "grad_norm": 0.8293351531028748,
      "learning_rate": 8.110117444240922e-05,
      "loss": 3.1985,
      "step": 80400
    },
    {
      "epoch": 8.658917231729632,
      "grad_norm": 0.8154245018959045,
      "learning_rate": 8.07779334123478e-05,
      "loss": 3.1987,
      "step": 80450
    },
    {
      "epoch": 8.664298783769238,
      "grad_norm": 0.7996413707733154,
      "learning_rate": 8.045469238228639e-05,
      "loss": 3.2052,
      "step": 80500
    },
    {
      "epoch": 8.669680335808847,
      "grad_norm": 0.8286173343658447,
      "learning_rate": 8.013145135222496e-05,
      "loss": 3.1941,
      "step": 80550
    },
    {
      "epoch": 8.675061887848456,
      "grad_norm": 0.8508105278015137,
      "learning_rate": 7.980821032216355e-05,
      "loss": 3.1993,
      "step": 80600
    },
    {
      "epoch": 8.680443439888064,
      "grad_norm": 0.8498315215110779,
      "learning_rate": 7.948496929210213e-05,
      "loss": 3.1933,
      "step": 80650
    },
    {
      "epoch": 8.685824991927673,
      "grad_norm": 0.8828619122505188,
      "learning_rate": 7.916172826204072e-05,
      "loss": 3.2053,
      "step": 80700
    },
    {
      "epoch": 8.69120654396728,
      "grad_norm": 0.8631619215011597,
      "learning_rate": 7.88384872319793e-05,
      "loss": 3.1763,
      "step": 80750
    },
    {
      "epoch": 8.696588096006888,
      "grad_norm": 0.8767603039741516,
      "learning_rate": 7.851524620191789e-05,
      "loss": 3.193,
      "step": 80800
    },
    {
      "epoch": 8.701969648046497,
      "grad_norm": 0.8803954720497131,
      "learning_rate": 7.819200517185649e-05,
      "loss": 3.1915,
      "step": 80850
    },
    {
      "epoch": 8.707351200086105,
      "grad_norm": 0.8561678528785706,
      "learning_rate": 7.786876414179505e-05,
      "loss": 3.194,
      "step": 80900
    },
    {
      "epoch": 8.712732752125714,
      "grad_norm": 0.8392429947853088,
      "learning_rate": 7.754552311173365e-05,
      "loss": 3.1924,
      "step": 80950
    },
    {
      "epoch": 8.718114304165322,
      "grad_norm": 0.8315561413764954,
      "learning_rate": 7.722228208167223e-05,
      "loss": 3.1895,
      "step": 81000
    },
    {
      "epoch": 8.718114304165322,
      "eval_accuracy": 0.3915695789733193,
      "eval_loss": 3.3245222568511963,
      "eval_runtime": 184.4,
      "eval_samples_per_second": 97.674,
      "eval_steps_per_second": 6.106,
      "step": 81000
    },
    {
      "epoch": 8.723495856204929,
      "grad_norm": 0.8995868563652039,
      "learning_rate": 7.689904105161081e-05,
      "loss": 3.1793,
      "step": 81050
    },
    {
      "epoch": 8.728877408244538,
      "grad_norm": 0.8229780793190002,
      "learning_rate": 7.657580002154939e-05,
      "loss": 3.1932,
      "step": 81100
    },
    {
      "epoch": 8.734258960284146,
      "grad_norm": 0.8569908142089844,
      "learning_rate": 7.625255899148799e-05,
      "loss": 3.2011,
      "step": 81150
    },
    {
      "epoch": 8.739640512323755,
      "grad_norm": 0.8527806997299194,
      "learning_rate": 7.592931796142655e-05,
      "loss": 3.1766,
      "step": 81200
    },
    {
      "epoch": 8.745022064363363,
      "grad_norm": 0.8646016716957092,
      "learning_rate": 7.560607693136515e-05,
      "loss": 3.161,
      "step": 81250
    },
    {
      "epoch": 8.75040361640297,
      "grad_norm": 0.871335506439209,
      "learning_rate": 7.528283590130373e-05,
      "loss": 3.2003,
      "step": 81300
    },
    {
      "epoch": 8.755785168442578,
      "grad_norm": 0.8726530075073242,
      "learning_rate": 7.495959487124231e-05,
      "loss": 3.1955,
      "step": 81350
    },
    {
      "epoch": 8.761166720482187,
      "grad_norm": 0.896847665309906,
      "learning_rate": 7.46363538411809e-05,
      "loss": 3.2012,
      "step": 81400
    },
    {
      "epoch": 8.766548272521796,
      "grad_norm": 0.8841673731803894,
      "learning_rate": 7.431311281111949e-05,
      "loss": 3.1901,
      "step": 81450
    },
    {
      "epoch": 8.771929824561404,
      "grad_norm": 0.8515608310699463,
      "learning_rate": 7.398987178105807e-05,
      "loss": 3.1872,
      "step": 81500
    },
    {
      "epoch": 8.777311376601011,
      "grad_norm": 0.8549227714538574,
      "learning_rate": 7.366663075099665e-05,
      "loss": 3.1852,
      "step": 81550
    },
    {
      "epoch": 8.78269292864062,
      "grad_norm": 0.911566436290741,
      "learning_rate": 7.334338972093523e-05,
      "loss": 3.1831,
      "step": 81600
    },
    {
      "epoch": 8.788074480680228,
      "grad_norm": 0.9200389385223389,
      "learning_rate": 7.302014869087382e-05,
      "loss": 3.1933,
      "step": 81650
    },
    {
      "epoch": 8.793456032719837,
      "grad_norm": 0.861751139163971,
      "learning_rate": 7.26969076608124e-05,
      "loss": 3.1973,
      "step": 81700
    },
    {
      "epoch": 8.798837584759445,
      "grad_norm": 0.8833962678909302,
      "learning_rate": 7.237366663075099e-05,
      "loss": 3.1909,
      "step": 81750
    },
    {
      "epoch": 8.804219136799054,
      "grad_norm": 0.8531435132026672,
      "learning_rate": 7.205042560068958e-05,
      "loss": 3.1699,
      "step": 81800
    },
    {
      "epoch": 8.80960068883866,
      "grad_norm": 0.8738068342208862,
      "learning_rate": 7.172718457062816e-05,
      "loss": 3.2018,
      "step": 81850
    },
    {
      "epoch": 8.814982240878269,
      "grad_norm": 0.8231048583984375,
      "learning_rate": 7.140394354056674e-05,
      "loss": 3.1933,
      "step": 81900
    },
    {
      "epoch": 8.820363792917878,
      "grad_norm": 0.8450343012809753,
      "learning_rate": 7.108070251050532e-05,
      "loss": 3.1977,
      "step": 81950
    },
    {
      "epoch": 8.825745344957486,
      "grad_norm": 0.9066534042358398,
      "learning_rate": 7.07574614804439e-05,
      "loss": 3.1749,
      "step": 82000
    },
    {
      "epoch": 8.825745344957486,
      "eval_accuracy": 0.39188217333156156,
      "eval_loss": 3.3206756114959717,
      "eval_runtime": 184.4876,
      "eval_samples_per_second": 97.627,
      "eval_steps_per_second": 6.103,
      "step": 82000
    },
    {
      "epoch": 8.831126896997095,
      "grad_norm": 0.8430816531181335,
      "learning_rate": 7.04342204503825e-05,
      "loss": 3.2051,
      "step": 82050
    },
    {
      "epoch": 8.836508449036701,
      "grad_norm": 0.8838109970092773,
      "learning_rate": 7.011097942032108e-05,
      "loss": 3.1874,
      "step": 82100
    },
    {
      "epoch": 8.84189000107631,
      "grad_norm": 0.8406835794448853,
      "learning_rate": 6.978773839025966e-05,
      "loss": 3.1736,
      "step": 82150
    },
    {
      "epoch": 8.847271553115919,
      "grad_norm": 0.8512722253799438,
      "learning_rate": 6.946449736019824e-05,
      "loss": 3.1916,
      "step": 82200
    },
    {
      "epoch": 8.852653105155527,
      "grad_norm": 0.8770856261253357,
      "learning_rate": 6.914125633013683e-05,
      "loss": 3.186,
      "step": 82250
    },
    {
      "epoch": 8.858034657195136,
      "grad_norm": 0.8982064127922058,
      "learning_rate": 6.881801530007542e-05,
      "loss": 3.2041,
      "step": 82300
    },
    {
      "epoch": 8.863416209234742,
      "grad_norm": 0.9468360543251038,
      "learning_rate": 6.8494774270014e-05,
      "loss": 3.2093,
      "step": 82350
    },
    {
      "epoch": 8.868797761274351,
      "grad_norm": 0.8822013735771179,
      "learning_rate": 6.81715332399526e-05,
      "loss": 3.1974,
      "step": 82400
    },
    {
      "epoch": 8.87417931331396,
      "grad_norm": 0.8615790009498596,
      "learning_rate": 6.784829220989116e-05,
      "loss": 3.186,
      "step": 82450
    },
    {
      "epoch": 8.879560865353568,
      "grad_norm": 0.8560094833374023,
      "learning_rate": 6.752505117982974e-05,
      "loss": 3.1784,
      "step": 82500
    },
    {
      "epoch": 8.884942417393177,
      "grad_norm": 0.886281430721283,
      "learning_rate": 6.720181014976834e-05,
      "loss": 3.1754,
      "step": 82550
    },
    {
      "epoch": 8.890323969432785,
      "grad_norm": 0.8993915319442749,
      "learning_rate": 6.687856911970692e-05,
      "loss": 3.1816,
      "step": 82600
    },
    {
      "epoch": 8.895705521472392,
      "grad_norm": 0.8716897964477539,
      "learning_rate": 6.655532808964551e-05,
      "loss": 3.2091,
      "step": 82650
    },
    {
      "epoch": 8.901087073512,
      "grad_norm": 0.8357254266738892,
      "learning_rate": 6.62320870595841e-05,
      "loss": 3.1786,
      "step": 82700
    },
    {
      "epoch": 8.906468625551609,
      "grad_norm": 0.8596709966659546,
      "learning_rate": 6.590884602952268e-05,
      "loss": 3.2113,
      "step": 82750
    },
    {
      "epoch": 8.911850177591218,
      "grad_norm": 0.8767656087875366,
      "learning_rate": 6.558560499946126e-05,
      "loss": 3.1959,
      "step": 82800
    },
    {
      "epoch": 8.917231729630826,
      "grad_norm": 0.9359754323959351,
      "learning_rate": 6.526236396939984e-05,
      "loss": 3.1961,
      "step": 82850
    },
    {
      "epoch": 8.922613281670433,
      "grad_norm": 0.8610734939575195,
      "learning_rate": 6.493912293933843e-05,
      "loss": 3.181,
      "step": 82900
    },
    {
      "epoch": 8.927994833710041,
      "grad_norm": 0.8895869851112366,
      "learning_rate": 6.461588190927701e-05,
      "loss": 3.1939,
      "step": 82950
    },
    {
      "epoch": 8.93337638574965,
      "grad_norm": 0.874125599861145,
      "learning_rate": 6.42926408792156e-05,
      "loss": 3.1778,
      "step": 83000
    },
    {
      "epoch": 8.93337638574965,
      "eval_accuracy": 0.3924153330520672,
      "eval_loss": 3.316565752029419,
      "eval_runtime": 184.5368,
      "eval_samples_per_second": 97.601,
      "eval_steps_per_second": 6.102,
      "step": 83000
    },
    {
      "epoch": 8.938757937789259,
      "grad_norm": 0.8375032544136047,
      "learning_rate": 6.396939984915418e-05,
      "loss": 3.1796,
      "step": 83050
    },
    {
      "epoch": 8.944139489828867,
      "grad_norm": 0.9105697274208069,
      "learning_rate": 6.364615881909276e-05,
      "loss": 3.1757,
      "step": 83100
    },
    {
      "epoch": 8.949521041868476,
      "grad_norm": 0.8781415224075317,
      "learning_rate": 6.332291778903135e-05,
      "loss": 3.1948,
      "step": 83150
    },
    {
      "epoch": 8.954902593908082,
      "grad_norm": 0.8927822709083557,
      "learning_rate": 6.299967675896993e-05,
      "loss": 3.1918,
      "step": 83200
    },
    {
      "epoch": 8.960284145947691,
      "grad_norm": 0.8524188995361328,
      "learning_rate": 6.267643572890851e-05,
      "loss": 3.1988,
      "step": 83250
    },
    {
      "epoch": 8.9656656979873,
      "grad_norm": 0.8493085503578186,
      "learning_rate": 6.235319469884711e-05,
      "loss": 3.1824,
      "step": 83300
    },
    {
      "epoch": 8.971047250026908,
      "grad_norm": 0.8500775098800659,
      "learning_rate": 6.202995366878569e-05,
      "loss": 3.1862,
      "step": 83350
    },
    {
      "epoch": 8.976428802066517,
      "grad_norm": 0.849372148513794,
      "learning_rate": 6.17131774593255e-05,
      "loss": 3.1923,
      "step": 83400
    },
    {
      "epoch": 8.981810354106123,
      "grad_norm": 0.8546610474586487,
      "learning_rate": 6.138993642926409e-05,
      "loss": 3.1914,
      "step": 83450
    },
    {
      "epoch": 8.987191906145732,
      "grad_norm": 0.8478337526321411,
      "learning_rate": 6.106669539920267e-05,
      "loss": 3.1969,
      "step": 83500
    },
    {
      "epoch": 8.99257345818534,
      "grad_norm": 0.9085497856140137,
      "learning_rate": 6.074345436914125e-05,
      "loss": 3.1912,
      "step": 83550
    },
    {
      "epoch": 8.997955010224949,
      "grad_norm": 0.8430022597312927,
      "learning_rate": 6.0420213339079835e-05,
      "loss": 3.1908,
      "step": 83600
    },
    {
      "epoch": 9.003336562264558,
      "grad_norm": 0.850607693195343,
      "learning_rate": 6.0096972309018416e-05,
      "loss": 3.1463,
      "step": 83650
    },
    {
      "epoch": 9.008718114304166,
      "grad_norm": 0.8597699403762817,
      "learning_rate": 5.9773731278957004e-05,
      "loss": 3.1242,
      "step": 83700
    },
    {
      "epoch": 9.014099666343773,
      "grad_norm": 0.8384504318237305,
      "learning_rate": 5.9450490248895585e-05,
      "loss": 3.141,
      "step": 83750
    },
    {
      "epoch": 9.019481218383381,
      "grad_norm": 0.8542933464050293,
      "learning_rate": 5.9127249218834166e-05,
      "loss": 3.1227,
      "step": 83800
    },
    {
      "epoch": 9.02486277042299,
      "grad_norm": 0.8425549268722534,
      "learning_rate": 5.880400818877276e-05,
      "loss": 3.1308,
      "step": 83850
    },
    {
      "epoch": 9.030244322462599,
      "grad_norm": 0.8656365275382996,
      "learning_rate": 5.848076715871134e-05,
      "loss": 3.1285,
      "step": 83900
    },
    {
      "epoch": 9.035625874502207,
      "grad_norm": 0.8357840180397034,
      "learning_rate": 5.815752612864993e-05,
      "loss": 3.1229,
      "step": 83950
    },
    {
      "epoch": 9.041007426541814,
      "grad_norm": 0.9044317007064819,
      "learning_rate": 5.783428509858851e-05,
      "loss": 3.1312,
      "step": 84000
    },
    {
      "epoch": 9.041007426541814,
      "eval_accuracy": 0.3921031733053762,
      "eval_loss": 3.3195340633392334,
      "eval_runtime": 184.3311,
      "eval_samples_per_second": 97.71,
      "eval_steps_per_second": 6.109,
      "step": 84000
    },
    {
      "epoch": 9.046388978581422,
      "grad_norm": 0.8760092854499817,
      "learning_rate": 5.751104406852709e-05,
      "loss": 3.1438,
      "step": 84050
    },
    {
      "epoch": 9.051770530621031,
      "grad_norm": 0.9145100712776184,
      "learning_rate": 5.718780303846568e-05,
      "loss": 3.1259,
      "step": 84100
    },
    {
      "epoch": 9.05715208266064,
      "grad_norm": 0.8415228128433228,
      "learning_rate": 5.686456200840426e-05,
      "loss": 3.1187,
      "step": 84150
    },
    {
      "epoch": 9.062533634700248,
      "grad_norm": 0.812950611114502,
      "learning_rate": 5.654132097834285e-05,
      "loss": 3.1357,
      "step": 84200
    },
    {
      "epoch": 9.067915186739857,
      "grad_norm": 0.882986843585968,
      "learning_rate": 5.622454476888266e-05,
      "loss": 3.1312,
      "step": 84250
    },
    {
      "epoch": 9.073296738779463,
      "grad_norm": 0.8532711863517761,
      "learning_rate": 5.5901303738821244e-05,
      "loss": 3.1437,
      "step": 84300
    },
    {
      "epoch": 9.078678290819072,
      "grad_norm": 0.9017111659049988,
      "learning_rate": 5.5578062708759825e-05,
      "loss": 3.1232,
      "step": 84350
    },
    {
      "epoch": 9.08405984285868,
      "grad_norm": 0.8572760820388794,
      "learning_rate": 5.525482167869841e-05,
      "loss": 3.143,
      "step": 84400
    },
    {
      "epoch": 9.089441394898289,
      "grad_norm": 0.8974411487579346,
      "learning_rate": 5.4931580648636994e-05,
      "loss": 3.1204,
      "step": 84450
    },
    {
      "epoch": 9.094822946937898,
      "grad_norm": 0.8530295491218567,
      "learning_rate": 5.460833961857558e-05,
      "loss": 3.1277,
      "step": 84500
    },
    {
      "epoch": 9.100204498977504,
      "grad_norm": 0.8683794736862183,
      "learning_rate": 5.428509858851416e-05,
      "loss": 3.1435,
      "step": 84550
    },
    {
      "epoch": 9.105586051017113,
      "grad_norm": 0.8959080576896667,
      "learning_rate": 5.3961857558452744e-05,
      "loss": 3.1421,
      "step": 84600
    },
    {
      "epoch": 9.110967603056721,
      "grad_norm": 0.8570570945739746,
      "learning_rate": 5.363861652839133e-05,
      "loss": 3.1392,
      "step": 84650
    },
    {
      "epoch": 9.11634915509633,
      "grad_norm": 0.8528423309326172,
      "learning_rate": 5.331537549832991e-05,
      "loss": 3.1359,
      "step": 84700
    },
    {
      "epoch": 9.121730707135939,
      "grad_norm": 0.8884320259094238,
      "learning_rate": 5.299213446826851e-05,
      "loss": 3.1325,
      "step": 84750
    },
    {
      "epoch": 9.127112259175545,
      "grad_norm": 0.9350272417068481,
      "learning_rate": 5.266889343820709e-05,
      "loss": 3.1415,
      "step": 84800
    },
    {
      "epoch": 9.132493811215154,
      "grad_norm": 0.8802053928375244,
      "learning_rate": 5.234565240814567e-05,
      "loss": 3.1281,
      "step": 84850
    },
    {
      "epoch": 9.137875363254762,
      "grad_norm": 0.851226806640625,
      "learning_rate": 5.202241137808426e-05,
      "loss": 3.141,
      "step": 84900
    },
    {
      "epoch": 9.143256915294371,
      "grad_norm": 0.8808145523071289,
      "learning_rate": 5.169917034802284e-05,
      "loss": 3.1285,
      "step": 84950
    },
    {
      "epoch": 9.14863846733398,
      "grad_norm": 0.8982207775115967,
      "learning_rate": 5.1375929317961426e-05,
      "loss": 3.1279,
      "step": 85000
    },
    {
      "epoch": 9.14863846733398,
      "eval_accuracy": 0.39251409852709057,
      "eval_loss": 3.3192667961120605,
      "eval_runtime": 184.6616,
      "eval_samples_per_second": 97.535,
      "eval_steps_per_second": 6.098,
      "step": 85000
    },
    {
      "epoch": 9.154020019373588,
      "grad_norm": 0.839519202709198,
      "learning_rate": 5.105268828790001e-05,
      "loss": 3.136,
      "step": 85050
    },
    {
      "epoch": 9.159401571413195,
      "grad_norm": 0.9074612259864807,
      "learning_rate": 5.072944725783859e-05,
      "loss": 3.1313,
      "step": 85100
    },
    {
      "epoch": 9.164783123452803,
      "grad_norm": 0.8593998551368713,
      "learning_rate": 5.0406206227777177e-05,
      "loss": 3.1297,
      "step": 85150
    },
    {
      "epoch": 9.170164675492412,
      "grad_norm": 0.8691095113754272,
      "learning_rate": 5.008296519771576e-05,
      "loss": 3.1255,
      "step": 85200
    },
    {
      "epoch": 9.17554622753202,
      "grad_norm": 0.8539316058158875,
      "learning_rate": 4.9759724167654346e-05,
      "loss": 3.1327,
      "step": 85250
    },
    {
      "epoch": 9.180927779571629,
      "grad_norm": 0.887241005897522,
      "learning_rate": 4.943648313759293e-05,
      "loss": 3.1384,
      "step": 85300
    },
    {
      "epoch": 9.186309331611236,
      "grad_norm": 0.9014782309532166,
      "learning_rate": 4.911324210753151e-05,
      "loss": 3.1303,
      "step": 85350
    },
    {
      "epoch": 9.191690883650844,
      "grad_norm": 0.9132791757583618,
      "learning_rate": 4.8790001077470096e-05,
      "loss": 3.1451,
      "step": 85400
    },
    {
      "epoch": 9.197072435690453,
      "grad_norm": 0.8838841915130615,
      "learning_rate": 4.846676004740868e-05,
      "loss": 3.1275,
      "step": 85450
    },
    {
      "epoch": 9.202453987730062,
      "grad_norm": 0.8870031237602234,
      "learning_rate": 4.8143519017347265e-05,
      "loss": 3.1342,
      "step": 85500
    },
    {
      "epoch": 9.20783553976967,
      "grad_norm": 0.8817411661148071,
      "learning_rate": 4.7820277987285846e-05,
      "loss": 3.143,
      "step": 85550
    },
    {
      "epoch": 9.213217091809279,
      "grad_norm": 0.8416873812675476,
      "learning_rate": 4.749703695722443e-05,
      "loss": 3.1388,
      "step": 85600
    },
    {
      "epoch": 9.218598643848885,
      "grad_norm": 0.8405388593673706,
      "learning_rate": 4.717379592716302e-05,
      "loss": 3.135,
      "step": 85650
    },
    {
      "epoch": 9.223980195888494,
      "grad_norm": 0.888099193572998,
      "learning_rate": 4.68505548971016e-05,
      "loss": 3.135,
      "step": 85700
    },
    {
      "epoch": 9.229361747928102,
      "grad_norm": 0.9294061064720154,
      "learning_rate": 4.652731386704019e-05,
      "loss": 3.1517,
      "step": 85750
    },
    {
      "epoch": 9.234743299967711,
      "grad_norm": 0.8870118856430054,
      "learning_rate": 4.620407283697877e-05,
      "loss": 3.1369,
      "step": 85800
    },
    {
      "epoch": 9.24012485200732,
      "grad_norm": 0.8432061076164246,
      "learning_rate": 4.588083180691735e-05,
      "loss": 3.1341,
      "step": 85850
    },
    {
      "epoch": 9.245506404046926,
      "grad_norm": 0.8873381018638611,
      "learning_rate": 4.555759077685594e-05,
      "loss": 3.1428,
      "step": 85900
    },
    {
      "epoch": 9.250887956086535,
      "grad_norm": 0.8409569263458252,
      "learning_rate": 4.523434974679452e-05,
      "loss": 3.1444,
      "step": 85950
    },
    {
      "epoch": 9.256269508126143,
      "grad_norm": 0.8745248317718506,
      "learning_rate": 4.491110871673311e-05,
      "loss": 3.1351,
      "step": 86000
    },
    {
      "epoch": 9.256269508126143,
      "eval_accuracy": 0.3925824411935259,
      "eval_loss": 3.3168060779571533,
      "eval_runtime": 184.3857,
      "eval_samples_per_second": 97.681,
      "eval_steps_per_second": 6.107,
      "step": 86000
    },
    {
      "epoch": 9.261651060165752,
      "grad_norm": 0.8389315009117126,
      "learning_rate": 4.458786768667169e-05,
      "loss": 3.1205,
      "step": 86050
    },
    {
      "epoch": 9.26703261220536,
      "grad_norm": 0.9058299660682678,
      "learning_rate": 4.426462665661027e-05,
      "loss": 3.1442,
      "step": 86100
    },
    {
      "epoch": 9.272414164244967,
      "grad_norm": 0.8677758574485779,
      "learning_rate": 4.394138562654886e-05,
      "loss": 3.1457,
      "step": 86150
    },
    {
      "epoch": 9.277795716284576,
      "grad_norm": 0.8671353459358215,
      "learning_rate": 4.361814459648744e-05,
      "loss": 3.1574,
      "step": 86200
    },
    {
      "epoch": 9.283177268324184,
      "grad_norm": 0.8805967569351196,
      "learning_rate": 4.329490356642603e-05,
      "loss": 3.1301,
      "step": 86250
    },
    {
      "epoch": 9.288558820363793,
      "grad_norm": 0.8606386184692383,
      "learning_rate": 4.297166253636461e-05,
      "loss": 3.1358,
      "step": 86300
    },
    {
      "epoch": 9.293940372403402,
      "grad_norm": 0.9080101251602173,
      "learning_rate": 4.264842150630319e-05,
      "loss": 3.146,
      "step": 86350
    },
    {
      "epoch": 9.29932192444301,
      "grad_norm": 0.8556746244430542,
      "learning_rate": 4.2325180476241786e-05,
      "loss": 3.1332,
      "step": 86400
    },
    {
      "epoch": 9.304703476482617,
      "grad_norm": 0.8798325657844543,
      "learning_rate": 4.200193944618036e-05,
      "loss": 3.1362,
      "step": 86450
    },
    {
      "epoch": 9.310085028522225,
      "grad_norm": 0.9042881727218628,
      "learning_rate": 4.1678698416118955e-05,
      "loss": 3.1225,
      "step": 86500
    },
    {
      "epoch": 9.315466580561834,
      "grad_norm": 0.8741815090179443,
      "learning_rate": 4.1355457386057536e-05,
      "loss": 3.1384,
      "step": 86550
    },
    {
      "epoch": 9.320848132601443,
      "grad_norm": 0.8849769830703735,
      "learning_rate": 4.103221635599612e-05,
      "loss": 3.1365,
      "step": 86600
    },
    {
      "epoch": 9.326229684641051,
      "grad_norm": 0.8580855131149292,
      "learning_rate": 4.0708975325934705e-05,
      "loss": 3.1356,
      "step": 86650
    },
    {
      "epoch": 9.331611236680658,
      "grad_norm": 0.8469288349151611,
      "learning_rate": 4.0385734295873286e-05,
      "loss": 3.1351,
      "step": 86700
    },
    {
      "epoch": 9.336992788720266,
      "grad_norm": 0.8752751350402832,
      "learning_rate": 4.0062493265811874e-05,
      "loss": 3.1498,
      "step": 86750
    },
    {
      "epoch": 9.342374340759875,
      "grad_norm": 0.8474909663200378,
      "learning_rate": 3.9739252235750455e-05,
      "loss": 3.1362,
      "step": 86800
    },
    {
      "epoch": 9.347755892799483,
      "grad_norm": 0.8363193273544312,
      "learning_rate": 3.9416011205689036e-05,
      "loss": 3.1321,
      "step": 86850
    },
    {
      "epoch": 9.353137444839092,
      "grad_norm": 0.8611142039299011,
      "learning_rate": 3.9092770175627624e-05,
      "loss": 3.1403,
      "step": 86900
    },
    {
      "epoch": 9.3585189968787,
      "grad_norm": 0.8760079145431519,
      "learning_rate": 3.8769529145566205e-05,
      "loss": 3.1331,
      "step": 86950
    },
    {
      "epoch": 9.363900548918307,
      "grad_norm": 0.8837893605232239,
      "learning_rate": 3.844628811550479e-05,
      "loss": 3.1305,
      "step": 87000
    },
    {
      "epoch": 9.363900548918307,
      "eval_accuracy": 0.39302976513265814,
      "eval_loss": 3.3144354820251465,
      "eval_runtime": 184.4805,
      "eval_samples_per_second": 97.631,
      "eval_steps_per_second": 6.104,
      "step": 87000
    },
    {
      "epoch": 9.369282100957916,
      "grad_norm": 0.8648733496665955,
      "learning_rate": 3.8123047085443374e-05,
      "loss": 3.1364,
      "step": 87050
    },
    {
      "epoch": 9.374663652997524,
      "grad_norm": 0.8800923228263855,
      "learning_rate": 3.7799806055381955e-05,
      "loss": 3.1434,
      "step": 87100
    },
    {
      "epoch": 9.380045205037133,
      "grad_norm": 0.9065032005310059,
      "learning_rate": 3.747656502532054e-05,
      "loss": 3.1432,
      "step": 87150
    },
    {
      "epoch": 9.385426757076742,
      "grad_norm": 0.892939567565918,
      "learning_rate": 3.7153323995259124e-05,
      "loss": 3.1356,
      "step": 87200
    },
    {
      "epoch": 9.390808309116348,
      "grad_norm": 0.8660373091697693,
      "learning_rate": 3.683008296519771e-05,
      "loss": 3.1342,
      "step": 87250
    },
    {
      "epoch": 9.396189861155957,
      "grad_norm": 0.8589975237846375,
      "learning_rate": 3.65068419351363e-05,
      "loss": 3.1287,
      "step": 87300
    },
    {
      "epoch": 9.401571413195565,
      "grad_norm": 0.8542919754981995,
      "learning_rate": 3.618360090507488e-05,
      "loss": 3.1137,
      "step": 87350
    },
    {
      "epoch": 9.406952965235174,
      "grad_norm": 0.8634254932403564,
      "learning_rate": 3.586035987501347e-05,
      "loss": 3.1374,
      "step": 87400
    },
    {
      "epoch": 9.412334517274783,
      "grad_norm": 0.9162845015525818,
      "learning_rate": 3.553711884495205e-05,
      "loss": 3.146,
      "step": 87450
    },
    {
      "epoch": 9.417716069314391,
      "grad_norm": 0.866362452507019,
      "learning_rate": 3.521387781489063e-05,
      "loss": 3.1375,
      "step": 87500
    },
    {
      "epoch": 9.423097621353998,
      "grad_norm": 0.8890534043312073,
      "learning_rate": 3.489063678482922e-05,
      "loss": 3.1481,
      "step": 87550
    },
    {
      "epoch": 9.428479173393606,
      "grad_norm": 0.8560742735862732,
      "learning_rate": 3.45673957547678e-05,
      "loss": 3.1409,
      "step": 87600
    },
    {
      "epoch": 9.433860725433215,
      "grad_norm": 0.8744609951972961,
      "learning_rate": 3.424415472470639e-05,
      "loss": 3.1541,
      "step": 87650
    },
    {
      "epoch": 9.439242277472824,
      "grad_norm": 0.8752997517585754,
      "learning_rate": 3.392091369464497e-05,
      "loss": 3.1376,
      "step": 87700
    },
    {
      "epoch": 9.444623829512432,
      "grad_norm": 0.8732566237449646,
      "learning_rate": 3.359767266458356e-05,
      "loss": 3.1539,
      "step": 87750
    },
    {
      "epoch": 9.450005381552039,
      "grad_norm": 0.8782443404197693,
      "learning_rate": 3.327443163452214e-05,
      "loss": 3.134,
      "step": 87800
    },
    {
      "epoch": 9.455386933591647,
      "grad_norm": 0.8991506695747375,
      "learning_rate": 3.2951190604460726e-05,
      "loss": 3.1536,
      "step": 87850
    },
    {
      "epoch": 9.460768485631256,
      "grad_norm": 0.8943271040916443,
      "learning_rate": 3.262794957439931e-05,
      "loss": 3.1479,
      "step": 87900
    },
    {
      "epoch": 9.466150037670864,
      "grad_norm": 0.8531937599182129,
      "learning_rate": 3.230470854433789e-05,
      "loss": 3.1212,
      "step": 87950
    },
    {
      "epoch": 9.471531589710473,
      "grad_norm": 0.8615516424179077,
      "learning_rate": 3.1981467514276476e-05,
      "loss": 3.1363,
      "step": 88000
    },
    {
      "epoch": 9.471531589710473,
      "eval_accuracy": 0.3932671716925328,
      "eval_loss": 3.3122997283935547,
      "eval_runtime": 184.5104,
      "eval_samples_per_second": 97.615,
      "eval_steps_per_second": 6.103,
      "step": 88000
    },
    {
      "epoch": 9.476913141750082,
      "grad_norm": 0.9828490614891052,
      "learning_rate": 3.165822648421506e-05,
      "loss": 3.1508,
      "step": 88050
    },
    {
      "epoch": 9.482294693789688,
      "grad_norm": 0.8655065894126892,
      "learning_rate": 3.1334985454153645e-05,
      "loss": 3.1506,
      "step": 88100
    },
    {
      "epoch": 9.487676245829297,
      "grad_norm": 0.8861255049705505,
      "learning_rate": 3.101174442409223e-05,
      "loss": 3.1412,
      "step": 88150
    },
    {
      "epoch": 9.493057797868905,
      "grad_norm": 0.8610599637031555,
      "learning_rate": 3.0688503394030814e-05,
      "loss": 3.1456,
      "step": 88200
    },
    {
      "epoch": 9.498439349908514,
      "grad_norm": 0.9001054167747498,
      "learning_rate": 3.0371727184570624e-05,
      "loss": 3.127,
      "step": 88250
    },
    {
      "epoch": 9.503820901948123,
      "grad_norm": 0.8203874230384827,
      "learning_rate": 3.0048486154509208e-05,
      "loss": 3.1277,
      "step": 88300
    },
    {
      "epoch": 9.50920245398773,
      "grad_norm": 0.8560073971748352,
      "learning_rate": 2.9725245124447793e-05,
      "loss": 3.1447,
      "step": 88350
    },
    {
      "epoch": 9.514584006027338,
      "grad_norm": 0.8469882011413574,
      "learning_rate": 2.940200409438638e-05,
      "loss": 3.1499,
      "step": 88400
    },
    {
      "epoch": 9.519965558066946,
      "grad_norm": 0.904205858707428,
      "learning_rate": 2.9078763064324965e-05,
      "loss": 3.1326,
      "step": 88450
    },
    {
      "epoch": 9.525347110106555,
      "grad_norm": 0.8689672946929932,
      "learning_rate": 2.8755522034263546e-05,
      "loss": 3.1371,
      "step": 88500
    },
    {
      "epoch": 9.530728662146164,
      "grad_norm": 0.8763381838798523,
      "learning_rate": 2.843228100420213e-05,
      "loss": 3.1343,
      "step": 88550
    },
    {
      "epoch": 9.536110214185772,
      "grad_norm": 0.822755753993988,
      "learning_rate": 2.8109039974140715e-05,
      "loss": 3.1309,
      "step": 88600
    },
    {
      "epoch": 9.541491766225379,
      "grad_norm": 0.8990127444267273,
      "learning_rate": 2.77857989440793e-05,
      "loss": 3.1487,
      "step": 88650
    },
    {
      "epoch": 9.546873318264987,
      "grad_norm": 0.8536121249198914,
      "learning_rate": 2.7462557914017884e-05,
      "loss": 3.1484,
      "step": 88700
    },
    {
      "epoch": 9.552254870304596,
      "grad_norm": 0.851606011390686,
      "learning_rate": 2.7139316883956465e-05,
      "loss": 3.1491,
      "step": 88750
    },
    {
      "epoch": 9.557636422344205,
      "grad_norm": 0.8838151097297668,
      "learning_rate": 2.681607585389505e-05,
      "loss": 3.1446,
      "step": 88800
    },
    {
      "epoch": 9.563017974383813,
      "grad_norm": 0.9009897112846375,
      "learning_rate": 2.6492834823833638e-05,
      "loss": 3.1429,
      "step": 88850
    },
    {
      "epoch": 9.56839952642342,
      "grad_norm": 0.8954452276229858,
      "learning_rate": 2.6169593793772222e-05,
      "loss": 3.1443,
      "step": 88900
    },
    {
      "epoch": 9.573781078463028,
      "grad_norm": 0.9105966687202454,
      "learning_rate": 2.5846352763710807e-05,
      "loss": 3.1486,
      "step": 88950
    },
    {
      "epoch": 9.579162630502637,
      "grad_norm": 0.9291102290153503,
      "learning_rate": 2.5523111733649388e-05,
      "loss": 3.1362,
      "step": 89000
    },
    {
      "epoch": 9.579162630502637,
      "eval_accuracy": 0.3935795487449994,
      "eval_loss": 3.310000419616699,
      "eval_runtime": 184.5847,
      "eval_samples_per_second": 97.576,
      "eval_steps_per_second": 6.1,
      "step": 89000
    },
    {
      "epoch": 9.584544182542245,
      "grad_norm": 0.8935585618019104,
      "learning_rate": 2.5199870703587972e-05,
      "loss": 3.1386,
      "step": 89050
    },
    {
      "epoch": 9.589925734581854,
      "grad_norm": 0.9456694722175598,
      "learning_rate": 2.4876629673526557e-05,
      "loss": 3.1347,
      "step": 89100
    },
    {
      "epoch": 9.59530728662146,
      "grad_norm": 0.8893503546714783,
      "learning_rate": 2.455338864346514e-05,
      "loss": 3.1465,
      "step": 89150
    },
    {
      "epoch": 9.60068883866107,
      "grad_norm": 0.851827085018158,
      "learning_rate": 2.423014761340373e-05,
      "loss": 3.1396,
      "step": 89200
    },
    {
      "epoch": 9.606070390700678,
      "grad_norm": 0.8665865659713745,
      "learning_rate": 2.3906906583342307e-05,
      "loss": 3.1369,
      "step": 89250
    },
    {
      "epoch": 9.611451942740286,
      "grad_norm": 0.8616881966590881,
      "learning_rate": 2.3583665553280895e-05,
      "loss": 3.1535,
      "step": 89300
    },
    {
      "epoch": 9.616833494779895,
      "grad_norm": 0.8565948009490967,
      "learning_rate": 2.326042452321948e-05,
      "loss": 3.1368,
      "step": 89350
    },
    {
      "epoch": 9.622215046819504,
      "grad_norm": 0.8922091722488403,
      "learning_rate": 2.2937183493158064e-05,
      "loss": 3.1374,
      "step": 89400
    },
    {
      "epoch": 9.62759659885911,
      "grad_norm": 0.8471571803092957,
      "learning_rate": 2.2613942463096648e-05,
      "loss": 3.135,
      "step": 89450
    },
    {
      "epoch": 9.632978150898719,
      "grad_norm": 0.8804218173027039,
      "learning_rate": 2.229070143303523e-05,
      "loss": 3.133,
      "step": 89500
    },
    {
      "epoch": 9.638359702938327,
      "grad_norm": 0.9225805401802063,
      "learning_rate": 2.1967460402973814e-05,
      "loss": 3.1555,
      "step": 89550
    },
    {
      "epoch": 9.643741254977936,
      "grad_norm": 0.8777856826782227,
      "learning_rate": 2.1644219372912398e-05,
      "loss": 3.1359,
      "step": 89600
    },
    {
      "epoch": 9.649122807017545,
      "grad_norm": 0.9119404554367065,
      "learning_rate": 2.1320978342850986e-05,
      "loss": 3.1221,
      "step": 89650
    },
    {
      "epoch": 9.654504359057151,
      "grad_norm": 0.8720681667327881,
      "learning_rate": 2.099773731278957e-05,
      "loss": 3.1498,
      "step": 89700
    },
    {
      "epoch": 9.65988591109676,
      "grad_norm": 0.8849010467529297,
      "learning_rate": 2.0674496282728152e-05,
      "loss": 3.1424,
      "step": 89750
    },
    {
      "epoch": 9.665267463136368,
      "grad_norm": 0.8780276775360107,
      "learning_rate": 2.0351255252666736e-05,
      "loss": 3.1519,
      "step": 89800
    },
    {
      "epoch": 9.670649015175977,
      "grad_norm": 0.9049113988876343,
      "learning_rate": 2.003447904320655e-05,
      "loss": 3.1525,
      "step": 89850
    },
    {
      "epoch": 9.676030567215586,
      "grad_norm": 0.8769359588623047,
      "learning_rate": 1.9711238013145134e-05,
      "loss": 3.1469,
      "step": 89900
    },
    {
      "epoch": 9.681412119255192,
      "grad_norm": 0.8801373839378357,
      "learning_rate": 1.938799698308372e-05,
      "loss": 3.1382,
      "step": 89950
    },
    {
      "epoch": 9.6867936712948,
      "grad_norm": 0.917766809463501,
      "learning_rate": 1.9064755953022303e-05,
      "loss": 3.1398,
      "step": 90000
    },
    {
      "epoch": 9.6867936712948,
      "eval_accuracy": 0.393675815203603,
      "eval_loss": 3.3074893951416016,
      "eval_runtime": 184.642,
      "eval_samples_per_second": 97.546,
      "eval_steps_per_second": 6.098,
      "step": 90000
    },
    {
      "epoch": 9.69217522333441,
      "grad_norm": 0.8497322797775269,
      "learning_rate": 1.8741514922960887e-05,
      "loss": 3.1579,
      "step": 90050
    },
    {
      "epoch": 9.697556775374018,
      "grad_norm": 0.8897688388824463,
      "learning_rate": 1.841827389289947e-05,
      "loss": 3.1277,
      "step": 90100
    },
    {
      "epoch": 9.702938327413626,
      "grad_norm": 0.8786342740058899,
      "learning_rate": 1.8095032862838053e-05,
      "loss": 3.1311,
      "step": 90150
    },
    {
      "epoch": 9.708319879453235,
      "grad_norm": 0.89149010181427,
      "learning_rate": 1.777179183277664e-05,
      "loss": 3.1373,
      "step": 90200
    },
    {
      "epoch": 9.713701431492842,
      "grad_norm": 0.8806278705596924,
      "learning_rate": 1.7448550802715222e-05,
      "loss": 3.1286,
      "step": 90250
    },
    {
      "epoch": 9.71908298353245,
      "grad_norm": 0.8583605289459229,
      "learning_rate": 1.7125309772653807e-05,
      "loss": 3.1318,
      "step": 90300
    },
    {
      "epoch": 9.724464535572059,
      "grad_norm": 0.8373787999153137,
      "learning_rate": 1.680206874259239e-05,
      "loss": 3.1226,
      "step": 90350
    },
    {
      "epoch": 9.729846087611667,
      "grad_norm": 0.8798319697380066,
      "learning_rate": 1.6478827712530976e-05,
      "loss": 3.1321,
      "step": 90400
    },
    {
      "epoch": 9.735227639651276,
      "grad_norm": 0.9543043375015259,
      "learning_rate": 1.615558668246956e-05,
      "loss": 3.1496,
      "step": 90450
    },
    {
      "epoch": 9.740609191690883,
      "grad_norm": 0.8514014482498169,
      "learning_rate": 1.5832345652408145e-05,
      "loss": 3.1425,
      "step": 90500
    },
    {
      "epoch": 9.745990743730491,
      "grad_norm": 0.8588919639587402,
      "learning_rate": 1.550910462234673e-05,
      "loss": 3.1499,
      "step": 90550
    },
    {
      "epoch": 9.7513722957701,
      "grad_norm": 0.9020848274230957,
      "learning_rate": 1.5185863592285312e-05,
      "loss": 3.155,
      "step": 90600
    },
    {
      "epoch": 9.756753847809708,
      "grad_norm": 0.9013638496398926,
      "learning_rate": 1.4862622562223896e-05,
      "loss": 3.1493,
      "step": 90650
    },
    {
      "epoch": 9.762135399849317,
      "grad_norm": 0.8628064393997192,
      "learning_rate": 1.4539381532162483e-05,
      "loss": 3.1342,
      "step": 90700
    },
    {
      "epoch": 9.767516951888926,
      "grad_norm": 0.828583300113678,
      "learning_rate": 1.4216140502101065e-05,
      "loss": 3.1164,
      "step": 90750
    },
    {
      "epoch": 9.772898503928532,
      "grad_norm": 0.871366024017334,
      "learning_rate": 1.389289947203965e-05,
      "loss": 3.1313,
      "step": 90800
    },
    {
      "epoch": 9.77828005596814,
      "grad_norm": 0.9028555154800415,
      "learning_rate": 1.3569658441978233e-05,
      "loss": 3.1167,
      "step": 90850
    },
    {
      "epoch": 9.78366160800775,
      "grad_norm": 0.8447543978691101,
      "learning_rate": 1.3246417411916819e-05,
      "loss": 3.1433,
      "step": 90900
    },
    {
      "epoch": 9.789043160047358,
      "grad_norm": 0.8645057678222656,
      "learning_rate": 1.2923176381855403e-05,
      "loss": 3.1471,
      "step": 90950
    },
    {
      "epoch": 9.794424712086967,
      "grad_norm": 0.8772447109222412,
      "learning_rate": 1.2599935351793986e-05,
      "loss": 3.1428,
      "step": 91000
    },
    {
      "epoch": 9.794424712086967,
      "eval_accuracy": 0.39401731123000394,
      "eval_loss": 3.305429697036743,
      "eval_runtime": 184.9509,
      "eval_samples_per_second": 97.383,
      "eval_steps_per_second": 6.088,
      "step": 91000
    },
    {
      "epoch": 9.799806264126573,
      "grad_norm": 0.8695673942565918,
      "learning_rate": 1.227669432173257e-05,
      "loss": 3.1474,
      "step": 91050
    },
    {
      "epoch": 9.805187816166182,
      "grad_norm": 0.9071619510650635,
      "learning_rate": 1.1953453291671153e-05,
      "loss": 3.1233,
      "step": 91100
    },
    {
      "epoch": 9.81056936820579,
      "grad_norm": 0.9191455245018005,
      "learning_rate": 1.163021226160974e-05,
      "loss": 3.1366,
      "step": 91150
    },
    {
      "epoch": 9.815950920245399,
      "grad_norm": 0.8642084002494812,
      "learning_rate": 1.1306971231548324e-05,
      "loss": 3.1358,
      "step": 91200
    },
    {
      "epoch": 9.821332472285007,
      "grad_norm": 0.869050145149231,
      "learning_rate": 1.0983730201486907e-05,
      "loss": 3.158,
      "step": 91250
    },
    {
      "epoch": 9.826714024324616,
      "grad_norm": 0.8625747561454773,
      "learning_rate": 1.0660489171425493e-05,
      "loss": 3.1432,
      "step": 91300
    },
    {
      "epoch": 9.832095576364223,
      "grad_norm": 0.9114882349967957,
      "learning_rate": 1.0337248141364076e-05,
      "loss": 3.1511,
      "step": 91350
    },
    {
      "epoch": 9.837477128403831,
      "grad_norm": 0.9020192623138428,
      "learning_rate": 1.001400711130266e-05,
      "loss": 3.1411,
      "step": 91400
    },
    {
      "epoch": 9.84285868044344,
      "grad_norm": 0.9011106491088867,
      "learning_rate": 9.690766081241245e-06,
      "loss": 3.1237,
      "step": 91450
    },
    {
      "epoch": 9.848240232483048,
      "grad_norm": 0.8959977030754089,
      "learning_rate": 9.367525051179828e-06,
      "loss": 3.1565,
      "step": 91500
    },
    {
      "epoch": 9.853621784522657,
      "grad_norm": 0.8750842213630676,
      "learning_rate": 9.044284021118414e-06,
      "loss": 3.13,
      "step": 91550
    },
    {
      "epoch": 9.859003336562264,
      "grad_norm": 0.8629763722419739,
      "learning_rate": 8.721042991056998e-06,
      "loss": 3.1534,
      "step": 91600
    },
    {
      "epoch": 9.864384888601872,
      "grad_norm": 0.91971355676651,
      "learning_rate": 8.397801960995581e-06,
      "loss": 3.1351,
      "step": 91650
    },
    {
      "epoch": 9.869766440641481,
      "grad_norm": 0.8595741987228394,
      "learning_rate": 8.074560930934166e-06,
      "loss": 3.1302,
      "step": 91700
    },
    {
      "epoch": 9.87514799268109,
      "grad_norm": 0.8426492214202881,
      "learning_rate": 7.75131990087275e-06,
      "loss": 3.1462,
      "step": 91750
    },
    {
      "epoch": 9.880529544720698,
      "grad_norm": 0.8684423565864563,
      "learning_rate": 7.428078870811335e-06,
      "loss": 3.1374,
      "step": 91800
    },
    {
      "epoch": 9.885911096760307,
      "grad_norm": 0.8544895648956299,
      "learning_rate": 7.104837840749918e-06,
      "loss": 3.1338,
      "step": 91850
    },
    {
      "epoch": 9.891292648799913,
      "grad_norm": 0.8213328719139099,
      "learning_rate": 6.7880616312897315e-06,
      "loss": 3.1278,
      "step": 91900
    },
    {
      "epoch": 9.896674200839522,
      "grad_norm": 0.8747987151145935,
      "learning_rate": 6.464820601228315e-06,
      "loss": 3.1541,
      "step": 91950
    },
    {
      "epoch": 9.90205575287913,
      "grad_norm": 0.8488390445709229,
      "learning_rate": 6.1415795711669e-06,
      "loss": 3.1449,
      "step": 92000
    },
    {
      "epoch": 9.90205575287913,
      "eval_accuracy": 0.3941807251732769,
      "eval_loss": 3.3039169311523438,
      "eval_runtime": 184.1755,
      "eval_samples_per_second": 97.793,
      "eval_steps_per_second": 6.114,
      "step": 92000
    },
    {
      "epoch": 9.907437304918739,
      "grad_norm": 0.8306775093078613,
      "learning_rate": 5.818338541105483e-06,
      "loss": 3.1335,
      "step": 92050
    },
    {
      "epoch": 9.912818856958348,
      "grad_norm": 0.8709449172019958,
      "learning_rate": 5.495097511044069e-06,
      "loss": 3.1341,
      "step": 92100
    },
    {
      "epoch": 9.918200408997954,
      "grad_norm": 0.8920506834983826,
      "learning_rate": 5.171856480982652e-06,
      "loss": 3.142,
      "step": 92150
    },
    {
      "epoch": 9.923581961037563,
      "grad_norm": 0.8973367810249329,
      "learning_rate": 4.848615450921237e-06,
      "loss": 3.126,
      "step": 92200
    },
    {
      "epoch": 9.928963513077171,
      "grad_norm": 0.8748192191123962,
      "learning_rate": 4.5253744208598205e-06,
      "loss": 3.1379,
      "step": 92250
    },
    {
      "epoch": 9.93434506511678,
      "grad_norm": 0.8766202330589294,
      "learning_rate": 4.202133390798405e-06,
      "loss": 3.1372,
      "step": 92300
    },
    {
      "epoch": 9.939726617156388,
      "grad_norm": 0.9124631881713867,
      "learning_rate": 3.8788923607369895e-06,
      "loss": 3.1426,
      "step": 92350
    },
    {
      "epoch": 9.945108169195997,
      "grad_norm": 0.8823198080062866,
      "learning_rate": 3.555651330675573e-06,
      "loss": 3.1306,
      "step": 92400
    },
    {
      "epoch": 9.950489721235604,
      "grad_norm": 0.8331865072250366,
      "learning_rate": 3.2324103006141576e-06,
      "loss": 3.1497,
      "step": 92450
    },
    {
      "epoch": 9.955871273275212,
      "grad_norm": 0.8679667115211487,
      "learning_rate": 2.9091692705527417e-06,
      "loss": 3.1467,
      "step": 92500
    },
    {
      "epoch": 9.961252825314821,
      "grad_norm": 0.8798403739929199,
      "learning_rate": 2.585928240491326e-06,
      "loss": 3.1475,
      "step": 92550
    },
    {
      "epoch": 9.96663437735443,
      "grad_norm": 0.8715695142745972,
      "learning_rate": 2.2626872104299102e-06,
      "loss": 3.1384,
      "step": 92600
    },
    {
      "epoch": 9.972015929394038,
      "grad_norm": 0.9157437682151794,
      "learning_rate": 1.9394461803684947e-06,
      "loss": 3.15,
      "step": 92650
    },
    {
      "epoch": 9.977397481433645,
      "grad_norm": 0.9365684390068054,
      "learning_rate": 1.6162051503070788e-06,
      "loss": 3.1393,
      "step": 92700
    },
    {
      "epoch": 9.982779033473253,
      "grad_norm": 0.8449177742004395,
      "learning_rate": 1.292964120245663e-06,
      "loss": 3.1475,
      "step": 92750
    },
    {
      "epoch": 9.988160585512862,
      "grad_norm": 0.874988853931427,
      "learning_rate": 9.697230901842474e-07,
      "loss": 3.1427,
      "step": 92800
    },
    {
      "epoch": 9.99354213755247,
      "grad_norm": 0.8465766906738281,
      "learning_rate": 6.464820601228315e-07,
      "loss": 3.1474,
      "step": 92850
    },
    {
      "epoch": 9.998923689592079,
      "grad_norm": 0.8720293045043945,
      "learning_rate": 3.2324103006141577e-07,
      "loss": 3.134,
      "step": 92900
    },
    {
      "epoch": 10.0,
      "step": 92910,
      "total_flos": 7.7681859821568e+17,
      "train_loss": 3.462871078238294,
      "train_runtime": 80757.2697,
      "train_samples_per_second": 36.814,
      "train_steps_per_second": 1.15
    }
  ],
  "logging_steps": 50,
  "max_steps": 92910,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 10000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.7681859821568e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}