{
  "best_metric": 3.3056087493896484,
  "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_495/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": 1.0683914422988892,
      "learning_rate": 0.0003,
      "loss": 8.6798,
      "step": 50
    },
    {
      "epoch": 0.010763104079216447,
      "grad_norm": 2.859095573425293,
      "learning_rate": 0.0006,
      "loss": 6.9757,
      "step": 100
    },
    {
      "epoch": 0.01614465611882467,
      "grad_norm": 1.489780068397522,
      "learning_rate": 0.0005996767589699385,
      "loss": 6.479,
      "step": 150
    },
    {
      "epoch": 0.021526208158432893,
      "grad_norm": 1.3649728298187256,
      "learning_rate": 0.0005993535179398771,
      "loss": 6.1969,
      "step": 200
    },
    {
      "epoch": 0.026907760198041114,
      "grad_norm": 1.613872766494751,
      "learning_rate": 0.0005990302769098158,
      "loss": 6.0747,
      "step": 250
    },
    {
      "epoch": 0.03228931223764934,
      "grad_norm": 3.6575379371643066,
      "learning_rate": 0.0005987070358797543,
      "loss": 5.9726,
      "step": 300
    },
    {
      "epoch": 0.03767086427725756,
      "grad_norm": 1.7698675394058228,
      "learning_rate": 0.0005983837948496929,
      "loss": 5.8947,
      "step": 350
    },
    {
      "epoch": 0.04305241631686579,
      "grad_norm": 1.8536632061004639,
      "learning_rate": 0.0005980605538196314,
      "loss": 5.7945,
      "step": 400
    },
    {
      "epoch": 0.048433968356474004,
      "grad_norm": 1.3679734468460083,
      "learning_rate": 0.0005977373127895701,
      "loss": 5.721,
      "step": 450
    },
    {
      "epoch": 0.05381552039608223,
      "grad_norm": 1.759971022605896,
      "learning_rate": 0.0005974140717595086,
      "loss": 5.6392,
      "step": 500
    },
    {
      "epoch": 0.05919707243569045,
      "grad_norm": 1.350098729133606,
      "learning_rate": 0.0005970908307294472,
      "loss": 5.5834,
      "step": 550
    },
    {
      "epoch": 0.06457862447529868,
      "grad_norm": 1.1349716186523438,
      "learning_rate": 0.0005967675896993858,
      "loss": 5.5375,
      "step": 600
    },
    {
      "epoch": 0.0699601765149069,
      "grad_norm": 1.276620626449585,
      "learning_rate": 0.0005964443486693243,
      "loss": 5.4386,
      "step": 650
    },
    {
      "epoch": 0.07534172855451512,
      "grad_norm": 0.9349568486213684,
      "learning_rate": 0.000596121107639263,
      "loss": 5.389,
      "step": 700
    },
    {
      "epoch": 0.08072328059412334,
      "grad_norm": 1.2208337783813477,
      "learning_rate": 0.0005957978666092015,
      "loss": 5.347,
      "step": 750
    },
    {
      "epoch": 0.08610483263373157,
      "grad_norm": 0.9988003969192505,
      "learning_rate": 0.0005954746255791401,
      "loss": 5.2746,
      "step": 800
    },
    {
      "epoch": 0.09148638467333979,
      "grad_norm": 1.0252668857574463,
      "learning_rate": 0.0005951513845490787,
      "loss": 5.2421,
      "step": 850
    },
    {
      "epoch": 0.09686793671294801,
      "grad_norm": 1.0475060939788818,
      "learning_rate": 0.0005948281435190174,
      "loss": 5.1758,
      "step": 900
    },
    {
      "epoch": 0.10224948875255624,
      "grad_norm": 1.1841458082199097,
      "learning_rate": 0.0005945049024889559,
      "loss": 5.1527,
      "step": 950
    },
    {
      "epoch": 0.10763104079216446,
      "grad_norm": 1.3012635707855225,
      "learning_rate": 0.0005941816614588944,
      "loss": 5.1178,
      "step": 1000
    },
    {
      "epoch": 0.10763104079216446,
      "eval_accuracy": 0.2259316197396655,
      "eval_loss": 5.038772106170654,
      "eval_runtime": 185.0,
      "eval_samples_per_second": 97.357,
      "eval_steps_per_second": 6.086,
      "step": 1000
    },
    {
      "epoch": 0.11301259283177269,
      "grad_norm": 1.3463271856307983,
      "learning_rate": 0.000593858420428833,
      "loss": 5.0694,
      "step": 1050
    },
    {
      "epoch": 0.1183941448713809,
      "grad_norm": 1.3660123348236084,
      "learning_rate": 0.0005935351793987716,
      "loss": 5.0413,
      "step": 1100
    },
    {
      "epoch": 0.12377569691098914,
      "grad_norm": 1.0985352993011475,
      "learning_rate": 0.0005932119383687103,
      "loss": 5.0437,
      "step": 1150
    },
    {
      "epoch": 0.12915724895059735,
      "grad_norm": 1.175125241279602,
      "learning_rate": 0.0005928886973386488,
      "loss": 4.9567,
      "step": 1200
    },
    {
      "epoch": 0.13453880099020557,
      "grad_norm": 1.1065387725830078,
      "learning_rate": 0.0005925654563085874,
      "loss": 4.9404,
      "step": 1250
    },
    {
      "epoch": 0.1399203530298138,
      "grad_norm": 2.134150505065918,
      "learning_rate": 0.000592242215278526,
      "loss": 4.8978,
      "step": 1300
    },
    {
      "epoch": 0.14530190506942203,
      "grad_norm": 0.826282262802124,
      "learning_rate": 0.0005919189742484645,
      "loss": 4.8955,
      "step": 1350
    },
    {
      "epoch": 0.15068345710903025,
      "grad_norm": 1.1777522563934326,
      "learning_rate": 0.0005915957332184032,
      "loss": 4.8565,
      "step": 1400
    },
    {
      "epoch": 0.15606500914863847,
      "grad_norm": 1.0407218933105469,
      "learning_rate": 0.0005912724921883417,
      "loss": 4.8394,
      "step": 1450
    },
    {
      "epoch": 0.16144656118824668,
      "grad_norm": 1.1135185956954956,
      "learning_rate": 0.0005909492511582803,
      "loss": 4.8083,
      "step": 1500
    },
    {
      "epoch": 0.1668281132278549,
      "grad_norm": 0.9728663563728333,
      "learning_rate": 0.0005906260101282189,
      "loss": 4.7959,
      "step": 1550
    },
    {
      "epoch": 0.17220966526746315,
      "grad_norm": 0.8725541234016418,
      "learning_rate": 0.0005903027690981575,
      "loss": 4.7704,
      "step": 1600
    },
    {
      "epoch": 0.17759121730707136,
      "grad_norm": 0.8491561412811279,
      "learning_rate": 0.000589979528068096,
      "loss": 4.7693,
      "step": 1650
    },
    {
      "epoch": 0.18297276934667958,
      "grad_norm": 1.0023645162582397,
      "learning_rate": 0.0005896562870380347,
      "loss": 4.7243,
      "step": 1700
    },
    {
      "epoch": 0.1883543213862878,
      "grad_norm": 1.2298816442489624,
      "learning_rate": 0.0005893330460079732,
      "loss": 4.7076,
      "step": 1750
    },
    {
      "epoch": 0.19373587342589602,
      "grad_norm": 1.063509225845337,
      "learning_rate": 0.0005890098049779118,
      "loss": 4.6721,
      "step": 1800
    },
    {
      "epoch": 0.19911742546550426,
      "grad_norm": 1.0815796852111816,
      "learning_rate": 0.0005886865639478504,
      "loss": 4.6689,
      "step": 1850
    },
    {
      "epoch": 0.20449897750511248,
      "grad_norm": 1.0041590929031372,
      "learning_rate": 0.0005883633229177889,
      "loss": 4.664,
      "step": 1900
    },
    {
      "epoch": 0.2098805295447207,
      "grad_norm": 0.741370439529419,
      "learning_rate": 0.0005880400818877276,
      "loss": 4.647,
      "step": 1950
    },
    {
      "epoch": 0.2152620815843289,
      "grad_norm": 0.9399243593215942,
      "learning_rate": 0.0005877168408576662,
      "loss": 4.5945,
      "step": 2000
    },
    {
      "epoch": 0.2152620815843289,
      "eval_accuracy": 0.2670903115197812,
      "eval_loss": 4.538747310638428,
      "eval_runtime": 182.7156,
      "eval_samples_per_second": 98.574,
      "eval_steps_per_second": 6.163,
      "step": 2000
    },
    {
      "epoch": 0.22064363362393713,
      "grad_norm": 0.8416916131973267,
      "learning_rate": 0.0005873935998276048,
      "loss": 4.5951,
      "step": 2050
    },
    {
      "epoch": 0.22602518566354537,
      "grad_norm": 1.0621873140335083,
      "learning_rate": 0.0005870703587975433,
      "loss": 4.5883,
      "step": 2100
    },
    {
      "epoch": 0.2314067377031536,
      "grad_norm": 0.958995521068573,
      "learning_rate": 0.0005867471177674818,
      "loss": 4.564,
      "step": 2150
    },
    {
      "epoch": 0.2367882897427618,
      "grad_norm": 1.0263909101486206,
      "learning_rate": 0.0005864238767374205,
      "loss": 4.5361,
      "step": 2200
    },
    {
      "epoch": 0.24216984178237003,
      "grad_norm": 0.865015983581543,
      "learning_rate": 0.0005861006357073591,
      "loss": 4.5177,
      "step": 2250
    },
    {
      "epoch": 0.24755139382197827,
      "grad_norm": 0.8319127559661865,
      "learning_rate": 0.0005857773946772977,
      "loss": 4.5171,
      "step": 2300
    },
    {
      "epoch": 0.2529329458615865,
      "grad_norm": 0.8533033132553101,
      "learning_rate": 0.0005854541536472362,
      "loss": 4.4716,
      "step": 2350
    },
    {
      "epoch": 0.2583144979011947,
      "grad_norm": 0.9105812907218933,
      "learning_rate": 0.0005851309126171749,
      "loss": 4.4832,
      "step": 2400
    },
    {
      "epoch": 0.2636960499408029,
      "grad_norm": 0.9672813415527344,
      "learning_rate": 0.0005848076715871134,
      "loss": 4.4371,
      "step": 2450
    },
    {
      "epoch": 0.26907760198041114,
      "grad_norm": 0.9679467082023621,
      "learning_rate": 0.000584484430557052,
      "loss": 4.4541,
      "step": 2500
    },
    {
      "epoch": 0.27445915402001936,
      "grad_norm": 0.8314814567565918,
      "learning_rate": 0.0005841611895269906,
      "loss": 4.4395,
      "step": 2550
    },
    {
      "epoch": 0.2798407060596276,
      "grad_norm": 0.7759853601455688,
      "learning_rate": 0.0005838379484969291,
      "loss": 4.4153,
      "step": 2600
    },
    {
      "epoch": 0.2852222580992358,
      "grad_norm": 0.7185021638870239,
      "learning_rate": 0.0005835147074668678,
      "loss": 4.3998,
      "step": 2650
    },
    {
      "epoch": 0.29060381013884407,
      "grad_norm": 0.7233301997184753,
      "learning_rate": 0.0005831914664368063,
      "loss": 4.3864,
      "step": 2700
    },
    {
      "epoch": 0.2959853621784523,
      "grad_norm": 0.768659234046936,
      "learning_rate": 0.0005828682254067449,
      "loss": 4.3805,
      "step": 2750
    },
    {
      "epoch": 0.3013669142180605,
      "grad_norm": 0.7125174403190613,
      "learning_rate": 0.0005825449843766835,
      "loss": 4.3937,
      "step": 2800
    },
    {
      "epoch": 0.3067484662576687,
      "grad_norm": 0.8199774026870728,
      "learning_rate": 0.0005822217433466221,
      "loss": 4.366,
      "step": 2850
    },
    {
      "epoch": 0.31213001829727693,
      "grad_norm": 0.7131398916244507,
      "learning_rate": 0.0005818985023165607,
      "loss": 4.3393,
      "step": 2900
    },
    {
      "epoch": 0.31751157033688515,
      "grad_norm": 0.8093417286872864,
      "learning_rate": 0.0005815752612864992,
      "loss": 4.3397,
      "step": 2950
    },
    {
      "epoch": 0.32289312237649337,
      "grad_norm": 0.9575026631355286,
      "learning_rate": 0.0005812520202564378,
      "loss": 4.3222,
      "step": 3000
    },
    {
      "epoch": 0.32289312237649337,
      "eval_accuracy": 0.297169233717903,
      "eval_loss": 4.248993873596191,
      "eval_runtime": 182.6748,
      "eval_samples_per_second": 98.596,
      "eval_steps_per_second": 6.164,
      "step": 3000
    },
    {
      "epoch": 0.3282746744161016,
      "grad_norm": 0.8685798645019531,
      "learning_rate": 0.0005809287792263764,
      "loss": 4.3126,
      "step": 3050
    },
    {
      "epoch": 0.3336562264557098,
      "grad_norm": 0.8055946826934814,
      "learning_rate": 0.0005806055381963151,
      "loss": 4.303,
      "step": 3100
    },
    {
      "epoch": 0.3390377784953181,
      "grad_norm": 0.8136595487594604,
      "learning_rate": 0.0005802822971662536,
      "loss": 4.2861,
      "step": 3150
    },
    {
      "epoch": 0.3444193305349263,
      "grad_norm": 0.6919870972633362,
      "learning_rate": 0.0005799590561361922,
      "loss": 4.3128,
      "step": 3200
    },
    {
      "epoch": 0.3498008825745345,
      "grad_norm": 1.4158786535263062,
      "learning_rate": 0.0005796358151061307,
      "loss": 4.3008,
      "step": 3250
    },
    {
      "epoch": 0.35518243461414273,
      "grad_norm": 0.8841201663017273,
      "learning_rate": 0.0005793125740760694,
      "loss": 4.2924,
      "step": 3300
    },
    {
      "epoch": 0.36056398665375095,
      "grad_norm": 0.6999953985214233,
      "learning_rate": 0.0005789893330460079,
      "loss": 4.2491,
      "step": 3350
    },
    {
      "epoch": 0.36594553869335916,
      "grad_norm": 0.8567237257957458,
      "learning_rate": 0.0005786660920159465,
      "loss": 4.2506,
      "step": 3400
    },
    {
      "epoch": 0.3713270907329674,
      "grad_norm": 0.779323160648346,
      "learning_rate": 0.0005783428509858851,
      "loss": 4.2543,
      "step": 3450
    },
    {
      "epoch": 0.3767086427725756,
      "grad_norm": 0.7086902260780334,
      "learning_rate": 0.0005780196099558237,
      "loss": 4.235,
      "step": 3500
    },
    {
      "epoch": 0.3820901948121838,
      "grad_norm": 0.5936145186424255,
      "learning_rate": 0.0005776963689257623,
      "loss": 4.2331,
      "step": 3550
    },
    {
      "epoch": 0.38747174685179203,
      "grad_norm": 0.632709264755249,
      "learning_rate": 0.0005773731278957008,
      "loss": 4.2431,
      "step": 3600
    },
    {
      "epoch": 0.3928532988914003,
      "grad_norm": 0.9441436529159546,
      "learning_rate": 0.0005770498868656394,
      "loss": 4.2204,
      "step": 3650
    },
    {
      "epoch": 0.3982348509310085,
      "grad_norm": 0.8455289602279663,
      "learning_rate": 0.000576726645835578,
      "loss": 4.2151,
      "step": 3700
    },
    {
      "epoch": 0.40361640297061674,
      "grad_norm": 0.6691498756408691,
      "learning_rate": 0.0005764034048055167,
      "loss": 4.1907,
      "step": 3750
    },
    {
      "epoch": 0.40899795501022496,
      "grad_norm": 0.764274001121521,
      "learning_rate": 0.0005760801637754552,
      "loss": 4.1993,
      "step": 3800
    },
    {
      "epoch": 0.4143795070498332,
      "grad_norm": 0.8627861142158508,
      "learning_rate": 0.0005757569227453937,
      "loss": 4.2007,
      "step": 3850
    },
    {
      "epoch": 0.4197610590894414,
      "grad_norm": 0.7309780120849609,
      "learning_rate": 0.0005754336817153324,
      "loss": 4.1748,
      "step": 3900
    },
    {
      "epoch": 0.4251426111290496,
      "grad_norm": 1.0626062154769897,
      "learning_rate": 0.0005751104406852709,
      "loss": 4.1841,
      "step": 3950
    },
    {
      "epoch": 0.4305241631686578,
      "grad_norm": 0.7397871017456055,
      "learning_rate": 0.0005747871996552096,
      "loss": 4.1793,
      "step": 4000
    },
    {
      "epoch": 0.4305241631686578,
      "eval_accuracy": 0.31092153838146963,
      "eval_loss": 4.1014580726623535,
      "eval_runtime": 183.3521,
      "eval_samples_per_second": 98.232,
      "eval_steps_per_second": 6.141,
      "step": 4000
    },
    {
      "epoch": 0.43590571520826604,
      "grad_norm": 0.7458042502403259,
      "learning_rate": 0.0005744639586251481,
      "loss": 4.1751,
      "step": 4050
    },
    {
      "epoch": 0.44128726724787426,
      "grad_norm": 0.6242780685424805,
      "learning_rate": 0.0005741407175950867,
      "loss": 4.1522,
      "step": 4100
    },
    {
      "epoch": 0.44666881928748253,
      "grad_norm": 0.8113278746604919,
      "learning_rate": 0.0005738174765650253,
      "loss": 4.1662,
      "step": 4150
    },
    {
      "epoch": 0.45205037132709075,
      "grad_norm": 0.7726860046386719,
      "learning_rate": 0.0005734942355349638,
      "loss": 4.1576,
      "step": 4200
    },
    {
      "epoch": 0.45743192336669897,
      "grad_norm": 0.5663923621177673,
      "learning_rate": 0.0005731709945049025,
      "loss": 4.1365,
      "step": 4250
    },
    {
      "epoch": 0.4628134754063072,
      "grad_norm": 0.9014385938644409,
      "learning_rate": 0.000572847753474841,
      "loss": 4.1356,
      "step": 4300
    },
    {
      "epoch": 0.4681950274459154,
      "grad_norm": 0.7335566282272339,
      "learning_rate": 0.0005725245124447796,
      "loss": 4.1453,
      "step": 4350
    },
    {
      "epoch": 0.4735765794855236,
      "grad_norm": 0.779979944229126,
      "learning_rate": 0.0005722012714147182,
      "loss": 4.1228,
      "step": 4400
    },
    {
      "epoch": 0.47895813152513184,
      "grad_norm": 0.8542835116386414,
      "learning_rate": 0.0005718780303846568,
      "loss": 4.1002,
      "step": 4450
    },
    {
      "epoch": 0.48433968356474005,
      "grad_norm": 0.6748495697975159,
      "learning_rate": 0.0005715547893545953,
      "loss": 4.1213,
      "step": 4500
    },
    {
      "epoch": 0.48972123560434827,
      "grad_norm": 0.7007385492324829,
      "learning_rate": 0.000571231548324534,
      "loss": 4.1179,
      "step": 4550
    },
    {
      "epoch": 0.49510278764395654,
      "grad_norm": 0.6953123807907104,
      "learning_rate": 0.0005709083072944725,
      "loss": 4.0866,
      "step": 4600
    },
    {
      "epoch": 0.5004843396835648,
      "grad_norm": 0.6920002102851868,
      "learning_rate": 0.0005705850662644111,
      "loss": 4.1128,
      "step": 4650
    },
    {
      "epoch": 0.505865891723173,
      "grad_norm": 0.6578840613365173,
      "learning_rate": 0.0005702618252343497,
      "loss": 4.1059,
      "step": 4700
    },
    {
      "epoch": 0.5112474437627812,
      "grad_norm": 0.7448212504386902,
      "learning_rate": 0.0005699385842042882,
      "loss": 4.1112,
      "step": 4750
    },
    {
      "epoch": 0.5166289958023894,
      "grad_norm": 0.6418218612670898,
      "learning_rate": 0.0005696153431742269,
      "loss": 4.083,
      "step": 4800
    },
    {
      "epoch": 0.5220105478419976,
      "grad_norm": 0.5668521523475647,
      "learning_rate": 0.0005692921021441655,
      "loss": 4.092,
      "step": 4850
    },
    {
      "epoch": 0.5273920998816058,
      "grad_norm": 0.6103580594062805,
      "learning_rate": 0.0005689688611141041,
      "loss": 4.0851,
      "step": 4900
    },
    {
      "epoch": 0.5327736519212141,
      "grad_norm": 0.691785454750061,
      "learning_rate": 0.0005686456200840426,
      "loss": 4.0865,
      "step": 4950
    },
    {
      "epoch": 0.5381552039608223,
      "grad_norm": 0.725322425365448,
      "learning_rate": 0.0005683223790539811,
      "loss": 4.0646,
      "step": 5000
    },
    {
      "epoch": 0.5381552039608223,
      "eval_accuracy": 0.3200998824267101,
      "eval_loss": 4.004289627075195,
      "eval_runtime": 183.8369,
      "eval_samples_per_second": 97.973,
      "eval_steps_per_second": 6.125,
      "step": 5000
    },
    {
      "epoch": 0.5435367560004305,
      "grad_norm": 0.7255427837371826,
      "learning_rate": 0.0005679991380239198,
      "loss": 4.0629,
      "step": 5050
    },
    {
      "epoch": 0.5489183080400387,
      "grad_norm": 0.8185265064239502,
      "learning_rate": 0.0005676758969938584,
      "loss": 4.0571,
      "step": 5100
    },
    {
      "epoch": 0.5542998600796469,
      "grad_norm": 0.6164122223854065,
      "learning_rate": 0.000567352655963797,
      "loss": 4.0833,
      "step": 5150
    },
    {
      "epoch": 0.5596814121192552,
      "grad_norm": 0.7035121917724609,
      "learning_rate": 0.0005670294149337355,
      "loss": 4.0478,
      "step": 5200
    },
    {
      "epoch": 0.5650629641588634,
      "grad_norm": 0.6536286473274231,
      "learning_rate": 0.0005667061739036742,
      "loss": 4.0478,
      "step": 5250
    },
    {
      "epoch": 0.5704445161984716,
      "grad_norm": 0.6543012261390686,
      "learning_rate": 0.0005663829328736127,
      "loss": 4.0573,
      "step": 5300
    },
    {
      "epoch": 0.5758260682380799,
      "grad_norm": 0.7880141139030457,
      "learning_rate": 0.0005660596918435512,
      "loss": 4.037,
      "step": 5350
    },
    {
      "epoch": 0.5812076202776881,
      "grad_norm": 0.6059675216674805,
      "learning_rate": 0.0005657364508134899,
      "loss": 4.0243,
      "step": 5400
    },
    {
      "epoch": 0.5865891723172963,
      "grad_norm": 0.5372035503387451,
      "learning_rate": 0.0005654132097834284,
      "loss": 4.044,
      "step": 5450
    },
    {
      "epoch": 0.5919707243569046,
      "grad_norm": 0.652932345867157,
      "learning_rate": 0.0005650899687533671,
      "loss": 4.0212,
      "step": 5500
    },
    {
      "epoch": 0.5973522763965128,
      "grad_norm": 0.827506422996521,
      "learning_rate": 0.0005647667277233056,
      "loss": 4.0347,
      "step": 5550
    },
    {
      "epoch": 0.602733828436121,
      "grad_norm": 0.6992992758750916,
      "learning_rate": 0.0005644434866932442,
      "loss": 4.0602,
      "step": 5600
    },
    {
      "epoch": 0.6081153804757292,
      "grad_norm": 0.6638625860214233,
      "learning_rate": 0.0005641202456631828,
      "loss": 4.0149,
      "step": 5650
    },
    {
      "epoch": 0.6134969325153374,
      "grad_norm": 0.6855315566062927,
      "learning_rate": 0.0005637970046331214,
      "loss": 3.9967,
      "step": 5700
    },
    {
      "epoch": 0.6188784845549457,
      "grad_norm": 0.594383955001831,
      "learning_rate": 0.00056347376360306,
      "loss": 4.0143,
      "step": 5750
    },
    {
      "epoch": 0.6242600365945539,
      "grad_norm": 1.1790200471878052,
      "learning_rate": 0.0005631505225729985,
      "loss": 4.0077,
      "step": 5800
    },
    {
      "epoch": 0.6296415886341621,
      "grad_norm": 0.6079861521720886,
      "learning_rate": 0.0005628272815429371,
      "loss": 4.0057,
      "step": 5850
    },
    {
      "epoch": 0.6350231406737703,
      "grad_norm": 0.6142423748970032,
      "learning_rate": 0.0005625040405128757,
      "loss": 4.0097,
      "step": 5900
    },
    {
      "epoch": 0.6404046927133785,
      "grad_norm": 0.6306520700454712,
      "learning_rate": 0.0005621807994828143,
      "loss": 4.0023,
      "step": 5950
    },
    {
      "epoch": 0.6457862447529867,
      "grad_norm": 0.5871993899345398,
      "learning_rate": 0.0005618575584527529,
      "loss": 4.0013,
      "step": 6000
    },
    {
      "epoch": 0.6457862447529867,
      "eval_accuracy": 0.327245113635166,
      "eval_loss": 3.9299139976501465,
      "eval_runtime": 183.1308,
      "eval_samples_per_second": 98.35,
      "eval_steps_per_second": 6.149,
      "step": 6000
    },
    {
      "epoch": 0.651167796792595,
      "grad_norm": 0.6969402432441711,
      "learning_rate": 0.0005615407822432927,
      "loss": 4.0079,
      "step": 6050
    },
    {
      "epoch": 0.6565493488322032,
      "grad_norm": 0.6455370187759399,
      "learning_rate": 0.0005612175412132313,
      "loss": 3.976,
      "step": 6100
    },
    {
      "epoch": 0.6619309008718114,
      "grad_norm": 0.696642279624939,
      "learning_rate": 0.0005608943001831699,
      "loss": 3.9809,
      "step": 6150
    },
    {
      "epoch": 0.6673124529114196,
      "grad_norm": 0.6556864380836487,
      "learning_rate": 0.0005605710591531085,
      "loss": 3.9824,
      "step": 6200
    },
    {
      "epoch": 0.6726940049510278,
      "grad_norm": 0.5885508060455322,
      "learning_rate": 0.000560247818123047,
      "loss": 3.9573,
      "step": 6250
    },
    {
      "epoch": 0.6780755569906362,
      "grad_norm": 0.6368624567985535,
      "learning_rate": 0.0005599245770929855,
      "loss": 3.9602,
      "step": 6300
    },
    {
      "epoch": 0.6834571090302444,
      "grad_norm": 0.6564461588859558,
      "learning_rate": 0.0005596013360629242,
      "loss": 3.9735,
      "step": 6350
    },
    {
      "epoch": 0.6888386610698526,
      "grad_norm": 0.5797594785690308,
      "learning_rate": 0.0005592780950328628,
      "loss": 3.9884,
      "step": 6400
    },
    {
      "epoch": 0.6942202131094608,
      "grad_norm": 0.5808694958686829,
      "learning_rate": 0.0005589548540028014,
      "loss": 3.981,
      "step": 6450
    },
    {
      "epoch": 0.699601765149069,
      "grad_norm": 0.6396571397781372,
      "learning_rate": 0.0005586316129727399,
      "loss": 3.9754,
      "step": 6500
    },
    {
      "epoch": 0.7049833171886772,
      "grad_norm": 0.5532586574554443,
      "learning_rate": 0.0005583083719426786,
      "loss": 3.9448,
      "step": 6550
    },
    {
      "epoch": 0.7103648692282855,
      "grad_norm": 0.6465068459510803,
      "learning_rate": 0.0005579851309126171,
      "loss": 3.9362,
      "step": 6600
    },
    {
      "epoch": 0.7157464212678937,
      "grad_norm": 0.6378588080406189,
      "learning_rate": 0.0005576618898825558,
      "loss": 3.9604,
      "step": 6650
    },
    {
      "epoch": 0.7211279733075019,
      "grad_norm": 0.5902277231216431,
      "learning_rate": 0.0005573386488524943,
      "loss": 3.9683,
      "step": 6700
    },
    {
      "epoch": 0.7265095253471101,
      "grad_norm": 0.6843133568763733,
      "learning_rate": 0.0005570154078224328,
      "loss": 3.97,
      "step": 6750
    },
    {
      "epoch": 0.7318910773867183,
      "grad_norm": 0.5807896256446838,
      "learning_rate": 0.0005566921667923715,
      "loss": 3.9434,
      "step": 6800
    },
    {
      "epoch": 0.7372726294263265,
      "grad_norm": 0.5920153856277466,
      "learning_rate": 0.00055636892576231,
      "loss": 3.9583,
      "step": 6850
    },
    {
      "epoch": 0.7426541814659348,
      "grad_norm": 0.6305286288261414,
      "learning_rate": 0.0005560456847322487,
      "loss": 3.9369,
      "step": 6900
    },
    {
      "epoch": 0.748035733505543,
      "grad_norm": 0.6596742868423462,
      "learning_rate": 0.0005557224437021872,
      "loss": 3.9349,
      "step": 6950
    },
    {
      "epoch": 0.7534172855451512,
      "grad_norm": 0.5761475563049316,
      "learning_rate": 0.0005553992026721258,
      "loss": 3.9392,
      "step": 7000
    },
    {
      "epoch": 0.7534172855451512,
      "eval_accuracy": 0.33236201273390115,
      "eval_loss": 3.871555805206299,
      "eval_runtime": 182.3203,
      "eval_samples_per_second": 98.788,
      "eval_steps_per_second": 6.176,
      "step": 7000
    },
    {
      "epoch": 0.7587988375847594,
      "grad_norm": 0.7002006769180298,
      "learning_rate": 0.0005550759616420644,
      "loss": 3.9563,
      "step": 7050
    },
    {
      "epoch": 0.7641803896243676,
      "grad_norm": 0.6454929113388062,
      "learning_rate": 0.000554752720612003,
      "loss": 3.9207,
      "step": 7100
    },
    {
      "epoch": 0.7695619416639758,
      "grad_norm": 0.5613568425178528,
      "learning_rate": 0.0005544294795819415,
      "loss": 3.9345,
      "step": 7150
    },
    {
      "epoch": 0.7749434937035841,
      "grad_norm": 0.575527012348175,
      "learning_rate": 0.0005541062385518801,
      "loss": 3.9269,
      "step": 7200
    },
    {
      "epoch": 0.7803250457431924,
      "grad_norm": 0.5978026986122131,
      "learning_rate": 0.0005537829975218188,
      "loss": 3.9484,
      "step": 7250
    },
    {
      "epoch": 0.7857065977828006,
      "grad_norm": 0.5839754939079285,
      "learning_rate": 0.0005534597564917573,
      "loss": 3.9268,
      "step": 7300
    },
    {
      "epoch": 0.7910881498224088,
      "grad_norm": 0.6187883019447327,
      "learning_rate": 0.0005531365154616959,
      "loss": 3.9256,
      "step": 7350
    },
    {
      "epoch": 0.796469701862017,
      "grad_norm": 0.6549730896949768,
      "learning_rate": 0.0005528132744316344,
      "loss": 3.92,
      "step": 7400
    },
    {
      "epoch": 0.8018512539016253,
      "grad_norm": 0.5937124490737915,
      "learning_rate": 0.0005524900334015731,
      "loss": 3.8838,
      "step": 7450
    },
    {
      "epoch": 0.8072328059412335,
      "grad_norm": 0.5783231854438782,
      "learning_rate": 0.0005521667923715117,
      "loss": 3.9049,
      "step": 7500
    },
    {
      "epoch": 0.8126143579808417,
      "grad_norm": 0.5541702508926392,
      "learning_rate": 0.0005518435513414502,
      "loss": 3.9001,
      "step": 7550
    },
    {
      "epoch": 0.8179959100204499,
      "grad_norm": 0.5935503244400024,
      "learning_rate": 0.0005515203103113888,
      "loss": 3.9101,
      "step": 7600
    },
    {
      "epoch": 0.8233774620600581,
      "grad_norm": 0.6102548241615295,
      "learning_rate": 0.0005511970692813274,
      "loss": 3.9069,
      "step": 7650
    },
    {
      "epoch": 0.8287590140996663,
      "grad_norm": 0.5769250988960266,
      "learning_rate": 0.000550873828251266,
      "loss": 3.9094,
      "step": 7700
    },
    {
      "epoch": 0.8341405661392746,
      "grad_norm": 0.5763354301452637,
      "learning_rate": 0.0005505505872212045,
      "loss": 3.9138,
      "step": 7750
    },
    {
      "epoch": 0.8395221181788828,
      "grad_norm": 0.6475393176078796,
      "learning_rate": 0.0005502273461911432,
      "loss": 3.901,
      "step": 7800
    },
    {
      "epoch": 0.844903670218491,
      "grad_norm": 0.5787845849990845,
      "learning_rate": 0.0005499041051610817,
      "loss": 3.9031,
      "step": 7850
    },
    {
      "epoch": 0.8502852222580992,
      "grad_norm": 0.5721645355224609,
      "learning_rate": 0.0005495808641310204,
      "loss": 3.8975,
      "step": 7900
    },
    {
      "epoch": 0.8556667742977074,
      "grad_norm": 0.6674014925956726,
      "learning_rate": 0.0005492576231009589,
      "loss": 3.8725,
      "step": 7950
    },
    {
      "epoch": 0.8610483263373157,
      "grad_norm": 0.7338675260543823,
      "learning_rate": 0.0005489343820708974,
      "loss": 3.8996,
      "step": 8000
    },
    {
      "epoch": 0.8610483263373157,
      "eval_accuracy": 0.3371352427484791,
      "eval_loss": 3.8227832317352295,
      "eval_runtime": 184.0783,
      "eval_samples_per_second": 97.844,
      "eval_steps_per_second": 6.117,
      "step": 8000
    },
    {
      "epoch": 0.8664298783769239,
      "grad_norm": 0.5568352341651917,
      "learning_rate": 0.0005486111410408361,
      "loss": 3.8899,
      "step": 8050
    },
    {
      "epoch": 0.8718114304165321,
      "grad_norm": 0.5996348261833191,
      "learning_rate": 0.0005482879000107746,
      "loss": 3.8833,
      "step": 8100
    },
    {
      "epoch": 0.8771929824561403,
      "grad_norm": 0.6273235082626343,
      "learning_rate": 0.0005479646589807133,
      "loss": 3.8805,
      "step": 8150
    },
    {
      "epoch": 0.8825745344957485,
      "grad_norm": 0.6420896649360657,
      "learning_rate": 0.0005476414179506518,
      "loss": 3.8839,
      "step": 8200
    },
    {
      "epoch": 0.8879560865353568,
      "grad_norm": 0.5380816459655762,
      "learning_rate": 0.0005473246417411916,
      "loss": 3.8628,
      "step": 8250
    },
    {
      "epoch": 0.8933376385749651,
      "grad_norm": 0.6761050224304199,
      "learning_rate": 0.0005470014007111302,
      "loss": 3.8861,
      "step": 8300
    },
    {
      "epoch": 0.8987191906145733,
      "grad_norm": 0.6020174622535706,
      "learning_rate": 0.0005466781596810688,
      "loss": 3.8795,
      "step": 8350
    },
    {
      "epoch": 0.9041007426541815,
      "grad_norm": 0.6668645143508911,
      "learning_rate": 0.0005463549186510073,
      "loss": 3.8844,
      "step": 8400
    },
    {
      "epoch": 0.9094822946937897,
      "grad_norm": 0.5670798420906067,
      "learning_rate": 0.000546031677620946,
      "loss": 3.8835,
      "step": 8450
    },
    {
      "epoch": 0.9148638467333979,
      "grad_norm": 0.5120532512664795,
      "learning_rate": 0.0005457084365908845,
      "loss": 3.8562,
      "step": 8500
    },
    {
      "epoch": 0.9202453987730062,
      "grad_norm": 0.5259217023849487,
      "learning_rate": 0.0005453851955608232,
      "loss": 3.8672,
      "step": 8550
    },
    {
      "epoch": 0.9256269508126144,
      "grad_norm": 0.583608865737915,
      "learning_rate": 0.0005450619545307617,
      "loss": 3.8806,
      "step": 8600
    },
    {
      "epoch": 0.9310085028522226,
      "grad_norm": 0.64499431848526,
      "learning_rate": 0.0005447387135007003,
      "loss": 3.8743,
      "step": 8650
    },
    {
      "epoch": 0.9363900548918308,
      "grad_norm": 0.5465984344482422,
      "learning_rate": 0.0005444154724706389,
      "loss": 3.8684,
      "step": 8700
    },
    {
      "epoch": 0.941771606931439,
      "grad_norm": 0.5706291198730469,
      "learning_rate": 0.0005440922314405775,
      "loss": 3.8527,
      "step": 8750
    },
    {
      "epoch": 0.9471531589710472,
      "grad_norm": 0.5907847285270691,
      "learning_rate": 0.0005437689904105161,
      "loss": 3.8579,
      "step": 8800
    },
    {
      "epoch": 0.9525347110106555,
      "grad_norm": 0.6252136826515198,
      "learning_rate": 0.0005434457493804546,
      "loss": 3.8496,
      "step": 8850
    },
    {
      "epoch": 0.9579162630502637,
      "grad_norm": 0.5910587310791016,
      "learning_rate": 0.0005431225083503932,
      "loss": 3.8492,
      "step": 8900
    },
    {
      "epoch": 0.9632978150898719,
      "grad_norm": 0.5492550730705261,
      "learning_rate": 0.0005427992673203318,
      "loss": 3.8486,
      "step": 8950
    },
    {
      "epoch": 0.9686793671294801,
      "grad_norm": 0.6138157844543457,
      "learning_rate": 0.0005424760262902704,
      "loss": 3.8404,
      "step": 9000
    },
    {
      "epoch": 0.9686793671294801,
      "eval_accuracy": 0.3408974576419433,
      "eval_loss": 3.782958984375,
      "eval_runtime": 183.7954,
      "eval_samples_per_second": 97.995,
      "eval_steps_per_second": 6.126,
      "step": 9000
    },
    {
      "epoch": 0.9740609191690883,
      "grad_norm": 0.5365199446678162,
      "learning_rate": 0.000542152785260209,
      "loss": 3.8344,
      "step": 9050
    },
    {
      "epoch": 0.9794424712086965,
      "grad_norm": 0.5951078534126282,
      "learning_rate": 0.0005418295442301476,
      "loss": 3.849,
      "step": 9100
    },
    {
      "epoch": 0.9848240232483048,
      "grad_norm": 0.586760401725769,
      "learning_rate": 0.0005415063032000861,
      "loss": 3.8422,
      "step": 9150
    },
    {
      "epoch": 0.9902055752879131,
      "grad_norm": 0.5982652902603149,
      "learning_rate": 0.0005411830621700248,
      "loss": 3.8481,
      "step": 9200
    },
    {
      "epoch": 0.9955871273275213,
      "grad_norm": 0.5718969106674194,
      "learning_rate": 0.0005408598211399633,
      "loss": 3.8469,
      "step": 9250
    },
    {
      "epoch": 1.0009686793671295,
      "grad_norm": 0.6520361304283142,
      "learning_rate": 0.0005405365801099019,
      "loss": 3.8171,
      "step": 9300
    },
    {
      "epoch": 1.0063502314067376,
      "grad_norm": 0.5989835262298584,
      "learning_rate": 0.0005402133390798405,
      "loss": 3.7906,
      "step": 9350
    },
    {
      "epoch": 1.011731783446346,
      "grad_norm": 0.5910343527793884,
      "learning_rate": 0.000539890098049779,
      "loss": 3.7743,
      "step": 9400
    },
    {
      "epoch": 1.017113335485954,
      "grad_norm": 0.531580924987793,
      "learning_rate": 0.0005395668570197177,
      "loss": 3.7769,
      "step": 9450
    },
    {
      "epoch": 1.0224948875255624,
      "grad_norm": 0.5643661022186279,
      "learning_rate": 0.0005392436159896562,
      "loss": 3.7784,
      "step": 9500
    },
    {
      "epoch": 1.0278764395651705,
      "grad_norm": 0.5594536662101746,
      "learning_rate": 0.0005389203749595948,
      "loss": 3.7859,
      "step": 9550
    },
    {
      "epoch": 1.0332579916047788,
      "grad_norm": 0.5723885297775269,
      "learning_rate": 0.0005385971339295334,
      "loss": 3.7723,
      "step": 9600
    },
    {
      "epoch": 1.0386395436443872,
      "grad_norm": 0.5776793956756592,
      "learning_rate": 0.000538273892899472,
      "loss": 3.7622,
      "step": 9650
    },
    {
      "epoch": 1.0440210956839953,
      "grad_norm": 1.0429363250732422,
      "learning_rate": 0.0005379506518694106,
      "loss": 3.7716,
      "step": 9700
    },
    {
      "epoch": 1.0494026477236036,
      "grad_norm": 0.6570566296577454,
      "learning_rate": 0.0005376274108393491,
      "loss": 3.7688,
      "step": 9750
    },
    {
      "epoch": 1.0547841997632117,
      "grad_norm": 0.5422590374946594,
      "learning_rate": 0.0005373041698092877,
      "loss": 3.775,
      "step": 9800
    },
    {
      "epoch": 1.06016575180282,
      "grad_norm": 0.5695186853408813,
      "learning_rate": 0.0005369809287792263,
      "loss": 3.7836,
      "step": 9850
    },
    {
      "epoch": 1.0655473038424281,
      "grad_norm": 0.5725084543228149,
      "learning_rate": 0.000536657687749165,
      "loss": 3.7777,
      "step": 9900
    },
    {
      "epoch": 1.0709288558820365,
      "grad_norm": 0.5916622281074524,
      "learning_rate": 0.0005363344467191035,
      "loss": 3.7616,
      "step": 9950
    },
    {
      "epoch": 1.0763104079216446,
      "grad_norm": 0.6363723278045654,
      "learning_rate": 0.000536011205689042,
      "loss": 3.7704,
      "step": 10000
    },
    {
      "epoch": 1.0763104079216446,
      "eval_accuracy": 0.34435577040819043,
      "eval_loss": 3.753837823867798,
      "eval_runtime": 183.8836,
      "eval_samples_per_second": 97.948,
      "eval_steps_per_second": 6.123,
      "step": 10000
    },
    {
      "epoch": 1.081691959961253,
      "grad_norm": 0.6295488476753235,
      "learning_rate": 0.0005356879646589807,
      "loss": 3.7654,
      "step": 10050
    },
    {
      "epoch": 1.087073512000861,
      "grad_norm": 0.5673714280128479,
      "learning_rate": 0.0005353647236289192,
      "loss": 3.7797,
      "step": 10100
    },
    {
      "epoch": 1.0924550640404693,
      "grad_norm": 0.5380173325538635,
      "learning_rate": 0.0005350414825988579,
      "loss": 3.767,
      "step": 10150
    },
    {
      "epoch": 1.0978366160800774,
      "grad_norm": 0.5471936464309692,
      "learning_rate": 0.0005347182415687964,
      "loss": 3.7705,
      "step": 10200
    },
    {
      "epoch": 1.1032181681196858,
      "grad_norm": 0.5841044187545776,
      "learning_rate": 0.000534395000538735,
      "loss": 3.7676,
      "step": 10250
    },
    {
      "epoch": 1.1085997201592939,
      "grad_norm": 0.5935708284378052,
      "learning_rate": 0.0005340717595086736,
      "loss": 3.7552,
      "step": 10300
    },
    {
      "epoch": 1.1139812721989022,
      "grad_norm": 0.5655228495597839,
      "learning_rate": 0.0005337485184786122,
      "loss": 3.7727,
      "step": 10350
    },
    {
      "epoch": 1.1193628242385103,
      "grad_norm": 0.5181326270103455,
      "learning_rate": 0.0005334252774485507,
      "loss": 3.7737,
      "step": 10400
    },
    {
      "epoch": 1.1247443762781186,
      "grad_norm": 0.5479722619056702,
      "learning_rate": 0.0005331085012390905,
      "loss": 3.7639,
      "step": 10450
    },
    {
      "epoch": 1.1301259283177267,
      "grad_norm": 0.6414276361465454,
      "learning_rate": 0.0005327852602090292,
      "loss": 3.7678,
      "step": 10500
    },
    {
      "epoch": 1.135507480357335,
      "grad_norm": 0.5645982623100281,
      "learning_rate": 0.0005324620191789678,
      "loss": 3.781,
      "step": 10550
    },
    {
      "epoch": 1.1408890323969434,
      "grad_norm": 0.5996174812316895,
      "learning_rate": 0.0005321387781489063,
      "loss": 3.7616,
      "step": 10600
    },
    {
      "epoch": 1.1462705844365515,
      "grad_norm": 0.5798478126525879,
      "learning_rate": 0.0005318155371188449,
      "loss": 3.7509,
      "step": 10650
    },
    {
      "epoch": 1.1516521364761596,
      "grad_norm": 0.6145731210708618,
      "learning_rate": 0.0005314922960887834,
      "loss": 3.742,
      "step": 10700
    },
    {
      "epoch": 1.157033688515768,
      "grad_norm": 0.5426759123802185,
      "learning_rate": 0.0005311690550587221,
      "loss": 3.7573,
      "step": 10750
    },
    {
      "epoch": 1.1624152405553763,
      "grad_norm": 0.6367711424827576,
      "learning_rate": 0.0005308458140286607,
      "loss": 3.7447,
      "step": 10800
    },
    {
      "epoch": 1.1677967925949844,
      "grad_norm": 0.5617762804031372,
      "learning_rate": 0.0005305225729985993,
      "loss": 3.7552,
      "step": 10850
    },
    {
      "epoch": 1.1731783446345927,
      "grad_norm": 0.5642740726470947,
      "learning_rate": 0.0005301993319685378,
      "loss": 3.7717,
      "step": 10900
    },
    {
      "epoch": 1.1785598966742008,
      "grad_norm": 0.5389478206634521,
      "learning_rate": 0.0005298760909384765,
      "loss": 3.7591,
      "step": 10950
    },
    {
      "epoch": 1.1839414487138091,
      "grad_norm": 0.5747219920158386,
      "learning_rate": 0.0005295593147290163,
      "loss": 3.7569,
      "step": 11000
    },
    {
      "epoch": 1.1839414487138091,
      "eval_accuracy": 0.34640528983103497,
      "eval_loss": 3.7258360385894775,
      "eval_runtime": 181.9752,
      "eval_samples_per_second": 98.975,
      "eval_steps_per_second": 6.188,
      "step": 11000
    },
    {
      "epoch": 1.1893230007534172,
      "grad_norm": 0.535572350025177,
      "learning_rate": 0.0005292360736989548,
      "loss": 3.7503,
      "step": 11050
    },
    {
      "epoch": 1.1947045527930256,
      "grad_norm": 0.5539134740829468,
      "learning_rate": 0.0005289128326688933,
      "loss": 3.7425,
      "step": 11100
    },
    {
      "epoch": 1.2000861048326337,
      "grad_norm": 0.5157566070556641,
      "learning_rate": 0.000528589591638832,
      "loss": 3.7573,
      "step": 11150
    },
    {
      "epoch": 1.205467656872242,
      "grad_norm": 0.6375812888145447,
      "learning_rate": 0.0005282663506087705,
      "loss": 3.747,
      "step": 11200
    },
    {
      "epoch": 1.21084920891185,
      "grad_norm": 0.5762513279914856,
      "learning_rate": 0.0005279431095787092,
      "loss": 3.7664,
      "step": 11250
    },
    {
      "epoch": 1.2162307609514584,
      "grad_norm": 0.5227396488189697,
      "learning_rate": 0.0005276198685486477,
      "loss": 3.7466,
      "step": 11300
    },
    {
      "epoch": 1.2216123129910665,
      "grad_norm": 0.5795957446098328,
      "learning_rate": 0.0005272966275185863,
      "loss": 3.7608,
      "step": 11350
    },
    {
      "epoch": 1.2269938650306749,
      "grad_norm": 0.6271036863327026,
      "learning_rate": 0.0005269733864885249,
      "loss": 3.7467,
      "step": 11400
    },
    {
      "epoch": 1.232375417070283,
      "grad_norm": 0.6098865866661072,
      "learning_rate": 0.0005266501454584636,
      "loss": 3.7546,
      "step": 11450
    },
    {
      "epoch": 1.2377569691098913,
      "grad_norm": 0.5986586809158325,
      "learning_rate": 0.0005263269044284021,
      "loss": 3.7606,
      "step": 11500
    },
    {
      "epoch": 1.2431385211494996,
      "grad_norm": 0.564387857913971,
      "learning_rate": 0.0005260036633983406,
      "loss": 3.7539,
      "step": 11550
    },
    {
      "epoch": 1.2485200731891077,
      "grad_norm": 0.5610516667366028,
      "learning_rate": 0.0005256804223682792,
      "loss": 3.7399,
      "step": 11600
    },
    {
      "epoch": 1.2539016252287158,
      "grad_norm": 0.535050094127655,
      "learning_rate": 0.0005253571813382178,
      "loss": 3.7442,
      "step": 11650
    },
    {
      "epoch": 1.2592831772683242,
      "grad_norm": 0.600563645362854,
      "learning_rate": 0.0005250339403081564,
      "loss": 3.7306,
      "step": 11700
    },
    {
      "epoch": 1.2646647293079325,
      "grad_norm": 0.6017090678215027,
      "learning_rate": 0.000524710699278095,
      "loss": 3.7414,
      "step": 11750
    },
    {
      "epoch": 1.2700462813475406,
      "grad_norm": 0.7243601679801941,
      "learning_rate": 0.0005243874582480336,
      "loss": 3.748,
      "step": 11800
    },
    {
      "epoch": 1.275427833387149,
      "grad_norm": 0.5405592918395996,
      "learning_rate": 0.0005240642172179722,
      "loss": 3.7235,
      "step": 11850
    },
    {
      "epoch": 1.280809385426757,
      "grad_norm": 0.5994392037391663,
      "learning_rate": 0.0005237409761879107,
      "loss": 3.7396,
      "step": 11900
    },
    {
      "epoch": 1.2861909374663654,
      "grad_norm": 0.568977415561676,
      "learning_rate": 0.0005234177351578493,
      "loss": 3.7417,
      "step": 11950
    },
    {
      "epoch": 1.2915724895059735,
      "grad_norm": 0.5830429196357727,
      "learning_rate": 0.0005230944941277878,
      "loss": 3.7236,
      "step": 12000
    },
    {
      "epoch": 1.2915724895059735,
      "eval_accuracy": 0.3490606577563331,
      "eval_loss": 3.7004106044769287,
      "eval_runtime": 181.7656,
      "eval_samples_per_second": 99.089,
      "eval_steps_per_second": 6.195,
      "step": 12000
    },
    {
      "epoch": 1.2969540415455818,
      "grad_norm": 0.5280687212944031,
      "learning_rate": 0.0005227712530977265,
      "loss": 3.7264,
      "step": 12050
    },
    {
      "epoch": 1.30233559358519,
      "grad_norm": 0.5419384837150574,
      "learning_rate": 0.0005224480120676651,
      "loss": 3.7493,
      "step": 12100
    },
    {
      "epoch": 1.3077171456247982,
      "grad_norm": 0.5870584845542908,
      "learning_rate": 0.0005221247710376037,
      "loss": 3.7321,
      "step": 12150
    },
    {
      "epoch": 1.3130986976644063,
      "grad_norm": 0.6159378886222839,
      "learning_rate": 0.0005218015300075422,
      "loss": 3.7377,
      "step": 12200
    },
    {
      "epoch": 1.3184802497040147,
      "grad_norm": 0.6157721281051636,
      "learning_rate": 0.0005214782889774809,
      "loss": 3.7259,
      "step": 12250
    },
    {
      "epoch": 1.3238618017436228,
      "grad_norm": 0.6117875576019287,
      "learning_rate": 0.0005211550479474194,
      "loss": 3.7355,
      "step": 12300
    },
    {
      "epoch": 1.329243353783231,
      "grad_norm": 0.6459168791770935,
      "learning_rate": 0.000520831806917358,
      "loss": 3.7286,
      "step": 12350
    },
    {
      "epoch": 1.3346249058228392,
      "grad_norm": 0.5380271673202515,
      "learning_rate": 0.0005205085658872966,
      "loss": 3.7315,
      "step": 12400
    },
    {
      "epoch": 1.3400064578624475,
      "grad_norm": 0.5702387094497681,
      "learning_rate": 0.0005201853248572351,
      "loss": 3.7275,
      "step": 12450
    },
    {
      "epoch": 1.3453880099020559,
      "grad_norm": 0.6522827744483948,
      "learning_rate": 0.0005198620838271738,
      "loss": 3.731,
      "step": 12500
    },
    {
      "epoch": 1.350769561941664,
      "grad_norm": 0.6003380417823792,
      "learning_rate": 0.0005195388427971123,
      "loss": 3.7168,
      "step": 12550
    },
    {
      "epoch": 1.356151113981272,
      "grad_norm": 0.5687087178230286,
      "learning_rate": 0.000519215601767051,
      "loss": 3.7108,
      "step": 12600
    },
    {
      "epoch": 1.3615326660208804,
      "grad_norm": 0.602519690990448,
      "learning_rate": 0.0005188923607369895,
      "loss": 3.7322,
      "step": 12650
    },
    {
      "epoch": 1.3669142180604887,
      "grad_norm": 0.5567368268966675,
      "learning_rate": 0.000518569119706928,
      "loss": 3.7188,
      "step": 12700
    },
    {
      "epoch": 1.3722957701000968,
      "grad_norm": 0.6820892095565796,
      "learning_rate": 0.0005182458786768667,
      "loss": 3.7278,
      "step": 12750
    },
    {
      "epoch": 1.3776773221397052,
      "grad_norm": 0.6009151339530945,
      "learning_rate": 0.0005179226376468052,
      "loss": 3.7225,
      "step": 12800
    },
    {
      "epoch": 1.3830588741793133,
      "grad_norm": 0.5885601043701172,
      "learning_rate": 0.0005175993966167438,
      "loss": 3.7263,
      "step": 12850
    },
    {
      "epoch": 1.3884404262189216,
      "grad_norm": 0.6636022329330444,
      "learning_rate": 0.0005172761555866824,
      "loss": 3.7311,
      "step": 12900
    },
    {
      "epoch": 1.3938219782585297,
      "grad_norm": 0.6068388223648071,
      "learning_rate": 0.0005169529145566211,
      "loss": 3.7087,
      "step": 12950
    },
    {
      "epoch": 1.399203530298138,
      "grad_norm": 0.534371018409729,
      "learning_rate": 0.0005166296735265596,
      "loss": 3.7062,
      "step": 13000
    },
    {
      "epoch": 1.399203530298138,
      "eval_accuracy": 0.35125294707376586,
      "eval_loss": 3.679492712020874,
      "eval_runtime": 182.0328,
      "eval_samples_per_second": 98.944,
      "eval_steps_per_second": 6.186,
      "step": 13000
    },
    {
      "epoch": 1.4045850823377461,
      "grad_norm": 0.5494077801704407,
      "learning_rate": 0.0005163064324964982,
      "loss": 3.7203,
      "step": 13050
    },
    {
      "epoch": 1.4099666343773545,
      "grad_norm": 0.573820948600769,
      "learning_rate": 0.0005159831914664367,
      "loss": 3.6958,
      "step": 13100
    },
    {
      "epoch": 1.4153481864169626,
      "grad_norm": 0.6078840494155884,
      "learning_rate": 0.0005156599504363753,
      "loss": 3.7048,
      "step": 13150
    },
    {
      "epoch": 1.420729738456571,
      "grad_norm": 0.6435017585754395,
      "learning_rate": 0.000515336709406314,
      "loss": 3.719,
      "step": 13200
    },
    {
      "epoch": 1.426111290496179,
      "grad_norm": 0.5942484736442566,
      "learning_rate": 0.0005150134683762525,
      "loss": 3.724,
      "step": 13250
    },
    {
      "epoch": 1.4314928425357873,
      "grad_norm": 0.5379244685173035,
      "learning_rate": 0.0005146902273461911,
      "loss": 3.7028,
      "step": 13300
    },
    {
      "epoch": 1.4368743945753955,
      "grad_norm": 0.5623278617858887,
      "learning_rate": 0.0005143669863161297,
      "loss": 3.7117,
      "step": 13350
    },
    {
      "epoch": 1.4422559466150038,
      "grad_norm": 0.6037217378616333,
      "learning_rate": 0.0005140437452860683,
      "loss": 3.7123,
      "step": 13400
    },
    {
      "epoch": 1.447637498654612,
      "grad_norm": 0.6184905767440796,
      "learning_rate": 0.0005137205042560069,
      "loss": 3.7258,
      "step": 13450
    },
    {
      "epoch": 1.4530190506942202,
      "grad_norm": 0.5425773859024048,
      "learning_rate": 0.0005133972632259455,
      "loss": 3.6993,
      "step": 13500
    },
    {
      "epoch": 1.4584006027338283,
      "grad_norm": 0.6625587344169617,
      "learning_rate": 0.000513074022195884,
      "loss": 3.6997,
      "step": 13550
    },
    {
      "epoch": 1.4637821547734367,
      "grad_norm": 0.5868763327598572,
      "learning_rate": 0.0005127507811658226,
      "loss": 3.7111,
      "step": 13600
    },
    {
      "epoch": 1.469163706813045,
      "grad_norm": 0.5310501456260681,
      "learning_rate": 0.0005124275401357612,
      "loss": 3.7201,
      "step": 13650
    },
    {
      "epoch": 1.474545258852653,
      "grad_norm": 0.5664429664611816,
      "learning_rate": 0.0005121042991056997,
      "loss": 3.689,
      "step": 13700
    },
    {
      "epoch": 1.4799268108922612,
      "grad_norm": 0.5378715395927429,
      "learning_rate": 0.0005117810580756384,
      "loss": 3.7096,
      "step": 13750
    },
    {
      "epoch": 1.4853083629318695,
      "grad_norm": 0.5423511266708374,
      "learning_rate": 0.0005114578170455769,
      "loss": 3.7038,
      "step": 13800
    },
    {
      "epoch": 1.4906899149714778,
      "grad_norm": 0.5492138862609863,
      "learning_rate": 0.0005111345760155156,
      "loss": 3.7261,
      "step": 13850
    },
    {
      "epoch": 1.496071467011086,
      "grad_norm": 0.5724121928215027,
      "learning_rate": 0.0005108113349854541,
      "loss": 3.7067,
      "step": 13900
    },
    {
      "epoch": 1.501453019050694,
      "grad_norm": 0.5634517669677734,
      "learning_rate": 0.0005104880939553926,
      "loss": 3.6915,
      "step": 13950
    },
    {
      "epoch": 1.5068345710903024,
      "grad_norm": 0.6407066583633423,
      "learning_rate": 0.0005101648529253313,
      "loss": 3.7181,
      "step": 14000
    },
    {
      "epoch": 1.5068345710903024,
      "eval_accuracy": 0.35367851414133633,
      "eval_loss": 3.6570589542388916,
      "eval_runtime": 182.1403,
      "eval_samples_per_second": 98.885,
      "eval_steps_per_second": 6.182,
      "step": 14000
    },
    {
      "epoch": 1.5122161231299107,
      "grad_norm": 0.5960928797721863,
      "learning_rate": 0.0005098416118952699,
      "loss": 3.6972,
      "step": 14050
    },
    {
      "epoch": 1.5175976751695188,
      "grad_norm": 0.5764510035514832,
      "learning_rate": 0.0005095183708652085,
      "loss": 3.7012,
      "step": 14100
    },
    {
      "epoch": 1.5229792272091272,
      "grad_norm": 0.5632539391517639,
      "learning_rate": 0.000509195129835147,
      "loss": 3.7139,
      "step": 14150
    },
    {
      "epoch": 1.5283607792487355,
      "grad_norm": 0.5743679404258728,
      "learning_rate": 0.0005088718888050856,
      "loss": 3.6973,
      "step": 14200
    },
    {
      "epoch": 1.5337423312883436,
      "grad_norm": 0.5791763663291931,
      "learning_rate": 0.0005085486477750242,
      "loss": 3.7127,
      "step": 14250
    },
    {
      "epoch": 1.5391238833279517,
      "grad_norm": 0.5912649631500244,
      "learning_rate": 0.0005082254067449629,
      "loss": 3.6992,
      "step": 14300
    },
    {
      "epoch": 1.54450543536756,
      "grad_norm": 0.5650004744529724,
      "learning_rate": 0.0005079021657149014,
      "loss": 3.7003,
      "step": 14350
    },
    {
      "epoch": 1.5498869874071683,
      "grad_norm": 0.5783352851867676,
      "learning_rate": 0.0005075789246848399,
      "loss": 3.7108,
      "step": 14400
    },
    {
      "epoch": 1.5552685394467765,
      "grad_norm": 0.6116908192634583,
      "learning_rate": 0.0005072556836547785,
      "loss": 3.6884,
      "step": 14450
    },
    {
      "epoch": 1.5606500914863846,
      "grad_norm": 0.5558695197105408,
      "learning_rate": 0.0005069324426247171,
      "loss": 3.6849,
      "step": 14500
    },
    {
      "epoch": 1.566031643525993,
      "grad_norm": 0.6027122735977173,
      "learning_rate": 0.0005066092015946557,
      "loss": 3.6764,
      "step": 14550
    },
    {
      "epoch": 1.5714131955656012,
      "grad_norm": 0.6431997418403625,
      "learning_rate": 0.0005062859605645943,
      "loss": 3.7126,
      "step": 14600
    },
    {
      "epoch": 1.5767947476052093,
      "grad_norm": 0.548612117767334,
      "learning_rate": 0.0005059627195345329,
      "loss": 3.6901,
      "step": 14650
    },
    {
      "epoch": 1.5821762996448174,
      "grad_norm": 0.6005812883377075,
      "learning_rate": 0.0005056394785044715,
      "loss": 3.695,
      "step": 14700
    },
    {
      "epoch": 1.5875578516844258,
      "grad_norm": 0.6675664782524109,
      "learning_rate": 0.00050531623747441,
      "loss": 3.6855,
      "step": 14750
    },
    {
      "epoch": 1.592939403724034,
      "grad_norm": 0.6004554033279419,
      "learning_rate": 0.0005049929964443486,
      "loss": 3.6797,
      "step": 14800
    },
    {
      "epoch": 1.5983209557636422,
      "grad_norm": 0.578523576259613,
      "learning_rate": 0.0005046697554142871,
      "loss": 3.6766,
      "step": 14850
    },
    {
      "epoch": 1.6037025078032503,
      "grad_norm": 0.5477756857872009,
      "learning_rate": 0.0005043465143842258,
      "loss": 3.6804,
      "step": 14900
    },
    {
      "epoch": 1.6090840598428586,
      "grad_norm": 0.5516514778137207,
      "learning_rate": 0.0005040232733541644,
      "loss": 3.6764,
      "step": 14950
    },
    {
      "epoch": 1.614465611882467,
      "grad_norm": 0.5752219557762146,
      "learning_rate": 0.000503700032324103,
      "loss": 3.6986,
      "step": 15000
    },
    {
      "epoch": 1.614465611882467,
      "eval_accuracy": 0.3556933732929681,
      "eval_loss": 3.6386351585388184,
      "eval_runtime": 181.8961,
      "eval_samples_per_second": 99.018,
      "eval_steps_per_second": 6.19,
      "step": 15000
    },
    {
      "epoch": 1.619847163922075,
      "grad_norm": 0.5473290085792542,
      "learning_rate": 0.0005033832561146428,
      "loss": 3.6819,
      "step": 15050
    },
    {
      "epoch": 1.6252287159616834,
      "grad_norm": 0.5905124545097351,
      "learning_rate": 0.0005030600150845813,
      "loss": 3.672,
      "step": 15100
    },
    {
      "epoch": 1.6306102680012917,
      "grad_norm": 0.6535428762435913,
      "learning_rate": 0.00050273677405452,
      "loss": 3.6776,
      "step": 15150
    },
    {
      "epoch": 1.6359918200408998,
      "grad_norm": 0.5464186072349548,
      "learning_rate": 0.0005024135330244585,
      "loss": 3.6604,
      "step": 15200
    },
    {
      "epoch": 1.641373372080508,
      "grad_norm": 0.5981823801994324,
      "learning_rate": 0.0005020902919943972,
      "loss": 3.6632,
      "step": 15250
    },
    {
      "epoch": 1.6467549241201163,
      "grad_norm": 0.5628587603569031,
      "learning_rate": 0.0005017670509643357,
      "loss": 3.6775,
      "step": 15300
    },
    {
      "epoch": 1.6521364761597246,
      "grad_norm": 0.5168464183807373,
      "learning_rate": 0.0005014438099342743,
      "loss": 3.6883,
      "step": 15350
    },
    {
      "epoch": 1.6575180281993327,
      "grad_norm": 0.5790534019470215,
      "learning_rate": 0.0005011205689042129,
      "loss": 3.6784,
      "step": 15400
    },
    {
      "epoch": 1.6628995802389408,
      "grad_norm": 0.5653476119041443,
      "learning_rate": 0.0005007973278741514,
      "loss": 3.6495,
      "step": 15450
    },
    {
      "epoch": 1.6682811322785491,
      "grad_norm": 0.5761940479278564,
      "learning_rate": 0.00050047408684409,
      "loss": 3.6852,
      "step": 15500
    },
    {
      "epoch": 1.6736626843181575,
      "grad_norm": 0.5976137518882751,
      "learning_rate": 0.0005001508458140286,
      "loss": 3.6591,
      "step": 15550
    },
    {
      "epoch": 1.6790442363577656,
      "grad_norm": 0.5772688388824463,
      "learning_rate": 0.0004998276047839673,
      "loss": 3.654,
      "step": 15600
    },
    {
      "epoch": 1.6844257883973737,
      "grad_norm": 0.5581541061401367,
      "learning_rate": 0.0004995043637539058,
      "loss": 3.6748,
      "step": 15650
    },
    {
      "epoch": 1.689807340436982,
      "grad_norm": 0.5584582090377808,
      "learning_rate": 0.0004991811227238443,
      "loss": 3.666,
      "step": 15700
    },
    {
      "epoch": 1.6951888924765903,
      "grad_norm": 0.6056560277938843,
      "learning_rate": 0.0004988578816937829,
      "loss": 3.656,
      "step": 15750
    },
    {
      "epoch": 1.7005704445161984,
      "grad_norm": 0.5387526750564575,
      "learning_rate": 0.0004985346406637215,
      "loss": 3.6844,
      "step": 15800
    },
    {
      "epoch": 1.7059519965558065,
      "grad_norm": 0.6060013175010681,
      "learning_rate": 0.0004982113996336602,
      "loss": 3.6704,
      "step": 15850
    },
    {
      "epoch": 1.7113335485954149,
      "grad_norm": 0.5558328628540039,
      "learning_rate": 0.0004978881586035987,
      "loss": 3.6695,
      "step": 15900
    },
    {
      "epoch": 1.7167151006350232,
      "grad_norm": 0.6431989073753357,
      "learning_rate": 0.0004975649175735373,
      "loss": 3.6561,
      "step": 15950
    },
    {
      "epoch": 1.7220966526746313,
      "grad_norm": 0.5722593069076538,
      "learning_rate": 0.0004972416765434759,
      "loss": 3.6594,
      "step": 16000
    },
    {
      "epoch": 1.7220966526746313,
      "eval_accuracy": 0.35697623793939365,
      "eval_loss": 3.6226398944854736,
      "eval_runtime": 182.9771,
      "eval_samples_per_second": 98.433,
      "eval_steps_per_second": 6.154,
      "step": 16000
    },
    {
      "epoch": 1.7274782047142396,
      "grad_norm": 0.5558207035064697,
      "learning_rate": 0.0004969184355134145,
      "loss": 3.6741,
      "step": 16050
    },
    {
      "epoch": 1.732859756753848,
      "grad_norm": 0.6220034956932068,
      "learning_rate": 0.0004965951944833531,
      "loss": 3.6571,
      "step": 16100
    },
    {
      "epoch": 1.738241308793456,
      "grad_norm": 0.5504184365272522,
      "learning_rate": 0.0004962719534532916,
      "loss": 3.6812,
      "step": 16150
    },
    {
      "epoch": 1.7436228608330642,
      "grad_norm": 0.6175248622894287,
      "learning_rate": 0.0004959487124232302,
      "loss": 3.6989,
      "step": 16200
    },
    {
      "epoch": 1.7490044128726725,
      "grad_norm": 0.6186313629150391,
      "learning_rate": 0.0004956254713931688,
      "loss": 3.6716,
      "step": 16250
    },
    {
      "epoch": 1.7543859649122808,
      "grad_norm": 0.5942351818084717,
      "learning_rate": 0.0004953022303631074,
      "loss": 3.6773,
      "step": 16300
    },
    {
      "epoch": 1.759767516951889,
      "grad_norm": 0.5420483350753784,
      "learning_rate": 0.0004949789893330459,
      "loss": 3.6632,
      "step": 16350
    },
    {
      "epoch": 1.765149068991497,
      "grad_norm": 0.5678284764289856,
      "learning_rate": 0.0004946557483029846,
      "loss": 3.6603,
      "step": 16400
    },
    {
      "epoch": 1.7705306210311054,
      "grad_norm": 0.5654410719871521,
      "learning_rate": 0.0004943325072729231,
      "loss": 3.6556,
      "step": 16450
    },
    {
      "epoch": 1.7759121730707137,
      "grad_norm": 0.5799243450164795,
      "learning_rate": 0.0004940092662428617,
      "loss": 3.6594,
      "step": 16500
    },
    {
      "epoch": 1.7812937251103218,
      "grad_norm": 0.5611895322799683,
      "learning_rate": 0.0004936860252128003,
      "loss": 3.6558,
      "step": 16550
    },
    {
      "epoch": 1.78667527714993,
      "grad_norm": 0.6114226579666138,
      "learning_rate": 0.0004933627841827388,
      "loss": 3.6482,
      "step": 16600
    },
    {
      "epoch": 1.7920568291895382,
      "grad_norm": 0.5653012990951538,
      "learning_rate": 0.0004930395431526775,
      "loss": 3.6621,
      "step": 16650
    },
    {
      "epoch": 1.7974383812291466,
      "grad_norm": 0.5843732357025146,
      "learning_rate": 0.0004927163021226161,
      "loss": 3.6607,
      "step": 16700
    },
    {
      "epoch": 1.8028199332687547,
      "grad_norm": 0.622482180595398,
      "learning_rate": 0.0004923930610925547,
      "loss": 3.6551,
      "step": 16750
    },
    {
      "epoch": 1.8082014853083628,
      "grad_norm": 0.5526164770126343,
      "learning_rate": 0.0004920698200624932,
      "loss": 3.6533,
      "step": 16800
    },
    {
      "epoch": 1.813583037347971,
      "grad_norm": 0.6223166584968567,
      "learning_rate": 0.0004917465790324317,
      "loss": 3.6685,
      "step": 16850
    },
    {
      "epoch": 1.8189645893875794,
      "grad_norm": 0.6319237947463989,
      "learning_rate": 0.0004914233380023704,
      "loss": 3.6679,
      "step": 16900
    },
    {
      "epoch": 1.8243461414271875,
      "grad_norm": 0.5310279726982117,
      "learning_rate": 0.0004911000969723089,
      "loss": 3.6702,
      "step": 16950
    },
    {
      "epoch": 1.8297276934667959,
      "grad_norm": 0.5771204233169556,
      "learning_rate": 0.0004907768559422476,
      "loss": 3.6535,
      "step": 17000
    },
    {
      "epoch": 1.8297276934667959,
      "eval_accuracy": 0.3584541345194462,
      "eval_loss": 3.609053373336792,
      "eval_runtime": 183.1048,
      "eval_samples_per_second": 98.364,
      "eval_steps_per_second": 6.149,
      "step": 17000
    },
    {
      "epoch": 1.8351092455064042,
      "grad_norm": 0.6009894609451294,
      "learning_rate": 0.0004904536149121861,
      "loss": 3.6407,
      "step": 17050
    },
    {
      "epoch": 1.8404907975460123,
      "grad_norm": 0.5864176750183105,
      "learning_rate": 0.0004901303738821248,
      "loss": 3.6577,
      "step": 17100
    },
    {
      "epoch": 1.8458723495856204,
      "grad_norm": 0.5710988640785217,
      "learning_rate": 0.0004898135976726646,
      "loss": 3.6456,
      "step": 17150
    },
    {
      "epoch": 1.8512539016252287,
      "grad_norm": 0.6018273830413818,
      "learning_rate": 0.0004894903566426031,
      "loss": 3.6359,
      "step": 17200
    },
    {
      "epoch": 1.856635453664837,
      "grad_norm": 0.613945484161377,
      "learning_rate": 0.0004891671156125417,
      "loss": 3.6632,
      "step": 17250
    },
    {
      "epoch": 1.8620170057044452,
      "grad_norm": 0.5496900677680969,
      "learning_rate": 0.0004888438745824803,
      "loss": 3.6434,
      "step": 17300
    },
    {
      "epoch": 1.8673985577440533,
      "grad_norm": 0.5686019659042358,
      "learning_rate": 0.0004885206335524189,
      "loss": 3.6718,
      "step": 17350
    },
    {
      "epoch": 1.8727801097836616,
      "grad_norm": 0.5641106963157654,
      "learning_rate": 0.0004881973925223575,
      "loss": 3.6485,
      "step": 17400
    },
    {
      "epoch": 1.87816166182327,
      "grad_norm": 0.5751502513885498,
      "learning_rate": 0.00048787415149229604,
      "loss": 3.6619,
      "step": 17450
    },
    {
      "epoch": 1.883543213862878,
      "grad_norm": 0.5386393666267395,
      "learning_rate": 0.00048755091046223464,
      "loss": 3.6627,
      "step": 17500
    },
    {
      "epoch": 1.8889247659024861,
      "grad_norm": 0.5446299910545349,
      "learning_rate": 0.0004872276694321732,
      "loss": 3.6486,
      "step": 17550
    },
    {
      "epoch": 1.8943063179420945,
      "grad_norm": 0.5315663814544678,
      "learning_rate": 0.00048690442840211177,
      "loss": 3.6565,
      "step": 17600
    },
    {
      "epoch": 1.8996878699817028,
      "grad_norm": 0.5348063111305237,
      "learning_rate": 0.0004865811873720504,
      "loss": 3.6573,
      "step": 17650
    },
    {
      "epoch": 1.905069422021311,
      "grad_norm": 0.5544908046722412,
      "learning_rate": 0.00048625794634198896,
      "loss": 3.6473,
      "step": 17700
    },
    {
      "epoch": 1.910450974060919,
      "grad_norm": 0.5778284668922424,
      "learning_rate": 0.00048593470531192756,
      "loss": 3.6594,
      "step": 17750
    },
    {
      "epoch": 1.9158325261005273,
      "grad_norm": 0.566500723361969,
      "learning_rate": 0.00048561146428186615,
      "loss": 3.6411,
      "step": 17800
    },
    {
      "epoch": 1.9212140781401357,
      "grad_norm": 0.5555658340454102,
      "learning_rate": 0.0004852882232518047,
      "loss": 3.6261,
      "step": 17850
    },
    {
      "epoch": 1.9265956301797438,
      "grad_norm": 0.5867842435836792,
      "learning_rate": 0.00048496498222174334,
      "loss": 3.6291,
      "step": 17900
    },
    {
      "epoch": 1.931977182219352,
      "grad_norm": 0.5909476280212402,
      "learning_rate": 0.00048464174119168193,
      "loss": 3.6282,
      "step": 17950
    },
    {
      "epoch": 1.9373587342589604,
      "grad_norm": 0.642209529876709,
      "learning_rate": 0.0004843185001616205,
      "loss": 3.6528,
      "step": 18000
    },
    {
      "epoch": 1.9373587342589604,
      "eval_accuracy": 0.3601386888921219,
      "eval_loss": 3.5923776626586914,
      "eval_runtime": 182.6986,
      "eval_samples_per_second": 98.583,
      "eval_steps_per_second": 6.163,
      "step": 18000
    },
    {
      "epoch": 1.9427402862985685,
      "grad_norm": 0.5417137742042542,
      "learning_rate": 0.00048399525913155907,
      "loss": 3.6509,
      "step": 18050
    },
    {
      "epoch": 1.9481218383381766,
      "grad_norm": 0.5468534231185913,
      "learning_rate": 0.0004836720181014976,
      "loss": 3.6373,
      "step": 18100
    },
    {
      "epoch": 1.953503390377785,
      "grad_norm": 0.5604886412620544,
      "learning_rate": 0.0004833487770714362,
      "loss": 3.6485,
      "step": 18150
    },
    {
      "epoch": 1.9588849424173933,
      "grad_norm": 0.5776407718658447,
      "learning_rate": 0.00048302553604137485,
      "loss": 3.6288,
      "step": 18200
    },
    {
      "epoch": 1.9642664944570014,
      "grad_norm": 0.6095754504203796,
      "learning_rate": 0.0004827022950113134,
      "loss": 3.6571,
      "step": 18250
    },
    {
      "epoch": 1.9696480464966095,
      "grad_norm": 0.5740944147109985,
      "learning_rate": 0.000482379053981252,
      "loss": 3.6374,
      "step": 18300
    },
    {
      "epoch": 1.9750295985362178,
      "grad_norm": 0.5488967299461365,
      "learning_rate": 0.0004820558129511906,
      "loss": 3.6425,
      "step": 18350
    },
    {
      "epoch": 1.9804111505758262,
      "grad_norm": 0.5918818712234497,
      "learning_rate": 0.0004817325719211291,
      "loss": 3.6471,
      "step": 18400
    },
    {
      "epoch": 1.9857927026154343,
      "grad_norm": 0.5412439703941345,
      "learning_rate": 0.0004814093308910677,
      "loss": 3.6463,
      "step": 18450
    },
    {
      "epoch": 1.9911742546550424,
      "grad_norm": 0.5630549192428589,
      "learning_rate": 0.00048108608986100637,
      "loss": 3.6531,
      "step": 18500
    },
    {
      "epoch": 1.9965558066946507,
      "grad_norm": 0.5589700937271118,
      "learning_rate": 0.0004807628488309449,
      "loss": 3.65,
      "step": 18550
    },
    {
      "epoch": 2.001937358734259,
      "grad_norm": 0.5413640141487122,
      "learning_rate": 0.0004804396078008835,
      "loss": 3.6016,
      "step": 18600
    },
    {
      "epoch": 2.007318910773867,
      "grad_norm": 0.5609384179115295,
      "learning_rate": 0.00048011636677082204,
      "loss": 3.5453,
      "step": 18650
    },
    {
      "epoch": 2.0127004628134753,
      "grad_norm": 0.6163759827613831,
      "learning_rate": 0.00047979312574076064,
      "loss": 3.5521,
      "step": 18700
    },
    {
      "epoch": 2.018082014853084,
      "grad_norm": 0.5741806626319885,
      "learning_rate": 0.0004794698847106992,
      "loss": 3.5383,
      "step": 18750
    },
    {
      "epoch": 2.023463566892692,
      "grad_norm": 0.6155529022216797,
      "learning_rate": 0.0004791466436806378,
      "loss": 3.5698,
      "step": 18800
    },
    {
      "epoch": 2.0288451189323,
      "grad_norm": 0.6185212731361389,
      "learning_rate": 0.0004788234026505764,
      "loss": 3.5489,
      "step": 18850
    },
    {
      "epoch": 2.034226670971908,
      "grad_norm": 0.5259607434272766,
      "learning_rate": 0.00047850016162051496,
      "loss": 3.5425,
      "step": 18900
    },
    {
      "epoch": 2.0396082230115167,
      "grad_norm": 0.5763450264930725,
      "learning_rate": 0.00047817692059045356,
      "loss": 3.5663,
      "step": 18950
    },
    {
      "epoch": 2.044989775051125,
      "grad_norm": 0.5624204277992249,
      "learning_rate": 0.00047785367956039215,
      "loss": 3.5646,
      "step": 19000
    },
    {
      "epoch": 2.044989775051125,
      "eval_accuracy": 0.36152814202149347,
      "eval_loss": 3.5838420391082764,
      "eval_runtime": 183.3409,
      "eval_samples_per_second": 98.238,
      "eval_steps_per_second": 6.142,
      "step": 19000
    },
    {
      "epoch": 2.050371327090733,
      "grad_norm": 0.5626105666160583,
      "learning_rate": 0.00047753043853033075,
      "loss": 3.5464,
      "step": 19050
    },
    {
      "epoch": 2.055752879130341,
      "grad_norm": 0.5418869853019714,
      "learning_rate": 0.00047720719750026934,
      "loss": 3.5534,
      "step": 19100
    },
    {
      "epoch": 2.0611344311699495,
      "grad_norm": 0.6087026000022888,
      "learning_rate": 0.00047689042129080915,
      "loss": 3.5573,
      "step": 19150
    },
    {
      "epoch": 2.0665159832095576,
      "grad_norm": 0.5950330495834351,
      "learning_rate": 0.00047656718026074774,
      "loss": 3.5528,
      "step": 19200
    },
    {
      "epoch": 2.0718975352491658,
      "grad_norm": 0.5562347769737244,
      "learning_rate": 0.00047624393923068634,
      "loss": 3.5703,
      "step": 19250
    },
    {
      "epoch": 2.0772790872887743,
      "grad_norm": 0.5437284111976624,
      "learning_rate": 0.0004759206982006249,
      "loss": 3.5563,
      "step": 19300
    },
    {
      "epoch": 2.0826606393283824,
      "grad_norm": 0.6348011493682861,
      "learning_rate": 0.00047559745717056347,
      "loss": 3.5651,
      "step": 19350
    },
    {
      "epoch": 2.0880421913679905,
      "grad_norm": 0.5337033271789551,
      "learning_rate": 0.000475274216140502,
      "loss": 3.556,
      "step": 19400
    },
    {
      "epoch": 2.0934237434075986,
      "grad_norm": 0.6107613444328308,
      "learning_rate": 0.00047495097511044066,
      "loss": 3.5584,
      "step": 19450
    },
    {
      "epoch": 2.098805295447207,
      "grad_norm": 0.6084286570549011,
      "learning_rate": 0.00047462773408037925,
      "loss": 3.5613,
      "step": 19500
    },
    {
      "epoch": 2.1041868474868153,
      "grad_norm": 0.5908589959144592,
      "learning_rate": 0.0004743044930503178,
      "loss": 3.5732,
      "step": 19550
    },
    {
      "epoch": 2.1095683995264234,
      "grad_norm": 0.5735676884651184,
      "learning_rate": 0.0004739812520202564,
      "loss": 3.5512,
      "step": 19600
    },
    {
      "epoch": 2.1149499515660315,
      "grad_norm": 0.5397831201553345,
      "learning_rate": 0.000473658010990195,
      "loss": 3.5579,
      "step": 19650
    },
    {
      "epoch": 2.12033150360564,
      "grad_norm": 0.6003082394599915,
      "learning_rate": 0.0004733347699601336,
      "loss": 3.5704,
      "step": 19700
    },
    {
      "epoch": 2.125713055645248,
      "grad_norm": 0.6014171242713928,
      "learning_rate": 0.0004730115289300722,
      "loss": 3.5688,
      "step": 19750
    },
    {
      "epoch": 2.1310946076848563,
      "grad_norm": 0.5529822707176208,
      "learning_rate": 0.00047268828790001077,
      "loss": 3.5427,
      "step": 19800
    },
    {
      "epoch": 2.1364761597244644,
      "grad_norm": 0.6028132438659668,
      "learning_rate": 0.0004723650468699493,
      "loss": 3.5596,
      "step": 19850
    },
    {
      "epoch": 2.141857711764073,
      "grad_norm": 0.5742948651313782,
      "learning_rate": 0.0004720418058398879,
      "loss": 3.5591,
      "step": 19900
    },
    {
      "epoch": 2.147239263803681,
      "grad_norm": 0.6175876259803772,
      "learning_rate": 0.00047171856480982644,
      "loss": 3.5649,
      "step": 19950
    },
    {
      "epoch": 2.152620815843289,
      "grad_norm": 0.5835720300674438,
      "learning_rate": 0.0004713953237797651,
      "loss": 3.5654,
      "step": 20000
    },
    {
      "epoch": 2.152620815843289,
      "eval_accuracy": 0.36219320634780594,
      "eval_loss": 3.5732574462890625,
      "eval_runtime": 183.9589,
      "eval_samples_per_second": 97.908,
      "eval_steps_per_second": 6.121,
      "step": 20000
    },
    {
      "epoch": 2.1580023678828972,
      "grad_norm": 0.5701449513435364,
      "learning_rate": 0.0004710720827497037,
      "loss": 3.5597,
      "step": 20050
    },
    {
      "epoch": 2.163383919922506,
      "grad_norm": 0.5607191920280457,
      "learning_rate": 0.00047074884171964223,
      "loss": 3.5617,
      "step": 20100
    },
    {
      "epoch": 2.168765471962114,
      "grad_norm": 0.5911417603492737,
      "learning_rate": 0.0004704256006895808,
      "loss": 3.5635,
      "step": 20150
    },
    {
      "epoch": 2.174147024001722,
      "grad_norm": 0.5466926097869873,
      "learning_rate": 0.00047010235965951936,
      "loss": 3.5751,
      "step": 20200
    },
    {
      "epoch": 2.1795285760413305,
      "grad_norm": 0.58399897813797,
      "learning_rate": 0.00046977911862945796,
      "loss": 3.5768,
      "step": 20250
    },
    {
      "epoch": 2.1849101280809387,
      "grad_norm": 0.6332233548164368,
      "learning_rate": 0.0004694558775993966,
      "loss": 3.5679,
      "step": 20300
    },
    {
      "epoch": 2.1902916801205468,
      "grad_norm": 0.5964486598968506,
      "learning_rate": 0.00046913263656933515,
      "loss": 3.5699,
      "step": 20350
    },
    {
      "epoch": 2.195673232160155,
      "grad_norm": 0.5788668394088745,
      "learning_rate": 0.00046880939553927374,
      "loss": 3.5586,
      "step": 20400
    },
    {
      "epoch": 2.2010547841997634,
      "grad_norm": 0.5830016732215881,
      "learning_rate": 0.00046848615450921234,
      "loss": 3.5675,
      "step": 20450
    },
    {
      "epoch": 2.2064363362393715,
      "grad_norm": 0.6435354948043823,
      "learning_rate": 0.0004681629134791509,
      "loss": 3.572,
      "step": 20500
    },
    {
      "epoch": 2.2118178882789796,
      "grad_norm": 0.5667049884796143,
      "learning_rate": 0.00046783967244908947,
      "loss": 3.5638,
      "step": 20550
    },
    {
      "epoch": 2.2171994403185877,
      "grad_norm": 0.6753138899803162,
      "learning_rate": 0.0004675164314190281,
      "loss": 3.5719,
      "step": 20600
    },
    {
      "epoch": 2.2225809923581963,
      "grad_norm": 0.6224473118782043,
      "learning_rate": 0.00046719319038896666,
      "loss": 3.5729,
      "step": 20650
    },
    {
      "epoch": 2.2279625443978044,
      "grad_norm": 0.6385303735733032,
      "learning_rate": 0.00046686994935890526,
      "loss": 3.5713,
      "step": 20700
    },
    {
      "epoch": 2.2333440964374125,
      "grad_norm": 0.5930754542350769,
      "learning_rate": 0.0004665467083288438,
      "loss": 3.5553,
      "step": 20750
    },
    {
      "epoch": 2.2387256484770206,
      "grad_norm": 0.5501213669776917,
      "learning_rate": 0.0004662234672987824,
      "loss": 3.5775,
      "step": 20800
    },
    {
      "epoch": 2.244107200516629,
      "grad_norm": 0.5862070322036743,
      "learning_rate": 0.00046590022626872104,
      "loss": 3.5752,
      "step": 20850
    },
    {
      "epoch": 2.2494887525562373,
      "grad_norm": 0.5672935247421265,
      "learning_rate": 0.0004655769852386596,
      "loss": 3.5702,
      "step": 20900
    },
    {
      "epoch": 2.2548703045958454,
      "grad_norm": 0.564063310623169,
      "learning_rate": 0.0004652537442085982,
      "loss": 3.5571,
      "step": 20950
    },
    {
      "epoch": 2.2602518566354535,
      "grad_norm": 0.6813990473747253,
      "learning_rate": 0.00046493050317853677,
      "loss": 3.5613,
      "step": 21000
    },
    {
      "epoch": 2.2602518566354535,
      "eval_accuracy": 0.36378736151782,
      "eval_loss": 3.5629472732543945,
      "eval_runtime": 183.0527,
      "eval_samples_per_second": 98.392,
      "eval_steps_per_second": 6.151,
      "step": 21000
    },
    {
      "epoch": 2.265633408675062,
      "grad_norm": 0.5663150548934937,
      "learning_rate": 0.0004646072621484753,
      "loss": 3.5268,
      "step": 21050
    },
    {
      "epoch": 2.27101496071467,
      "grad_norm": 0.5659851431846619,
      "learning_rate": 0.00046429048593901517,
      "loss": 3.5753,
      "step": 21100
    },
    {
      "epoch": 2.2763965127542782,
      "grad_norm": 0.5690809488296509,
      "learning_rate": 0.0004639672449089537,
      "loss": 3.5542,
      "step": 21150
    },
    {
      "epoch": 2.281778064793887,
      "grad_norm": 0.6252328157424927,
      "learning_rate": 0.0004636440038788923,
      "loss": 3.565,
      "step": 21200
    },
    {
      "epoch": 2.287159616833495,
      "grad_norm": 0.5633666515350342,
      "learning_rate": 0.00046332076284883095,
      "loss": 3.5501,
      "step": 21250
    },
    {
      "epoch": 2.292541168873103,
      "grad_norm": 0.634242057800293,
      "learning_rate": 0.0004629975218187695,
      "loss": 3.5546,
      "step": 21300
    },
    {
      "epoch": 2.297922720912711,
      "grad_norm": 0.5788732767105103,
      "learning_rate": 0.0004626742807887081,
      "loss": 3.5609,
      "step": 21350
    },
    {
      "epoch": 2.303304272952319,
      "grad_norm": 0.5648044347763062,
      "learning_rate": 0.00046235103975864663,
      "loss": 3.546,
      "step": 21400
    },
    {
      "epoch": 2.3086858249919278,
      "grad_norm": 0.6103492379188538,
      "learning_rate": 0.0004620277987285852,
      "loss": 3.5766,
      "step": 21450
    },
    {
      "epoch": 2.314067377031536,
      "grad_norm": 0.6276610493659973,
      "learning_rate": 0.0004617045576985239,
      "loss": 3.5737,
      "step": 21500
    },
    {
      "epoch": 2.319448929071144,
      "grad_norm": 0.5594583749771118,
      "learning_rate": 0.0004613813166684624,
      "loss": 3.5475,
      "step": 21550
    },
    {
      "epoch": 2.3248304811107525,
      "grad_norm": 0.5839714407920837,
      "learning_rate": 0.000461058075638401,
      "loss": 3.5706,
      "step": 21600
    },
    {
      "epoch": 2.3302120331503606,
      "grad_norm": 0.5914338827133179,
      "learning_rate": 0.00046073483460833955,
      "loss": 3.5553,
      "step": 21650
    },
    {
      "epoch": 2.3355935851899687,
      "grad_norm": 0.5473167896270752,
      "learning_rate": 0.00046041159357827814,
      "loss": 3.5412,
      "step": 21700
    },
    {
      "epoch": 2.340975137229577,
      "grad_norm": 0.5862297415733337,
      "learning_rate": 0.00046008835254821674,
      "loss": 3.5511,
      "step": 21750
    },
    {
      "epoch": 2.3463566892691854,
      "grad_norm": 0.5849961042404175,
      "learning_rate": 0.0004597651115181554,
      "loss": 3.5424,
      "step": 21800
    },
    {
      "epoch": 2.3517382413087935,
      "grad_norm": 0.6141690015792847,
      "learning_rate": 0.00045944187048809393,
      "loss": 3.5617,
      "step": 21850
    },
    {
      "epoch": 2.3571197933484016,
      "grad_norm": 0.5442442893981934,
      "learning_rate": 0.0004591186294580325,
      "loss": 3.5594,
      "step": 21900
    },
    {
      "epoch": 2.3625013453880097,
      "grad_norm": 0.5921179056167603,
      "learning_rate": 0.00045879538842797106,
      "loss": 3.5472,
      "step": 21950
    },
    {
      "epoch": 2.3678828974276183,
      "grad_norm": 0.5729749798774719,
      "learning_rate": 0.00045847214739790966,
      "loss": 3.5588,
      "step": 22000
    },
    {
      "epoch": 2.3678828974276183,
      "eval_accuracy": 0.36438647354122905,
      "eval_loss": 3.5525951385498047,
      "eval_runtime": 183.627,
      "eval_samples_per_second": 98.085,
      "eval_steps_per_second": 6.132,
      "step": 22000
    },
    {
      "epoch": 2.3732644494672264,
      "grad_norm": 0.5790433287620544,
      "learning_rate": 0.0004581489063678482,
      "loss": 3.5655,
      "step": 22050
    },
    {
      "epoch": 2.3786460015068345,
      "grad_norm": 0.5972291827201843,
      "learning_rate": 0.00045782566533778685,
      "loss": 3.553,
      "step": 22100
    },
    {
      "epoch": 2.384027553546443,
      "grad_norm": 0.6700719594955444,
      "learning_rate": 0.00045750242430772544,
      "loss": 3.5554,
      "step": 22150
    },
    {
      "epoch": 2.389409105586051,
      "grad_norm": 0.5736136436462402,
      "learning_rate": 0.000457179183277664,
      "loss": 3.5505,
      "step": 22200
    },
    {
      "epoch": 2.3947906576256592,
      "grad_norm": 0.5969544053077698,
      "learning_rate": 0.0004568559422476026,
      "loss": 3.5582,
      "step": 22250
    },
    {
      "epoch": 2.4001722096652673,
      "grad_norm": 0.5612258911132812,
      "learning_rate": 0.00045653270121754117,
      "loss": 3.552,
      "step": 22300
    },
    {
      "epoch": 2.4055537617048754,
      "grad_norm": 0.5993139743804932,
      "learning_rate": 0.0004562094601874797,
      "loss": 3.5753,
      "step": 22350
    },
    {
      "epoch": 2.410935313744484,
      "grad_norm": 0.5695291757583618,
      "learning_rate": 0.00045588621915741836,
      "loss": 3.5636,
      "step": 22400
    },
    {
      "epoch": 2.416316865784092,
      "grad_norm": 0.6173685193061829,
      "learning_rate": 0.00045556297812735696,
      "loss": 3.5615,
      "step": 22450
    },
    {
      "epoch": 2.4216984178237,
      "grad_norm": 0.5470847487449646,
      "learning_rate": 0.0004552397370972955,
      "loss": 3.5524,
      "step": 22500
    },
    {
      "epoch": 2.4270799698633088,
      "grad_norm": 0.5483660697937012,
      "learning_rate": 0.0004549164960672341,
      "loss": 3.5562,
      "step": 22550
    },
    {
      "epoch": 2.432461521902917,
      "grad_norm": 0.630761444568634,
      "learning_rate": 0.00045459325503717263,
      "loss": 3.564,
      "step": 22600
    },
    {
      "epoch": 2.437843073942525,
      "grad_norm": 0.5955401062965393,
      "learning_rate": 0.0004542700140071113,
      "loss": 3.5453,
      "step": 22650
    },
    {
      "epoch": 2.443224625982133,
      "grad_norm": 0.5875324606895447,
      "learning_rate": 0.0004539467729770499,
      "loss": 3.5226,
      "step": 22700
    },
    {
      "epoch": 2.4486061780217416,
      "grad_norm": 0.5740516185760498,
      "learning_rate": 0.0004536235319469884,
      "loss": 3.5704,
      "step": 22750
    },
    {
      "epoch": 2.4539877300613497,
      "grad_norm": 0.5694280862808228,
      "learning_rate": 0.000453300290916927,
      "loss": 3.5505,
      "step": 22800
    },
    {
      "epoch": 2.459369282100958,
      "grad_norm": 0.6105132699012756,
      "learning_rate": 0.0004529770498868656,
      "loss": 3.5632,
      "step": 22850
    },
    {
      "epoch": 2.464750834140566,
      "grad_norm": 0.6120337843894958,
      "learning_rate": 0.00045265380885680414,
      "loss": 3.5523,
      "step": 22900
    },
    {
      "epoch": 2.4701323861801745,
      "grad_norm": 0.5750212669372559,
      "learning_rate": 0.0004523305678267428,
      "loss": 3.5434,
      "step": 22950
    },
    {
      "epoch": 2.4755139382197826,
      "grad_norm": 0.5812209248542786,
      "learning_rate": 0.0004520073267966814,
      "loss": 3.5626,
      "step": 23000
    },
    {
      "epoch": 2.4755139382197826,
      "eval_accuracy": 0.36594868476222564,
      "eval_loss": 3.5408027172088623,
      "eval_runtime": 182.9812,
      "eval_samples_per_second": 98.431,
      "eval_steps_per_second": 6.154,
      "step": 23000
    },
    {
      "epoch": 2.4808954902593907,
      "grad_norm": 0.6218942403793335,
      "learning_rate": 0.00045168408576661993,
      "loss": 3.5572,
      "step": 23050
    },
    {
      "epoch": 2.4862770422989993,
      "grad_norm": 0.5775257349014282,
      "learning_rate": 0.0004513608447365585,
      "loss": 3.5484,
      "step": 23100
    },
    {
      "epoch": 2.4916585943386074,
      "grad_norm": 0.5762562155723572,
      "learning_rate": 0.00045103760370649706,
      "loss": 3.5744,
      "step": 23150
    },
    {
      "epoch": 2.4970401463782155,
      "grad_norm": 0.5888144373893738,
      "learning_rate": 0.00045071436267643566,
      "loss": 3.5604,
      "step": 23200
    },
    {
      "epoch": 2.5024216984178236,
      "grad_norm": 0.6638675928115845,
      "learning_rate": 0.0004503911216463743,
      "loss": 3.5623,
      "step": 23250
    },
    {
      "epoch": 2.5078032504574317,
      "grad_norm": 0.620266318321228,
      "learning_rate": 0.00045006788061631285,
      "loss": 3.557,
      "step": 23300
    },
    {
      "epoch": 2.5131848024970402,
      "grad_norm": 0.5920720100402832,
      "learning_rate": 0.00044974463958625144,
      "loss": 3.5369,
      "step": 23350
    },
    {
      "epoch": 2.5185663545366483,
      "grad_norm": 0.5830405950546265,
      "learning_rate": 0.00044942139855619004,
      "loss": 3.5637,
      "step": 23400
    },
    {
      "epoch": 2.5239479065762565,
      "grad_norm": 0.5792795419692993,
      "learning_rate": 0.0004490981575261286,
      "loss": 3.5551,
      "step": 23450
    },
    {
      "epoch": 2.529329458615865,
      "grad_norm": 0.5703508257865906,
      "learning_rate": 0.0004487749164960672,
      "loss": 3.5656,
      "step": 23500
    },
    {
      "epoch": 2.534711010655473,
      "grad_norm": 0.5803419947624207,
      "learning_rate": 0.0004484516754660058,
      "loss": 3.5491,
      "step": 23550
    },
    {
      "epoch": 2.540092562695081,
      "grad_norm": 0.6119604706764221,
      "learning_rate": 0.00044812843443594436,
      "loss": 3.556,
      "step": 23600
    },
    {
      "epoch": 2.5454741147346893,
      "grad_norm": 0.5884591937065125,
      "learning_rate": 0.00044780519340588296,
      "loss": 3.5301,
      "step": 23650
    },
    {
      "epoch": 2.550855666774298,
      "grad_norm": 0.5345640778541565,
      "learning_rate": 0.0004474819523758215,
      "loss": 3.5249,
      "step": 23700
    },
    {
      "epoch": 2.556237218813906,
      "grad_norm": 0.5864933729171753,
      "learning_rate": 0.0004471587113457601,
      "loss": 3.5468,
      "step": 23750
    },
    {
      "epoch": 2.561618770853514,
      "grad_norm": 0.5660606622695923,
      "learning_rate": 0.00044683547031569874,
      "loss": 3.5576,
      "step": 23800
    },
    {
      "epoch": 2.567000322893122,
      "grad_norm": 0.5605663061141968,
      "learning_rate": 0.0004465122292856373,
      "loss": 3.5404,
      "step": 23850
    },
    {
      "epoch": 2.5723818749327307,
      "grad_norm": 0.6553633213043213,
      "learning_rate": 0.0004461889882555759,
      "loss": 3.5576,
      "step": 23900
    },
    {
      "epoch": 2.577763426972339,
      "grad_norm": 0.6218010783195496,
      "learning_rate": 0.00044586574722551447,
      "loss": 3.559,
      "step": 23950
    },
    {
      "epoch": 2.583144979011947,
      "grad_norm": 0.6036173701286316,
      "learning_rate": 0.000445542506195453,
      "loss": 3.5623,
      "step": 24000
    },
    {
      "epoch": 2.583144979011947,
      "eval_accuracy": 0.3662233592626207,
      "eval_loss": 3.533743381500244,
      "eval_runtime": 183.448,
      "eval_samples_per_second": 98.18,
      "eval_steps_per_second": 6.138,
      "step": 24000
    },
    {
      "epoch": 2.5885265310515555,
      "grad_norm": 0.6276186108589172,
      "learning_rate": 0.0004452192651653916,
      "loss": 3.5581,
      "step": 24050
    },
    {
      "epoch": 2.5939080830911636,
      "grad_norm": 0.6104943156242371,
      "learning_rate": 0.00044489602413533025,
      "loss": 3.5323,
      "step": 24100
    },
    {
      "epoch": 2.5992896351307717,
      "grad_norm": 0.5991529822349548,
      "learning_rate": 0.0004445727831052688,
      "loss": 3.5428,
      "step": 24150
    },
    {
      "epoch": 2.60467118717038,
      "grad_norm": 0.6116634607315063,
      "learning_rate": 0.0004442495420752074,
      "loss": 3.5436,
      "step": 24200
    },
    {
      "epoch": 2.610052739209988,
      "grad_norm": 0.637483537197113,
      "learning_rate": 0.00044392630104514593,
      "loss": 3.5658,
      "step": 24250
    },
    {
      "epoch": 2.6154342912495965,
      "grad_norm": 0.6104105710983276,
      "learning_rate": 0.0004436030600150845,
      "loss": 3.5557,
      "step": 24300
    },
    {
      "epoch": 2.6208158432892046,
      "grad_norm": 0.603550910949707,
      "learning_rate": 0.0004432798189850231,
      "loss": 3.5442,
      "step": 24350
    },
    {
      "epoch": 2.6261973953288127,
      "grad_norm": 0.5980435013771057,
      "learning_rate": 0.0004429565779549617,
      "loss": 3.5587,
      "step": 24400
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 0.556158185005188,
      "learning_rate": 0.0004426333369249003,
      "loss": 3.5477,
      "step": 24450
    },
    {
      "epoch": 2.6369604994080293,
      "grad_norm": 0.5729323029518127,
      "learning_rate": 0.0004423100958948389,
      "loss": 3.5398,
      "step": 24500
    },
    {
      "epoch": 2.6423420514476375,
      "grad_norm": 0.7184017896652222,
      "learning_rate": 0.00044198685486477744,
      "loss": 3.5459,
      "step": 24550
    },
    {
      "epoch": 2.6477236034872456,
      "grad_norm": 0.6241846680641174,
      "learning_rate": 0.00044166361383471604,
      "loss": 3.5393,
      "step": 24600
    },
    {
      "epoch": 2.653105155526854,
      "grad_norm": 0.625645637512207,
      "learning_rate": 0.0004413403728046547,
      "loss": 3.5611,
      "step": 24650
    },
    {
      "epoch": 2.658486707566462,
      "grad_norm": 0.5856419205665588,
      "learning_rate": 0.0004410171317745932,
      "loss": 3.5536,
      "step": 24700
    },
    {
      "epoch": 2.6638682596060703,
      "grad_norm": 0.6091840863227844,
      "learning_rate": 0.0004406938907445318,
      "loss": 3.541,
      "step": 24750
    },
    {
      "epoch": 2.6692498116456784,
      "grad_norm": 0.6365329027175903,
      "learning_rate": 0.00044037064971447036,
      "loss": 3.541,
      "step": 24800
    },
    {
      "epoch": 2.674631363685287,
      "grad_norm": 0.5892279744148254,
      "learning_rate": 0.00044004740868440896,
      "loss": 3.5459,
      "step": 24850
    },
    {
      "epoch": 2.680012915724895,
      "grad_norm": 0.6100271344184875,
      "learning_rate": 0.00043972416765434755,
      "loss": 3.5443,
      "step": 24900
    },
    {
      "epoch": 2.685394467764503,
      "grad_norm": 0.5854329466819763,
      "learning_rate": 0.00043940092662428615,
      "loss": 3.5453,
      "step": 24950
    },
    {
      "epoch": 2.6907760198041117,
      "grad_norm": 0.5759541392326355,
      "learning_rate": 0.00043907768559422474,
      "loss": 3.5571,
      "step": 25000
    },
    {
      "epoch": 2.6907760198041117,
      "eval_accuracy": 0.36772407294911424,
      "eval_loss": 3.526393413543701,
      "eval_runtime": 183.1728,
      "eval_samples_per_second": 98.328,
      "eval_steps_per_second": 6.147,
      "step": 25000
    },
    {
      "epoch": 2.69615757184372,
      "grad_norm": 0.5549103021621704,
      "learning_rate": 0.00043875444456416334,
      "loss": 3.5413,
      "step": 25050
    },
    {
      "epoch": 2.701539123883328,
      "grad_norm": 0.625041663646698,
      "learning_rate": 0.0004384312035341019,
      "loss": 3.5329,
      "step": 25100
    },
    {
      "epoch": 2.706920675922936,
      "grad_norm": 0.6317231059074402,
      "learning_rate": 0.0004381144273246417,
      "loss": 3.5234,
      "step": 25150
    },
    {
      "epoch": 2.712302227962544,
      "grad_norm": 0.5706122517585754,
      "learning_rate": 0.0004377911862945803,
      "loss": 3.5438,
      "step": 25200
    },
    {
      "epoch": 2.7176837800021527,
      "grad_norm": 0.6089178919792175,
      "learning_rate": 0.00043746794526451887,
      "loss": 3.5184,
      "step": 25250
    },
    {
      "epoch": 2.723065332041761,
      "grad_norm": 0.6188173890113831,
      "learning_rate": 0.00043714470423445747,
      "loss": 3.5516,
      "step": 25300
    },
    {
      "epoch": 2.728446884081369,
      "grad_norm": 0.5653114914894104,
      "learning_rate": 0.00043682146320439606,
      "loss": 3.5304,
      "step": 25350
    },
    {
      "epoch": 2.7338284361209775,
      "grad_norm": 0.6082587838172913,
      "learning_rate": 0.00043649822217433466,
      "loss": 3.5538,
      "step": 25400
    },
    {
      "epoch": 2.7392099881605856,
      "grad_norm": 0.5532857179641724,
      "learning_rate": 0.0004361749811442732,
      "loss": 3.5193,
      "step": 25450
    },
    {
      "epoch": 2.7445915402001937,
      "grad_norm": 0.6214589476585388,
      "learning_rate": 0.0004358517401142118,
      "loss": 3.536,
      "step": 25500
    },
    {
      "epoch": 2.749973092239802,
      "grad_norm": 0.5975282192230225,
      "learning_rate": 0.00043552849908415033,
      "loss": 3.5301,
      "step": 25550
    },
    {
      "epoch": 2.7553546442794103,
      "grad_norm": 0.6577773094177246,
      "learning_rate": 0.000435205258054089,
      "loss": 3.5566,
      "step": 25600
    },
    {
      "epoch": 2.7607361963190185,
      "grad_norm": 0.5862059593200684,
      "learning_rate": 0.0004348820170240276,
      "loss": 3.5315,
      "step": 25650
    },
    {
      "epoch": 2.7661177483586266,
      "grad_norm": 0.6143049001693726,
      "learning_rate": 0.0004345587759939661,
      "loss": 3.5541,
      "step": 25700
    },
    {
      "epoch": 2.7714993003982347,
      "grad_norm": 0.5826513767242432,
      "learning_rate": 0.0004342355349639047,
      "loss": 3.5328,
      "step": 25750
    },
    {
      "epoch": 2.776880852437843,
      "grad_norm": 0.6236239075660706,
      "learning_rate": 0.0004339122939338433,
      "loss": 3.5282,
      "step": 25800
    },
    {
      "epoch": 2.7822624044774513,
      "grad_norm": 0.5690028667449951,
      "learning_rate": 0.00043358905290378184,
      "loss": 3.5382,
      "step": 25850
    },
    {
      "epoch": 2.7876439565170594,
      "grad_norm": 0.630249559879303,
      "learning_rate": 0.0004332658118737205,
      "loss": 3.5471,
      "step": 25900
    },
    {
      "epoch": 2.793025508556668,
      "grad_norm": 0.561437726020813,
      "learning_rate": 0.0004329425708436591,
      "loss": 3.5452,
      "step": 25950
    },
    {
      "epoch": 2.798407060596276,
      "grad_norm": 0.5968010425567627,
      "learning_rate": 0.00043261932981359763,
      "loss": 3.5268,
      "step": 26000
    },
    {
      "epoch": 2.798407060596276,
      "eval_accuracy": 0.3683658855574344,
      "eval_loss": 3.5156972408294678,
      "eval_runtime": 183.4726,
      "eval_samples_per_second": 98.167,
      "eval_steps_per_second": 6.137,
      "step": 26000
    },
    {
      "epoch": 2.803788612635884,
      "grad_norm": 0.5962703227996826,
      "learning_rate": 0.0004322960887835362,
      "loss": 3.5435,
      "step": 26050
    },
    {
      "epoch": 2.8091701646754923,
      "grad_norm": 0.5704419016838074,
      "learning_rate": 0.00043197284775347476,
      "loss": 3.5439,
      "step": 26100
    },
    {
      "epoch": 2.8145517167151004,
      "grad_norm": 0.6311613917350769,
      "learning_rate": 0.00043164960672341336,
      "loss": 3.5348,
      "step": 26150
    },
    {
      "epoch": 2.819933268754709,
      "grad_norm": 0.5643249154090881,
      "learning_rate": 0.00043133283051395317,
      "loss": 3.5529,
      "step": 26200
    },
    {
      "epoch": 2.825314820794317,
      "grad_norm": 0.6420338749885559,
      "learning_rate": 0.0004310095894838918,
      "loss": 3.545,
      "step": 26250
    },
    {
      "epoch": 2.830696372833925,
      "grad_norm": 0.6328233480453491,
      "learning_rate": 0.0004306863484538304,
      "loss": 3.5405,
      "step": 26300
    },
    {
      "epoch": 2.8360779248735337,
      "grad_norm": 0.6627986431121826,
      "learning_rate": 0.00043036310742376895,
      "loss": 3.5228,
      "step": 26350
    },
    {
      "epoch": 2.841459476913142,
      "grad_norm": 0.5869179964065552,
      "learning_rate": 0.00043003986639370754,
      "loss": 3.5245,
      "step": 26400
    },
    {
      "epoch": 2.84684102895275,
      "grad_norm": 0.5820196270942688,
      "learning_rate": 0.0004297166253636461,
      "loss": 3.5315,
      "step": 26450
    },
    {
      "epoch": 2.852222580992358,
      "grad_norm": 0.6153207421302795,
      "learning_rate": 0.0004293933843335847,
      "loss": 3.5206,
      "step": 26500
    },
    {
      "epoch": 2.857604133031966,
      "grad_norm": 0.5706645846366882,
      "learning_rate": 0.00042907014330352333,
      "loss": 3.5492,
      "step": 26550
    },
    {
      "epoch": 2.8629856850715747,
      "grad_norm": 0.5671915411949158,
      "learning_rate": 0.00042874690227346187,
      "loss": 3.5261,
      "step": 26600
    },
    {
      "epoch": 2.868367237111183,
      "grad_norm": 0.6378110647201538,
      "learning_rate": 0.00042842366124340046,
      "loss": 3.5554,
      "step": 26650
    },
    {
      "epoch": 2.873748789150791,
      "grad_norm": 0.5837578773498535,
      "learning_rate": 0.00042810042021333906,
      "loss": 3.5286,
      "step": 26700
    },
    {
      "epoch": 2.8791303411903995,
      "grad_norm": 0.5720592737197876,
      "learning_rate": 0.0004277771791832776,
      "loss": 3.5193,
      "step": 26750
    },
    {
      "epoch": 2.8845118932300076,
      "grad_norm": 0.6068984270095825,
      "learning_rate": 0.0004274539381532162,
      "loss": 3.5375,
      "step": 26800
    },
    {
      "epoch": 2.8898934452696157,
      "grad_norm": 0.590977132320404,
      "learning_rate": 0.00042713069712315484,
      "loss": 3.515,
      "step": 26850
    },
    {
      "epoch": 2.895274997309224,
      "grad_norm": 0.6040093302726746,
      "learning_rate": 0.0004268074560930934,
      "loss": 3.5472,
      "step": 26900
    },
    {
      "epoch": 2.9006565493488323,
      "grad_norm": 0.5838053822517395,
      "learning_rate": 0.000426484215063032,
      "loss": 3.5313,
      "step": 26950
    },
    {
      "epoch": 2.9060381013884404,
      "grad_norm": 0.6135042905807495,
      "learning_rate": 0.0004261609740329705,
      "loss": 3.5395,
      "step": 27000
    },
    {
      "epoch": 2.9060381013884404,
      "eval_accuracy": 0.3692769400217588,
      "eval_loss": 3.50553035736084,
      "eval_runtime": 183.0394,
      "eval_samples_per_second": 98.4,
      "eval_steps_per_second": 6.152,
      "step": 27000
    },
    {
      "epoch": 2.9114196534280485,
      "grad_norm": 0.6675634384155273,
      "learning_rate": 0.0004258377330029091,
      "loss": 3.5392,
      "step": 27050
    },
    {
      "epoch": 2.9168012054676566,
      "grad_norm": 0.5531997680664062,
      "learning_rate": 0.00042551449197284776,
      "loss": 3.5078,
      "step": 27100
    },
    {
      "epoch": 2.922182757507265,
      "grad_norm": 0.6126731634140015,
      "learning_rate": 0.0004251912509427863,
      "loss": 3.5329,
      "step": 27150
    },
    {
      "epoch": 2.9275643095468733,
      "grad_norm": 0.5538141131401062,
      "learning_rate": 0.0004248680099127249,
      "loss": 3.5216,
      "step": 27200
    },
    {
      "epoch": 2.9329458615864814,
      "grad_norm": 0.5796554088592529,
      "learning_rate": 0.0004245447688826635,
      "loss": 3.5247,
      "step": 27250
    },
    {
      "epoch": 2.93832741362609,
      "grad_norm": 0.5609702467918396,
      "learning_rate": 0.00042422152785260203,
      "loss": 3.5229,
      "step": 27300
    },
    {
      "epoch": 2.943708965665698,
      "grad_norm": 0.6337541937828064,
      "learning_rate": 0.0004238982868225406,
      "loss": 3.4924,
      "step": 27350
    },
    {
      "epoch": 2.949090517705306,
      "grad_norm": 0.6238342523574829,
      "learning_rate": 0.0004235750457924793,
      "loss": 3.5259,
      "step": 27400
    },
    {
      "epoch": 2.9544720697449143,
      "grad_norm": 0.707619309425354,
      "learning_rate": 0.0004232518047624178,
      "loss": 3.5301,
      "step": 27450
    },
    {
      "epoch": 2.9598536217845224,
      "grad_norm": 0.5899302959442139,
      "learning_rate": 0.0004229285637323564,
      "loss": 3.5246,
      "step": 27500
    },
    {
      "epoch": 2.965235173824131,
      "grad_norm": 0.552964985370636,
      "learning_rate": 0.00042260532270229495,
      "loss": 3.5266,
      "step": 27550
    },
    {
      "epoch": 2.970616725863739,
      "grad_norm": 0.5927809476852417,
      "learning_rate": 0.00042228208167223354,
      "loss": 3.5179,
      "step": 27600
    },
    {
      "epoch": 2.975998277903347,
      "grad_norm": 0.6026394367218018,
      "learning_rate": 0.00042195884064217214,
      "loss": 3.531,
      "step": 27650
    },
    {
      "epoch": 2.9813798299429557,
      "grad_norm": 0.6079704165458679,
      "learning_rate": 0.00042163559961211073,
      "loss": 3.5447,
      "step": 27700
    },
    {
      "epoch": 2.986761381982564,
      "grad_norm": 0.5754170417785645,
      "learning_rate": 0.00042131235858204933,
      "loss": 3.5106,
      "step": 27750
    },
    {
      "epoch": 2.992142934022172,
      "grad_norm": 0.5957307815551758,
      "learning_rate": 0.0004209891175519879,
      "loss": 3.5292,
      "step": 27800
    },
    {
      "epoch": 2.9975244860617805,
      "grad_norm": 0.5584477782249451,
      "learning_rate": 0.00042066587652192646,
      "loss": 3.517,
      "step": 27850
    },
    {
      "epoch": 3.0029060381013886,
      "grad_norm": 0.5913463234901428,
      "learning_rate": 0.00042034263549186506,
      "loss": 3.4829,
      "step": 27900
    },
    {
      "epoch": 3.0082875901409967,
      "grad_norm": 0.5728430151939392,
      "learning_rate": 0.0004200193944618036,
      "loss": 3.4424,
      "step": 27950
    },
    {
      "epoch": 3.0136691421806048,
      "grad_norm": 0.6232100129127502,
      "learning_rate": 0.00041969615343174225,
      "loss": 3.4403,
      "step": 28000
    },
    {
      "epoch": 3.0136691421806048,
      "eval_accuracy": 0.3704998282741108,
      "eval_loss": 3.501554250717163,
      "eval_runtime": 181.6654,
      "eval_samples_per_second": 99.144,
      "eval_steps_per_second": 6.198,
      "step": 28000
    },
    {
      "epoch": 3.0190506942202133,
      "grad_norm": 0.5570206046104431,
      "learning_rate": 0.00041937291240168084,
      "loss": 3.436,
      "step": 28050
    },
    {
      "epoch": 3.0244322462598214,
      "grad_norm": 0.6119909882545471,
      "learning_rate": 0.0004190496713716194,
      "loss": 3.4374,
      "step": 28100
    },
    {
      "epoch": 3.0298137982994295,
      "grad_norm": 0.6302096843719482,
      "learning_rate": 0.000418726430341558,
      "loss": 3.4448,
      "step": 28150
    },
    {
      "epoch": 3.0351953503390376,
      "grad_norm": 0.5777623653411865,
      "learning_rate": 0.00041840318931149657,
      "loss": 3.4356,
      "step": 28200
    },
    {
      "epoch": 3.040576902378646,
      "grad_norm": 0.5832511782646179,
      "learning_rate": 0.00041807994828143517,
      "loss": 3.4348,
      "step": 28250
    },
    {
      "epoch": 3.0459584544182543,
      "grad_norm": 0.6186639666557312,
      "learning_rate": 0.00041775670725137376,
      "loss": 3.4531,
      "step": 28300
    },
    {
      "epoch": 3.0513400064578624,
      "grad_norm": 0.5793587565422058,
      "learning_rate": 0.00041743346622131236,
      "loss": 3.4351,
      "step": 28350
    },
    {
      "epoch": 3.0567215584974705,
      "grad_norm": 0.6191322207450867,
      "learning_rate": 0.0004171102251912509,
      "loss": 3.4609,
      "step": 28400
    },
    {
      "epoch": 3.062103110537079,
      "grad_norm": 0.5854801535606384,
      "learning_rate": 0.0004167869841611895,
      "loss": 3.445,
      "step": 28450
    },
    {
      "epoch": 3.067484662576687,
      "grad_norm": 0.5954617857933044,
      "learning_rate": 0.00041646374313112803,
      "loss": 3.4516,
      "step": 28500
    },
    {
      "epoch": 3.0728662146162953,
      "grad_norm": 0.6144933104515076,
      "learning_rate": 0.0004161405021010667,
      "loss": 3.4306,
      "step": 28550
    },
    {
      "epoch": 3.0782477666559034,
      "grad_norm": 0.6332607269287109,
      "learning_rate": 0.0004158172610710053,
      "loss": 3.4483,
      "step": 28600
    },
    {
      "epoch": 3.083629318695512,
      "grad_norm": 0.6351560354232788,
      "learning_rate": 0.0004154940200409438,
      "loss": 3.4597,
      "step": 28650
    },
    {
      "epoch": 3.08901087073512,
      "grad_norm": 0.580599308013916,
      "learning_rate": 0.0004151707790108824,
      "loss": 3.4463,
      "step": 28700
    },
    {
      "epoch": 3.094392422774728,
      "grad_norm": 0.636760950088501,
      "learning_rate": 0.00041484753798082095,
      "loss": 3.4619,
      "step": 28750
    },
    {
      "epoch": 3.0997739748143363,
      "grad_norm": 0.6538554430007935,
      "learning_rate": 0.00041452429695075955,
      "loss": 3.448,
      "step": 28800
    },
    {
      "epoch": 3.105155526853945,
      "grad_norm": 0.5733194351196289,
      "learning_rate": 0.0004142010559206982,
      "loss": 3.4343,
      "step": 28850
    },
    {
      "epoch": 3.110537078893553,
      "grad_norm": 0.6369328498840332,
      "learning_rate": 0.00041387781489063673,
      "loss": 3.4369,
      "step": 28900
    },
    {
      "epoch": 3.115918630933161,
      "grad_norm": 0.6201010942459106,
      "learning_rate": 0.00041355457386057533,
      "loss": 3.453,
      "step": 28950
    },
    {
      "epoch": 3.121300182972769,
      "grad_norm": 0.6407283544540405,
      "learning_rate": 0.0004132313328305139,
      "loss": 3.4636,
      "step": 29000
    },
    {
      "epoch": 3.121300182972769,
      "eval_accuracy": 0.3711315361638642,
      "eval_loss": 3.497187614440918,
      "eval_runtime": 181.453,
      "eval_samples_per_second": 99.26,
      "eval_steps_per_second": 6.205,
      "step": 29000
    },
    {
      "epoch": 3.1266817350123777,
      "grad_norm": 0.6396225690841675,
      "learning_rate": 0.00041290809180045246,
      "loss": 3.4613,
      "step": 29050
    },
    {
      "epoch": 3.132063287051986,
      "grad_norm": 0.6482574939727783,
      "learning_rate": 0.0004125848507703911,
      "loss": 3.4346,
      "step": 29100
    },
    {
      "epoch": 3.137444839091594,
      "grad_norm": 0.6442139744758606,
      "learning_rate": 0.0004122616097403297,
      "loss": 3.4496,
      "step": 29150
    },
    {
      "epoch": 3.1428263911312024,
      "grad_norm": 0.5868032574653625,
      "learning_rate": 0.0004119448335308695,
      "loss": 3.4653,
      "step": 29200
    },
    {
      "epoch": 3.1482079431708105,
      "grad_norm": 0.8085774779319763,
      "learning_rate": 0.0004116215925008081,
      "loss": 3.4591,
      "step": 29250
    },
    {
      "epoch": 3.1535894952104186,
      "grad_norm": 0.5951862931251526,
      "learning_rate": 0.00041129835147074665,
      "loss": 3.4455,
      "step": 29300
    },
    {
      "epoch": 3.1589710472500268,
      "grad_norm": 0.6378307938575745,
      "learning_rate": 0.00041097511044068524,
      "loss": 3.4538,
      "step": 29350
    },
    {
      "epoch": 3.1643525992896353,
      "grad_norm": 0.6090945601463318,
      "learning_rate": 0.0004106518694106238,
      "loss": 3.448,
      "step": 29400
    },
    {
      "epoch": 3.1697341513292434,
      "grad_norm": 0.6292023062705994,
      "learning_rate": 0.0004103286283805624,
      "loss": 3.4687,
      "step": 29450
    },
    {
      "epoch": 3.1751157033688515,
      "grad_norm": 0.6239890456199646,
      "learning_rate": 0.00041000538735050103,
      "loss": 3.4909,
      "step": 29500
    },
    {
      "epoch": 3.1804972554084596,
      "grad_norm": 0.5989189147949219,
      "learning_rate": 0.00040968214632043957,
      "loss": 3.4548,
      "step": 29550
    },
    {
      "epoch": 3.185878807448068,
      "grad_norm": 0.6178577542304993,
      "learning_rate": 0.00040935890529037816,
      "loss": 3.4575,
      "step": 29600
    },
    {
      "epoch": 3.1912603594876763,
      "grad_norm": 0.6444182395935059,
      "learning_rate": 0.00040903566426031676,
      "loss": 3.4525,
      "step": 29650
    },
    {
      "epoch": 3.1966419115272844,
      "grad_norm": 0.6180800199508667,
      "learning_rate": 0.0004087124232302553,
      "loss": 3.436,
      "step": 29700
    },
    {
      "epoch": 3.2020234635668925,
      "grad_norm": 0.595629096031189,
      "learning_rate": 0.0004083891822001939,
      "loss": 3.4835,
      "step": 29750
    },
    {
      "epoch": 3.207405015606501,
      "grad_norm": 0.6241080164909363,
      "learning_rate": 0.00040806594117013254,
      "loss": 3.454,
      "step": 29800
    },
    {
      "epoch": 3.212786567646109,
      "grad_norm": 0.6285580992698669,
      "learning_rate": 0.0004077427001400711,
      "loss": 3.4568,
      "step": 29850
    },
    {
      "epoch": 3.2181681196857173,
      "grad_norm": 0.648627758026123,
      "learning_rate": 0.0004074194591100097,
      "loss": 3.441,
      "step": 29900
    },
    {
      "epoch": 3.2235496717253254,
      "grad_norm": 0.64311283826828,
      "learning_rate": 0.0004070962180799482,
      "loss": 3.4699,
      "step": 29950
    },
    {
      "epoch": 3.228931223764934,
      "grad_norm": 0.5918191075325012,
      "learning_rate": 0.0004067729770498868,
      "loss": 3.4348,
      "step": 30000
    },
    {
      "epoch": 3.228931223764934,
      "eval_accuracy": 0.37159700513526145,
      "eval_loss": 3.49194073677063,
      "eval_runtime": 181.5347,
      "eval_samples_per_second": 99.215,
      "eval_steps_per_second": 6.203,
      "step": 30000
    },
    {
      "epoch": 3.234312775804542,
      "grad_norm": 0.6755573153495789,
      "learning_rate": 0.00040644973601982546,
      "loss": 3.4519,
      "step": 30050
    },
    {
      "epoch": 3.23969432784415,
      "grad_norm": 0.6835849285125732,
      "learning_rate": 0.000406126494989764,
      "loss": 3.4827,
      "step": 30100
    },
    {
      "epoch": 3.2450758798837587,
      "grad_norm": 0.6226943135261536,
      "learning_rate": 0.0004058032539597026,
      "loss": 3.4645,
      "step": 30150
    },
    {
      "epoch": 3.250457431923367,
      "grad_norm": 0.6609615087509155,
      "learning_rate": 0.00040548001292964114,
      "loss": 3.4642,
      "step": 30200
    },
    {
      "epoch": 3.255838983962975,
      "grad_norm": 0.5687957406044006,
      "learning_rate": 0.00040515677189957973,
      "loss": 3.4688,
      "step": 30250
    },
    {
      "epoch": 3.261220536002583,
      "grad_norm": 0.6109760999679565,
      "learning_rate": 0.0004048335308695183,
      "loss": 3.4437,
      "step": 30300
    },
    {
      "epoch": 3.2666020880421915,
      "grad_norm": 0.5959461331367493,
      "learning_rate": 0.0004045102898394569,
      "loss": 3.4724,
      "step": 30350
    },
    {
      "epoch": 3.2719836400817996,
      "grad_norm": 0.6271927952766418,
      "learning_rate": 0.0004041870488093955,
      "loss": 3.4651,
      "step": 30400
    },
    {
      "epoch": 3.2773651921214078,
      "grad_norm": 0.6263259649276733,
      "learning_rate": 0.0004038638077793341,
      "loss": 3.4471,
      "step": 30450
    },
    {
      "epoch": 3.282746744161016,
      "grad_norm": 0.6434895396232605,
      "learning_rate": 0.00040354056674927265,
      "loss": 3.4604,
      "step": 30500
    },
    {
      "epoch": 3.2881282962006244,
      "grad_norm": 0.661169707775116,
      "learning_rate": 0.00040321732571921124,
      "loss": 3.4547,
      "step": 30550
    },
    {
      "epoch": 3.2935098482402325,
      "grad_norm": 0.683679461479187,
      "learning_rate": 0.0004028940846891498,
      "loss": 3.4602,
      "step": 30600
    },
    {
      "epoch": 3.2988914002798406,
      "grad_norm": 0.6047713756561279,
      "learning_rate": 0.00040257084365908843,
      "loss": 3.4512,
      "step": 30650
    },
    {
      "epoch": 3.304272952319449,
      "grad_norm": 0.6232369542121887,
      "learning_rate": 0.00040224760262902703,
      "loss": 3.4713,
      "step": 30700
    },
    {
      "epoch": 3.3096545043590573,
      "grad_norm": 0.6141099333763123,
      "learning_rate": 0.00040192436159896557,
      "loss": 3.4627,
      "step": 30750
    },
    {
      "epoch": 3.3150360563986654,
      "grad_norm": 0.6429236531257629,
      "learning_rate": 0.00040160112056890416,
      "loss": 3.4586,
      "step": 30800
    },
    {
      "epoch": 3.3204176084382735,
      "grad_norm": 0.5798470377922058,
      "learning_rate": 0.00040127787953884276,
      "loss": 3.4562,
      "step": 30850
    },
    {
      "epoch": 3.3257991604778816,
      "grad_norm": 0.6334143280982971,
      "learning_rate": 0.00040095463850878135,
      "loss": 3.4408,
      "step": 30900
    },
    {
      "epoch": 3.33118071251749,
      "grad_norm": 0.6591746211051941,
      "learning_rate": 0.00040063139747871995,
      "loss": 3.4526,
      "step": 30950
    },
    {
      "epoch": 3.3365622645570983,
      "grad_norm": 0.6119304895401001,
      "learning_rate": 0.00040030815644865854,
      "loss": 3.4676,
      "step": 31000
    },
    {
      "epoch": 3.3365622645570983,
      "eval_accuracy": 0.3719315473768422,
      "eval_loss": 3.486844062805176,
      "eval_runtime": 182.7259,
      "eval_samples_per_second": 98.568,
      "eval_steps_per_second": 6.162,
      "step": 31000
    },
    {
      "epoch": 3.3419438165967064,
      "grad_norm": 0.6275706887245178,
      "learning_rate": 0.0003999849154185971,
      "loss": 3.4726,
      "step": 31050
    },
    {
      "epoch": 3.347325368636315,
      "grad_norm": 0.5887126326560974,
      "learning_rate": 0.0003996616743885357,
      "loss": 3.4399,
      "step": 31100
    },
    {
      "epoch": 3.352706920675923,
      "grad_norm": 0.6963056921958923,
      "learning_rate": 0.0003993384333584742,
      "loss": 3.4559,
      "step": 31150
    },
    {
      "epoch": 3.358088472715531,
      "grad_norm": 0.642962634563446,
      "learning_rate": 0.00039901519232841287,
      "loss": 3.4906,
      "step": 31200
    },
    {
      "epoch": 3.3634700247551392,
      "grad_norm": 0.6062067747116089,
      "learning_rate": 0.00039869195129835146,
      "loss": 3.4634,
      "step": 31250
    },
    {
      "epoch": 3.368851576794748,
      "grad_norm": 0.6534258723258972,
      "learning_rate": 0.00039836871026829,
      "loss": 3.4594,
      "step": 31300
    },
    {
      "epoch": 3.374233128834356,
      "grad_norm": 0.60047847032547,
      "learning_rate": 0.0003980454692382286,
      "loss": 3.4708,
      "step": 31350
    },
    {
      "epoch": 3.379614680873964,
      "grad_norm": 0.5911248326301575,
      "learning_rate": 0.0003977222282081672,
      "loss": 3.4498,
      "step": 31400
    },
    {
      "epoch": 3.384996232913572,
      "grad_norm": 0.5842762589454651,
      "learning_rate": 0.00039739898717810573,
      "loss": 3.472,
      "step": 31450
    },
    {
      "epoch": 3.3903777849531807,
      "grad_norm": 0.6286013126373291,
      "learning_rate": 0.0003970757461480444,
      "loss": 3.4783,
      "step": 31500
    },
    {
      "epoch": 3.3957593369927888,
      "grad_norm": 0.6078022718429565,
      "learning_rate": 0.000396752505117983,
      "loss": 3.4681,
      "step": 31550
    },
    {
      "epoch": 3.401140889032397,
      "grad_norm": 0.6259421706199646,
      "learning_rate": 0.0003964292640879215,
      "loss": 3.4482,
      "step": 31600
    },
    {
      "epoch": 3.4065224410720054,
      "grad_norm": 0.6382803916931152,
      "learning_rate": 0.0003961060230578601,
      "loss": 3.4746,
      "step": 31650
    },
    {
      "epoch": 3.4119039931116135,
      "grad_norm": 0.6442477107048035,
      "learning_rate": 0.00039578278202779865,
      "loss": 3.4505,
      "step": 31700
    },
    {
      "epoch": 3.4172855451512216,
      "grad_norm": 0.578318178653717,
      "learning_rate": 0.00039545954099773725,
      "loss": 3.4637,
      "step": 31750
    },
    {
      "epoch": 3.4226670971908297,
      "grad_norm": 0.6376295685768127,
      "learning_rate": 0.0003951362999676759,
      "loss": 3.4703,
      "step": 31800
    },
    {
      "epoch": 3.428048649230438,
      "grad_norm": 0.5918070673942566,
      "learning_rate": 0.00039481305893761444,
      "loss": 3.4539,
      "step": 31850
    },
    {
      "epoch": 3.4334302012700464,
      "grad_norm": 0.6081002354621887,
      "learning_rate": 0.0003944962827281543,
      "loss": 3.4505,
      "step": 31900
    },
    {
      "epoch": 3.4388117533096545,
      "grad_norm": 0.6261876225471497,
      "learning_rate": 0.00039417304169809284,
      "loss": 3.4755,
      "step": 31950
    },
    {
      "epoch": 3.4441933053492626,
      "grad_norm": 0.6414044499397278,
      "learning_rate": 0.00039384980066803143,
      "loss": 3.4556,
      "step": 32000
    },
    {
      "epoch": 3.4441933053492626,
      "eval_accuracy": 0.37261671248740036,
      "eval_loss": 3.481884002685547,
      "eval_runtime": 181.6828,
      "eval_samples_per_second": 99.134,
      "eval_steps_per_second": 6.198,
      "step": 32000
    },
    {
      "epoch": 3.449574857388871,
      "grad_norm": 0.6231973767280579,
      "learning_rate": 0.00039352655963796997,
      "loss": 3.4596,
      "step": 32050
    },
    {
      "epoch": 3.4549564094284793,
      "grad_norm": 0.6971520781517029,
      "learning_rate": 0.00039320331860790857,
      "loss": 3.4584,
      "step": 32100
    },
    {
      "epoch": 3.4603379614680874,
      "grad_norm": 0.6719905734062195,
      "learning_rate": 0.0003928800775778472,
      "loss": 3.4615,
      "step": 32150
    },
    {
      "epoch": 3.4657195135076955,
      "grad_norm": 0.7227779626846313,
      "learning_rate": 0.00039255683654778576,
      "loss": 3.4486,
      "step": 32200
    },
    {
      "epoch": 3.471101065547304,
      "grad_norm": 0.6186626553535461,
      "learning_rate": 0.00039223359551772435,
      "loss": 3.4744,
      "step": 32250
    },
    {
      "epoch": 3.476482617586912,
      "grad_norm": 0.6083580851554871,
      "learning_rate": 0.00039191035448766294,
      "loss": 3.4683,
      "step": 32300
    },
    {
      "epoch": 3.4818641696265202,
      "grad_norm": 0.6660602688789368,
      "learning_rate": 0.0003915871134576015,
      "loss": 3.4638,
      "step": 32350
    },
    {
      "epoch": 3.4872457216661283,
      "grad_norm": 0.6296412944793701,
      "learning_rate": 0.0003912638724275401,
      "loss": 3.4557,
      "step": 32400
    },
    {
      "epoch": 3.492627273705737,
      "grad_norm": 0.6064853072166443,
      "learning_rate": 0.00039094063139747873,
      "loss": 3.4495,
      "step": 32450
    },
    {
      "epoch": 3.498008825745345,
      "grad_norm": 0.6372223496437073,
      "learning_rate": 0.00039061739036741727,
      "loss": 3.4593,
      "step": 32500
    },
    {
      "epoch": 3.503390377784953,
      "grad_norm": 0.6473979949951172,
      "learning_rate": 0.00039029414933735586,
      "loss": 3.4618,
      "step": 32550
    },
    {
      "epoch": 3.5087719298245617,
      "grad_norm": 0.5884780883789062,
      "learning_rate": 0.0003899709083072944,
      "loss": 3.4682,
      "step": 32600
    },
    {
      "epoch": 3.5141534818641698,
      "grad_norm": 0.6261137127876282,
      "learning_rate": 0.000389647667277233,
      "loss": 3.4469,
      "step": 32650
    },
    {
      "epoch": 3.519535033903778,
      "grad_norm": 0.6143213510513306,
      "learning_rate": 0.00038932442624717165,
      "loss": 3.4486,
      "step": 32700
    },
    {
      "epoch": 3.524916585943386,
      "grad_norm": 0.6216268539428711,
      "learning_rate": 0.0003890011852171102,
      "loss": 3.4687,
      "step": 32750
    },
    {
      "epoch": 3.530298137982994,
      "grad_norm": 0.6284915208816528,
      "learning_rate": 0.0003886779441870488,
      "loss": 3.45,
      "step": 32800
    },
    {
      "epoch": 3.5356796900226026,
      "grad_norm": 0.7261960506439209,
      "learning_rate": 0.0003883547031569874,
      "loss": 3.461,
      "step": 32850
    },
    {
      "epoch": 3.5410612420622107,
      "grad_norm": 0.6670135855674744,
      "learning_rate": 0.0003880314621269259,
      "loss": 3.4522,
      "step": 32900
    },
    {
      "epoch": 3.546442794101819,
      "grad_norm": 0.6466349959373474,
      "learning_rate": 0.0003877082210968645,
      "loss": 3.4644,
      "step": 32950
    },
    {
      "epoch": 3.5518243461414274,
      "grad_norm": 0.5869767069816589,
      "learning_rate": 0.00038738498006680316,
      "loss": 3.4758,
      "step": 33000
    },
    {
      "epoch": 3.5518243461414274,
      "eval_accuracy": 0.37266495436958996,
      "eval_loss": 3.4755313396453857,
      "eval_runtime": 181.8939,
      "eval_samples_per_second": 99.019,
      "eval_steps_per_second": 6.19,
      "step": 33000
    },
    {
      "epoch": 3.5572058981810355,
      "grad_norm": 0.6126701831817627,
      "learning_rate": 0.0003870617390367417,
      "loss": 3.4703,
      "step": 33050
    },
    {
      "epoch": 3.5625874502206436,
      "grad_norm": 0.6517377495765686,
      "learning_rate": 0.0003867384980066803,
      "loss": 3.4547,
      "step": 33100
    },
    {
      "epoch": 3.5679690022602517,
      "grad_norm": 0.6763390898704529,
      "learning_rate": 0.00038641525697661884,
      "loss": 3.4589,
      "step": 33150
    },
    {
      "epoch": 3.57335055429986,
      "grad_norm": 0.6317027807235718,
      "learning_rate": 0.00038609201594655743,
      "loss": 3.4612,
      "step": 33200
    },
    {
      "epoch": 3.5787321063394684,
      "grad_norm": 0.6396004557609558,
      "learning_rate": 0.000385768774916496,
      "loss": 3.4317,
      "step": 33250
    },
    {
      "epoch": 3.5841136583790765,
      "grad_norm": 0.6065496802330017,
      "learning_rate": 0.0003854455338864346,
      "loss": 3.4426,
      "step": 33300
    },
    {
      "epoch": 3.5894952104186846,
      "grad_norm": 0.6124275326728821,
      "learning_rate": 0.0003851222928563732,
      "loss": 3.4479,
      "step": 33350
    },
    {
      "epoch": 3.594876762458293,
      "grad_norm": 0.6135236620903015,
      "learning_rate": 0.0003847990518263118,
      "loss": 3.468,
      "step": 33400
    },
    {
      "epoch": 3.6002583144979012,
      "grad_norm": 0.6805423498153687,
      "learning_rate": 0.00038447581079625035,
      "loss": 3.4838,
      "step": 33450
    },
    {
      "epoch": 3.6056398665375093,
      "grad_norm": 0.5888933539390564,
      "learning_rate": 0.00038415256976618895,
      "loss": 3.4487,
      "step": 33500
    },
    {
      "epoch": 3.611021418577118,
      "grad_norm": 0.6695815920829773,
      "learning_rate": 0.0003838293287361275,
      "loss": 3.4717,
      "step": 33550
    },
    {
      "epoch": 3.616402970616726,
      "grad_norm": 0.6879743337631226,
      "learning_rate": 0.00038350608770606613,
      "loss": 3.4645,
      "step": 33600
    },
    {
      "epoch": 3.621784522656334,
      "grad_norm": 0.549407958984375,
      "learning_rate": 0.00038318284667600473,
      "loss": 3.4642,
      "step": 33650
    },
    {
      "epoch": 3.627166074695942,
      "grad_norm": 0.5964043736457825,
      "learning_rate": 0.00038285960564594327,
      "loss": 3.4735,
      "step": 33700
    },
    {
      "epoch": 3.6325476267355503,
      "grad_norm": 0.6675652265548706,
      "learning_rate": 0.00038253636461588186,
      "loss": 3.4586,
      "step": 33750
    },
    {
      "epoch": 3.637929178775159,
      "grad_norm": 0.6164919137954712,
      "learning_rate": 0.00038221312358582046,
      "loss": 3.4589,
      "step": 33800
    },
    {
      "epoch": 3.643310730814767,
      "grad_norm": 0.6311922073364258,
      "learning_rate": 0.00038188988255575905,
      "loss": 3.4587,
      "step": 33850
    },
    {
      "epoch": 3.648692282854375,
      "grad_norm": 0.6482579708099365,
      "learning_rate": 0.00038156664152569765,
      "loss": 3.4482,
      "step": 33900
    },
    {
      "epoch": 3.6540738348939836,
      "grad_norm": 0.7032673358917236,
      "learning_rate": 0.00038124340049563624,
      "loss": 3.4705,
      "step": 33950
    },
    {
      "epoch": 3.6594553869335917,
      "grad_norm": 0.592106819152832,
      "learning_rate": 0.0003809201594655748,
      "loss": 3.455,
      "step": 34000
    },
    {
      "epoch": 3.6594553869335917,
      "eval_accuracy": 0.37426226047335065,
      "eval_loss": 3.4661221504211426,
      "eval_runtime": 181.561,
      "eval_samples_per_second": 99.201,
      "eval_steps_per_second": 6.202,
      "step": 34000
    },
    {
      "epoch": 3.6648369389732,
      "grad_norm": 0.5778534412384033,
      "learning_rate": 0.0003805969184355134,
      "loss": 3.4494,
      "step": 34050
    },
    {
      "epoch": 3.670218491012808,
      "grad_norm": 0.5945811867713928,
      "learning_rate": 0.0003802736774054519,
      "loss": 3.4473,
      "step": 34100
    },
    {
      "epoch": 3.675600043052416,
      "grad_norm": 0.6176246404647827,
      "learning_rate": 0.00037995043637539057,
      "loss": 3.455,
      "step": 34150
    },
    {
      "epoch": 3.6809815950920246,
      "grad_norm": 0.6745234727859497,
      "learning_rate": 0.00037962719534532916,
      "loss": 3.4559,
      "step": 34200
    },
    {
      "epoch": 3.6863631471316327,
      "grad_norm": 0.5692402124404907,
      "learning_rate": 0.0003793039543152677,
      "loss": 3.441,
      "step": 34250
    },
    {
      "epoch": 3.691744699171241,
      "grad_norm": 0.6138830184936523,
      "learning_rate": 0.0003789807132852063,
      "loss": 3.4481,
      "step": 34300
    },
    {
      "epoch": 3.6971262512108494,
      "grad_norm": 0.6347447037696838,
      "learning_rate": 0.0003786574722551449,
      "loss": 3.4508,
      "step": 34350
    },
    {
      "epoch": 3.7025078032504575,
      "grad_norm": 0.6304515600204468,
      "learning_rate": 0.00037833423122508343,
      "loss": 3.4593,
      "step": 34400
    },
    {
      "epoch": 3.7078893552900656,
      "grad_norm": 0.6190226078033447,
      "learning_rate": 0.0003780109901950221,
      "loss": 3.4571,
      "step": 34450
    },
    {
      "epoch": 3.713270907329674,
      "grad_norm": 0.6307307481765747,
      "learning_rate": 0.0003776877491649607,
      "loss": 3.4505,
      "step": 34500
    },
    {
      "epoch": 3.7186524593692822,
      "grad_norm": 0.6326085925102234,
      "learning_rate": 0.0003773645081348992,
      "loss": 3.4599,
      "step": 34550
    },
    {
      "epoch": 3.7240340114088903,
      "grad_norm": 0.6394867897033691,
      "learning_rate": 0.0003770412671048378,
      "loss": 3.4568,
      "step": 34600
    },
    {
      "epoch": 3.7294155634484984,
      "grad_norm": 0.633033037185669,
      "learning_rate": 0.00037671802607477635,
      "loss": 3.4429,
      "step": 34650
    },
    {
      "epoch": 3.7347971154881066,
      "grad_norm": 0.6172804236412048,
      "learning_rate": 0.000376394785044715,
      "loss": 3.4633,
      "step": 34700
    },
    {
      "epoch": 3.740178667527715,
      "grad_norm": 0.6102865934371948,
      "learning_rate": 0.0003760715440146536,
      "loss": 3.4695,
      "step": 34750
    },
    {
      "epoch": 3.745560219567323,
      "grad_norm": 0.6239819526672363,
      "learning_rate": 0.00037574830298459214,
      "loss": 3.4567,
      "step": 34800
    },
    {
      "epoch": 3.7509417716069313,
      "grad_norm": 0.6692067980766296,
      "learning_rate": 0.00037542506195453073,
      "loss": 3.4591,
      "step": 34850
    },
    {
      "epoch": 3.75632332364654,
      "grad_norm": 0.589547336101532,
      "learning_rate": 0.0003751018209244693,
      "loss": 3.4723,
      "step": 34900
    },
    {
      "epoch": 3.761704875686148,
      "grad_norm": 0.6273283958435059,
      "learning_rate": 0.00037477857989440787,
      "loss": 3.4668,
      "step": 34950
    },
    {
      "epoch": 3.767086427725756,
      "grad_norm": 0.6687270402908325,
      "learning_rate": 0.00037446180368494767,
      "loss": 3.4728,
      "step": 35000
    },
    {
      "epoch": 3.767086427725756,
      "eval_accuracy": 0.3749564437735974,
      "eval_loss": 3.4588773250579834,
      "eval_runtime": 181.7474,
      "eval_samples_per_second": 99.099,
      "eval_steps_per_second": 6.195,
      "step": 35000
    },
    {
      "epoch": 3.772467979765364,
      "grad_norm": 0.657584547996521,
      "learning_rate": 0.00037413856265488627,
      "loss": 3.4584,
      "step": 35050
    },
    {
      "epoch": 3.7778495318049723,
      "grad_norm": 0.6070948839187622,
      "learning_rate": 0.0003738153216248249,
      "loss": 3.4495,
      "step": 35100
    },
    {
      "epoch": 3.783231083844581,
      "grad_norm": 0.6411889791488647,
      "learning_rate": 0.00037349208059476346,
      "loss": 3.4406,
      "step": 35150
    },
    {
      "epoch": 3.788612635884189,
      "grad_norm": 0.6415666341781616,
      "learning_rate": 0.00037316883956470205,
      "loss": 3.4803,
      "step": 35200
    },
    {
      "epoch": 3.793994187923797,
      "grad_norm": 0.6426709890365601,
      "learning_rate": 0.00037284559853464064,
      "loss": 3.4618,
      "step": 35250
    },
    {
      "epoch": 3.7993757399634056,
      "grad_norm": 0.6597058176994324,
      "learning_rate": 0.0003725223575045792,
      "loss": 3.4453,
      "step": 35300
    },
    {
      "epoch": 3.8047572920030137,
      "grad_norm": 0.6358980536460876,
      "learning_rate": 0.0003721991164745178,
      "loss": 3.4715,
      "step": 35350
    },
    {
      "epoch": 3.810138844042622,
      "grad_norm": 0.6410354375839233,
      "learning_rate": 0.00037187587544445643,
      "loss": 3.4647,
      "step": 35400
    },
    {
      "epoch": 3.8155203960822304,
      "grad_norm": 0.6233660578727722,
      "learning_rate": 0.00037155263441439497,
      "loss": 3.4698,
      "step": 35450
    },
    {
      "epoch": 3.8209019481218385,
      "grad_norm": 0.628305196762085,
      "learning_rate": 0.00037122939338433356,
      "loss": 3.4537,
      "step": 35500
    },
    {
      "epoch": 3.8262835001614466,
      "grad_norm": 0.6736071705818176,
      "learning_rate": 0.0003709061523542721,
      "loss": 3.4576,
      "step": 35550
    },
    {
      "epoch": 3.8316650522010547,
      "grad_norm": 0.6532096266746521,
      "learning_rate": 0.0003705829113242107,
      "loss": 3.4494,
      "step": 35600
    },
    {
      "epoch": 3.837046604240663,
      "grad_norm": 0.5983206629753113,
      "learning_rate": 0.00037025967029414935,
      "loss": 3.4538,
      "step": 35650
    },
    {
      "epoch": 3.8424281562802713,
      "grad_norm": 0.6398340463638306,
      "learning_rate": 0.0003699364292640879,
      "loss": 3.4738,
      "step": 35700
    },
    {
      "epoch": 3.8478097083198795,
      "grad_norm": 0.5982816219329834,
      "learning_rate": 0.0003696131882340265,
      "loss": 3.4687,
      "step": 35750
    },
    {
      "epoch": 3.8531912603594876,
      "grad_norm": 0.6313088536262512,
      "learning_rate": 0.0003692899472039651,
      "loss": 3.4527,
      "step": 35800
    },
    {
      "epoch": 3.858572812399096,
      "grad_norm": 0.6521557569503784,
      "learning_rate": 0.0003689667061739036,
      "loss": 3.4547,
      "step": 35850
    },
    {
      "epoch": 3.863954364438704,
      "grad_norm": 0.6223726272583008,
      "learning_rate": 0.0003686434651438422,
      "loss": 3.4714,
      "step": 35900
    },
    {
      "epoch": 3.8693359164783123,
      "grad_norm": 0.6917114853858948,
      "learning_rate": 0.00036832022411378086,
      "loss": 3.4696,
      "step": 35950
    },
    {
      "epoch": 3.8747174685179204,
      "grad_norm": 0.606867253780365,
      "learning_rate": 0.0003679969830837194,
      "loss": 3.4347,
      "step": 36000
    },
    {
      "epoch": 3.8747174685179204,
      "eval_accuracy": 0.3753462903350757,
      "eval_loss": 3.4530980587005615,
      "eval_runtime": 181.4546,
      "eval_samples_per_second": 99.259,
      "eval_steps_per_second": 6.205,
      "step": 36000
    },
    {
      "epoch": 3.8800990205575285,
      "grad_norm": 0.6278716325759888,
      "learning_rate": 0.000367673742053658,
      "loss": 3.4351,
      "step": 36050
    },
    {
      "epoch": 3.885480572597137,
      "grad_norm": 0.625916600227356,
      "learning_rate": 0.00036735050102359654,
      "loss": 3.4415,
      "step": 36100
    },
    {
      "epoch": 3.890862124636745,
      "grad_norm": 0.6142260432243347,
      "learning_rate": 0.00036702725999353513,
      "loss": 3.4539,
      "step": 36150
    },
    {
      "epoch": 3.8962436766763533,
      "grad_norm": 0.6766964793205261,
      "learning_rate": 0.0003667040189634737,
      "loss": 3.4584,
      "step": 36200
    },
    {
      "epoch": 3.901625228715962,
      "grad_norm": 0.6694559454917908,
      "learning_rate": 0.0003663807779334123,
      "loss": 3.4564,
      "step": 36250
    },
    {
      "epoch": 3.90700678075557,
      "grad_norm": 0.6292983889579773,
      "learning_rate": 0.0003660575369033509,
      "loss": 3.4572,
      "step": 36300
    },
    {
      "epoch": 3.912388332795178,
      "grad_norm": 0.6223490834236145,
      "learning_rate": 0.0003657342958732895,
      "loss": 3.4521,
      "step": 36350
    },
    {
      "epoch": 3.9177698848347866,
      "grad_norm": 0.6118103861808777,
      "learning_rate": 0.00036541105484322805,
      "loss": 3.4535,
      "step": 36400
    },
    {
      "epoch": 3.9231514368743947,
      "grad_norm": 0.645898163318634,
      "learning_rate": 0.00036508781381316665,
      "loss": 3.4518,
      "step": 36450
    },
    {
      "epoch": 3.928532988914003,
      "grad_norm": 0.6415484547615051,
      "learning_rate": 0.0003647645727831053,
      "loss": 3.4585,
      "step": 36500
    },
    {
      "epoch": 3.933914540953611,
      "grad_norm": 0.6804546117782593,
      "learning_rate": 0.00036444133175304384,
      "loss": 3.4533,
      "step": 36550
    },
    {
      "epoch": 3.939296092993219,
      "grad_norm": 0.6136488318443298,
      "learning_rate": 0.00036411809072298243,
      "loss": 3.4474,
      "step": 36600
    },
    {
      "epoch": 3.9446776450328276,
      "grad_norm": 0.6238126158714294,
      "learning_rate": 0.00036379484969292097,
      "loss": 3.458,
      "step": 36650
    },
    {
      "epoch": 3.9500591970724357,
      "grad_norm": 0.6238007545471191,
      "learning_rate": 0.00036347160866285956,
      "loss": 3.4624,
      "step": 36700
    },
    {
      "epoch": 3.955440749112044,
      "grad_norm": 0.6192507743835449,
      "learning_rate": 0.0003631483676327981,
      "loss": 3.4544,
      "step": 36750
    },
    {
      "epoch": 3.9608223011516523,
      "grad_norm": 0.6187044978141785,
      "learning_rate": 0.00036282512660273675,
      "loss": 3.4372,
      "step": 36800
    },
    {
      "epoch": 3.9662038531912605,
      "grad_norm": 0.6006636619567871,
      "learning_rate": 0.00036250188557267535,
      "loss": 3.4588,
      "step": 36850
    },
    {
      "epoch": 3.9715854052308686,
      "grad_norm": 0.6458033323287964,
      "learning_rate": 0.0003621786445426139,
      "loss": 3.4481,
      "step": 36900
    },
    {
      "epoch": 3.9769669572704767,
      "grad_norm": 0.6214568614959717,
      "learning_rate": 0.0003618554035125525,
      "loss": 3.4495,
      "step": 36950
    },
    {
      "epoch": 3.9823485093100848,
      "grad_norm": 0.6304638981819153,
      "learning_rate": 0.0003615321624824911,
      "loss": 3.46,
      "step": 37000
    },
    {
      "epoch": 3.9823485093100848,
      "eval_accuracy": 0.37597354345642875,
      "eval_loss": 3.446321725845337,
      "eval_runtime": 181.7245,
      "eval_samples_per_second": 99.112,
      "eval_steps_per_second": 6.196,
      "step": 37000
    },
    {
      "epoch": 3.9877300613496933,
      "grad_norm": 0.5896131992340088,
      "learning_rate": 0.0003612089214524296,
      "loss": 3.4304,
      "step": 37050
    },
    {
      "epoch": 3.9931116133893014,
      "grad_norm": 0.6256299614906311,
      "learning_rate": 0.00036088568042236827,
      "loss": 3.4539,
      "step": 37100
    },
    {
      "epoch": 3.9984931654289095,
      "grad_norm": 0.6513100862503052,
      "learning_rate": 0.00036056243939230686,
      "loss": 3.4543,
      "step": 37150
    },
    {
      "epoch": 4.003874717468518,
      "grad_norm": 0.6809729337692261,
      "learning_rate": 0.0003602391983622454,
      "loss": 3.382,
      "step": 37200
    },
    {
      "epoch": 4.009256269508126,
      "grad_norm": 0.6216638088226318,
      "learning_rate": 0.000359915957332184,
      "loss": 3.3691,
      "step": 37250
    },
    {
      "epoch": 4.014637821547734,
      "grad_norm": 0.687582790851593,
      "learning_rate": 0.00035959271630212254,
      "loss": 3.3739,
      "step": 37300
    },
    {
      "epoch": 4.020019373587343,
      "grad_norm": 0.6538896560668945,
      "learning_rate": 0.00035926947527206113,
      "loss": 3.3553,
      "step": 37350
    },
    {
      "epoch": 4.0254009256269505,
      "grad_norm": 0.6100491881370544,
      "learning_rate": 0.0003589462342419998,
      "loss": 3.3636,
      "step": 37400
    },
    {
      "epoch": 4.030782477666559,
      "grad_norm": 0.6782578825950623,
      "learning_rate": 0.0003586229932119383,
      "loss": 3.3712,
      "step": 37450
    },
    {
      "epoch": 4.036164029706168,
      "grad_norm": 0.7136054039001465,
      "learning_rate": 0.0003582997521818769,
      "loss": 3.3502,
      "step": 37500
    },
    {
      "epoch": 4.041545581745775,
      "grad_norm": 1.0063740015029907,
      "learning_rate": 0.0003579765111518155,
      "loss": 3.3518,
      "step": 37550
    },
    {
      "epoch": 4.046927133785384,
      "grad_norm": 0.6431665420532227,
      "learning_rate": 0.00035765327012175405,
      "loss": 3.3673,
      "step": 37600
    },
    {
      "epoch": 4.0523086858249915,
      "grad_norm": 0.6726707220077515,
      "learning_rate": 0.0003573300290916927,
      "loss": 3.365,
      "step": 37650
    },
    {
      "epoch": 4.0576902378646,
      "grad_norm": 0.6649792194366455,
      "learning_rate": 0.0003570067880616313,
      "loss": 3.373,
      "step": 37700
    },
    {
      "epoch": 4.063071789904209,
      "grad_norm": 0.6573074460029602,
      "learning_rate": 0.00035668354703156984,
      "loss": 3.3614,
      "step": 37750
    },
    {
      "epoch": 4.068453341943816,
      "grad_norm": 0.6372230052947998,
      "learning_rate": 0.00035636030600150843,
      "loss": 3.3596,
      "step": 37800
    },
    {
      "epoch": 4.073834893983425,
      "grad_norm": 0.6461084485054016,
      "learning_rate": 0.00035603706497144697,
      "loss": 3.3685,
      "step": 37850
    },
    {
      "epoch": 4.079216446023033,
      "grad_norm": 0.6296938061714172,
      "learning_rate": 0.00035571382394138557,
      "loss": 3.3641,
      "step": 37900
    },
    {
      "epoch": 4.084597998062641,
      "grad_norm": 0.658966600894928,
      "learning_rate": 0.0003553905829113242,
      "loss": 3.3773,
      "step": 37950
    },
    {
      "epoch": 4.08997955010225,
      "grad_norm": 0.6599716544151306,
      "learning_rate": 0.00035506734188126275,
      "loss": 3.3683,
      "step": 38000
    },
    {
      "epoch": 4.08997955010225,
      "eval_accuracy": 0.3761422813912046,
      "eval_loss": 3.44930100440979,
      "eval_runtime": 181.3188,
      "eval_samples_per_second": 99.333,
      "eval_steps_per_second": 6.21,
      "step": 38000
    },
    {
      "epoch": 4.095361102141858,
      "grad_norm": 0.6307918429374695,
      "learning_rate": 0.00035474410085120135,
      "loss": 3.3854,
      "step": 38050
    },
    {
      "epoch": 4.100742654181466,
      "grad_norm": 0.6225454211235046,
      "learning_rate": 0.00035442085982113994,
      "loss": 3.3698,
      "step": 38100
    },
    {
      "epoch": 4.106124206221074,
      "grad_norm": 0.6287394762039185,
      "learning_rate": 0.0003540976187910785,
      "loss": 3.3919,
      "step": 38150
    },
    {
      "epoch": 4.111505758260682,
      "grad_norm": 0.6210476756095886,
      "learning_rate": 0.0003537743777610171,
      "loss": 3.3621,
      "step": 38200
    },
    {
      "epoch": 4.1168873103002905,
      "grad_norm": 0.6711219549179077,
      "learning_rate": 0.00035345113673095573,
      "loss": 3.3686,
      "step": 38250
    },
    {
      "epoch": 4.122268862339899,
      "grad_norm": 0.6280274391174316,
      "learning_rate": 0.00035312789570089427,
      "loss": 3.3742,
      "step": 38300
    },
    {
      "epoch": 4.127650414379507,
      "grad_norm": 0.6348408460617065,
      "learning_rate": 0.00035280465467083286,
      "loss": 3.3871,
      "step": 38350
    },
    {
      "epoch": 4.133031966419115,
      "grad_norm": 0.6895505785942078,
      "learning_rate": 0.0003524814136407714,
      "loss": 3.3586,
      "step": 38400
    },
    {
      "epoch": 4.138413518458724,
      "grad_norm": 0.7086821794509888,
      "learning_rate": 0.00035215817261071,
      "loss": 3.3695,
      "step": 38450
    },
    {
      "epoch": 4.1437950704983315,
      "grad_norm": 0.6616617441177368,
      "learning_rate": 0.00035183493158064865,
      "loss": 3.382,
      "step": 38500
    },
    {
      "epoch": 4.14917662253794,
      "grad_norm": 0.6920830607414246,
      "learning_rate": 0.0003515116905505872,
      "loss": 3.4122,
      "step": 38550
    },
    {
      "epoch": 4.154558174577549,
      "grad_norm": 0.7117165923118591,
      "learning_rate": 0.0003511884495205258,
      "loss": 3.3802,
      "step": 38600
    },
    {
      "epoch": 4.159939726617156,
      "grad_norm": 0.6522443890571594,
      "learning_rate": 0.0003508652084904644,
      "loss": 3.3718,
      "step": 38650
    },
    {
      "epoch": 4.165321278656765,
      "grad_norm": 0.7154022455215454,
      "learning_rate": 0.0003505419674604029,
      "loss": 3.3929,
      "step": 38700
    },
    {
      "epoch": 4.1707028306963725,
      "grad_norm": 0.6403151750564575,
      "learning_rate": 0.0003502187264303415,
      "loss": 3.3938,
      "step": 38750
    },
    {
      "epoch": 4.176084382735981,
      "grad_norm": 0.5948874354362488,
      "learning_rate": 0.00034989548540028016,
      "loss": 3.3894,
      "step": 38800
    },
    {
      "epoch": 4.18146593477559,
      "grad_norm": 0.6313942074775696,
      "learning_rate": 0.0003495722443702187,
      "loss": 3.384,
      "step": 38850
    },
    {
      "epoch": 4.186847486815197,
      "grad_norm": 0.6668844223022461,
      "learning_rate": 0.0003492490033401573,
      "loss": 3.3911,
      "step": 38900
    },
    {
      "epoch": 4.192229038854806,
      "grad_norm": 0.6406861543655396,
      "learning_rate": 0.00034892576231009584,
      "loss": 3.3706,
      "step": 38950
    },
    {
      "epoch": 4.197610590894414,
      "grad_norm": 0.6086711883544922,
      "learning_rate": 0.00034860252128003443,
      "loss": 3.3682,
      "step": 39000
    },
    {
      "epoch": 4.197610590894414,
      "eval_accuracy": 0.3768287503364165,
      "eval_loss": 3.4479315280914307,
      "eval_runtime": 181.8639,
      "eval_samples_per_second": 99.036,
      "eval_steps_per_second": 6.191,
      "step": 39000
    },
    {
      "epoch": 4.202992142934022,
      "grad_norm": 0.6679412722587585,
      "learning_rate": 0.000348279280249973,
      "loss": 3.3828,
      "step": 39050
    },
    {
      "epoch": 4.208373694973631,
      "grad_norm": 0.6079999804496765,
      "learning_rate": 0.0003479560392199116,
      "loss": 3.3802,
      "step": 39100
    },
    {
      "epoch": 4.213755247013238,
      "grad_norm": 0.629065215587616,
      "learning_rate": 0.00034763926301045137,
      "loss": 3.3894,
      "step": 39150
    },
    {
      "epoch": 4.219136799052847,
      "grad_norm": 0.6328132152557373,
      "learning_rate": 0.00034731602198039,
      "loss": 3.3766,
      "step": 39200
    },
    {
      "epoch": 4.224518351092455,
      "grad_norm": 0.6607147455215454,
      "learning_rate": 0.0003469927809503286,
      "loss": 3.3742,
      "step": 39250
    },
    {
      "epoch": 4.229899903132063,
      "grad_norm": 0.6375463008880615,
      "learning_rate": 0.00034666953992026716,
      "loss": 3.3871,
      "step": 39300
    },
    {
      "epoch": 4.2352814551716715,
      "grad_norm": 0.6818897724151611,
      "learning_rate": 0.00034634629889020575,
      "loss": 3.3868,
      "step": 39350
    },
    {
      "epoch": 4.24066300721128,
      "grad_norm": 0.6507583856582642,
      "learning_rate": 0.00034602305786014435,
      "loss": 3.3852,
      "step": 39400
    },
    {
      "epoch": 4.246044559250888,
      "grad_norm": 0.649346113204956,
      "learning_rate": 0.00034569981683008294,
      "loss": 3.3816,
      "step": 39450
    },
    {
      "epoch": 4.251426111290496,
      "grad_norm": 0.6513179540634155,
      "learning_rate": 0.00034537657580002154,
      "loss": 3.3709,
      "step": 39500
    },
    {
      "epoch": 4.256807663330104,
      "grad_norm": 0.7123865485191345,
      "learning_rate": 0.00034505333476996013,
      "loss": 3.3986,
      "step": 39550
    },
    {
      "epoch": 4.2621892153697125,
      "grad_norm": 0.7108580470085144,
      "learning_rate": 0.00034473009373989867,
      "loss": 3.3923,
      "step": 39600
    },
    {
      "epoch": 4.267570767409321,
      "grad_norm": 0.6283742785453796,
      "learning_rate": 0.00034440685270983727,
      "loss": 3.401,
      "step": 39650
    },
    {
      "epoch": 4.272952319448929,
      "grad_norm": 0.6473816633224487,
      "learning_rate": 0.0003440836116797758,
      "loss": 3.391,
      "step": 39700
    },
    {
      "epoch": 4.278333871488537,
      "grad_norm": 0.6747614145278931,
      "learning_rate": 0.00034376037064971445,
      "loss": 3.401,
      "step": 39750
    },
    {
      "epoch": 4.283715423528146,
      "grad_norm": 0.6350759863853455,
      "learning_rate": 0.00034343712961965305,
      "loss": 3.4075,
      "step": 39800
    },
    {
      "epoch": 4.2890969755677535,
      "grad_norm": 0.7255776524543762,
      "learning_rate": 0.0003431138885895916,
      "loss": 3.3835,
      "step": 39850
    },
    {
      "epoch": 4.294478527607362,
      "grad_norm": 0.7065731287002563,
      "learning_rate": 0.00034279711238013145,
      "loss": 3.396,
      "step": 39900
    },
    {
      "epoch": 4.299860079646971,
      "grad_norm": 0.67261803150177,
      "learning_rate": 0.00034247387135007,
      "loss": 3.3983,
      "step": 39950
    },
    {
      "epoch": 4.305241631686578,
      "grad_norm": 0.6579726934432983,
      "learning_rate": 0.0003421506303200086,
      "loss": 3.3876,
      "step": 40000
    },
    {
      "epoch": 4.305241631686578,
      "eval_accuracy": 0.3775528131808122,
      "eval_loss": 3.441394805908203,
      "eval_runtime": 181.6119,
      "eval_samples_per_second": 99.173,
      "eval_steps_per_second": 6.2,
      "step": 40000
    },
    {
      "epoch": 4.310623183726187,
      "grad_norm": 0.6545993089675903,
      "learning_rate": 0.0003418273892899471,
      "loss": 3.382,
      "step": 40050
    },
    {
      "epoch": 4.3160047357657945,
      "grad_norm": 0.5974786877632141,
      "learning_rate": 0.0003415041482598858,
      "loss": 3.3722,
      "step": 40100
    },
    {
      "epoch": 4.321386287805403,
      "grad_norm": 0.6922400593757629,
      "learning_rate": 0.00034118090722982437,
      "loss": 3.3802,
      "step": 40150
    },
    {
      "epoch": 4.326767839845012,
      "grad_norm": 0.6838219165802002,
      "learning_rate": 0.0003408576661997629,
      "loss": 3.39,
      "step": 40200
    },
    {
      "epoch": 4.332149391884619,
      "grad_norm": 0.6342960000038147,
      "learning_rate": 0.0003405344251697015,
      "loss": 3.3819,
      "step": 40250
    },
    {
      "epoch": 4.337530943924228,
      "grad_norm": 0.7120860815048218,
      "learning_rate": 0.0003402111841396401,
      "loss": 3.374,
      "step": 40300
    },
    {
      "epoch": 4.342912495963836,
      "grad_norm": 0.648949921131134,
      "learning_rate": 0.00033988794310957864,
      "loss": 3.3776,
      "step": 40350
    },
    {
      "epoch": 4.348294048003444,
      "grad_norm": 0.7276732325553894,
      "learning_rate": 0.0003395647020795173,
      "loss": 3.4097,
      "step": 40400
    },
    {
      "epoch": 4.3536756000430525,
      "grad_norm": 0.6344953179359436,
      "learning_rate": 0.0003392414610494559,
      "loss": 3.3754,
      "step": 40450
    },
    {
      "epoch": 4.359057152082661,
      "grad_norm": 0.6971585154533386,
      "learning_rate": 0.0003389182200193944,
      "loss": 3.4039,
      "step": 40500
    },
    {
      "epoch": 4.364438704122269,
      "grad_norm": 0.6294052600860596,
      "learning_rate": 0.000338594978989333,
      "loss": 3.3791,
      "step": 40550
    },
    {
      "epoch": 4.369820256161877,
      "grad_norm": 0.6447839140892029,
      "learning_rate": 0.00033827173795927156,
      "loss": 3.3871,
      "step": 40600
    },
    {
      "epoch": 4.375201808201485,
      "grad_norm": 0.6472920179367065,
      "learning_rate": 0.00033794849692921015,
      "loss": 3.3886,
      "step": 40650
    },
    {
      "epoch": 4.3805833602410935,
      "grad_norm": 0.6552985310554504,
      "learning_rate": 0.0003376252558991488,
      "loss": 3.3904,
      "step": 40700
    },
    {
      "epoch": 4.385964912280702,
      "grad_norm": 0.6818690896034241,
      "learning_rate": 0.00033730201486908734,
      "loss": 3.3947,
      "step": 40750
    },
    {
      "epoch": 4.39134646432031,
      "grad_norm": 0.646091639995575,
      "learning_rate": 0.00033697877383902594,
      "loss": 3.4024,
      "step": 40800
    },
    {
      "epoch": 4.396728016359918,
      "grad_norm": 0.6999764442443848,
      "learning_rate": 0.00033665553280896453,
      "loss": 3.391,
      "step": 40850
    },
    {
      "epoch": 4.402109568399527,
      "grad_norm": 0.6619383096694946,
      "learning_rate": 0.00033633229177890307,
      "loss": 3.4125,
      "step": 40900
    },
    {
      "epoch": 4.4074911204391345,
      "grad_norm": 0.8046988844871521,
      "learning_rate": 0.00033600905074884167,
      "loss": 3.373,
      "step": 40950
    },
    {
      "epoch": 4.412872672478743,
      "grad_norm": 0.692973256111145,
      "learning_rate": 0.0003356858097187803,
      "loss": 3.3813,
      "step": 41000
    },
    {
      "epoch": 4.412872672478743,
      "eval_accuracy": 0.37769666960427856,
      "eval_loss": 3.4349679946899414,
      "eval_runtime": 181.1247,
      "eval_samples_per_second": 99.44,
      "eval_steps_per_second": 6.217,
      "step": 41000
    },
    {
      "epoch": 4.418254224518351,
      "grad_norm": 0.6343230605125427,
      "learning_rate": 0.00033536256868871886,
      "loss": 3.3868,
      "step": 41050
    },
    {
      "epoch": 4.423635776557959,
      "grad_norm": 0.6828956604003906,
      "learning_rate": 0.00033503932765865745,
      "loss": 3.3779,
      "step": 41100
    },
    {
      "epoch": 4.429017328597568,
      "grad_norm": 0.6604346632957458,
      "learning_rate": 0.000334716086628596,
      "loss": 3.3862,
      "step": 41150
    },
    {
      "epoch": 4.4343988806371755,
      "grad_norm": 0.7162014842033386,
      "learning_rate": 0.0003343928455985346,
      "loss": 3.4053,
      "step": 41200
    },
    {
      "epoch": 4.439780432676784,
      "grad_norm": 0.7162100672721863,
      "learning_rate": 0.00033406960456847324,
      "loss": 3.397,
      "step": 41250
    },
    {
      "epoch": 4.445161984716393,
      "grad_norm": 0.6890962719917297,
      "learning_rate": 0.0003337463635384118,
      "loss": 3.4001,
      "step": 41300
    },
    {
      "epoch": 4.450543536756,
      "grad_norm": 0.714141309261322,
      "learning_rate": 0.00033342312250835037,
      "loss": 3.3843,
      "step": 41350
    },
    {
      "epoch": 4.455925088795609,
      "grad_norm": 0.6847618222236633,
      "learning_rate": 0.00033309988147828896,
      "loss": 3.4102,
      "step": 41400
    },
    {
      "epoch": 4.461306640835216,
      "grad_norm": 0.6703453063964844,
      "learning_rate": 0.0003327766404482275,
      "loss": 3.4016,
      "step": 41450
    },
    {
      "epoch": 4.466688192874825,
      "grad_norm": 0.6432178616523743,
      "learning_rate": 0.0003324533994181661,
      "loss": 3.3857,
      "step": 41500
    },
    {
      "epoch": 4.4720697449144335,
      "grad_norm": 0.6851868033409119,
      "learning_rate": 0.00033213015838810475,
      "loss": 3.4059,
      "step": 41550
    },
    {
      "epoch": 4.477451296954041,
      "grad_norm": 0.6780423521995544,
      "learning_rate": 0.0003318069173580433,
      "loss": 3.3957,
      "step": 41600
    },
    {
      "epoch": 4.48283284899365,
      "grad_norm": 0.6518639922142029,
      "learning_rate": 0.0003314836763279819,
      "loss": 3.3808,
      "step": 41650
    },
    {
      "epoch": 4.488214401033258,
      "grad_norm": 0.630031406879425,
      "learning_rate": 0.0003311604352979204,
      "loss": 3.3863,
      "step": 41700
    },
    {
      "epoch": 4.493595953072866,
      "grad_norm": 0.6582158207893372,
      "learning_rate": 0.000330837194267859,
      "loss": 3.3722,
      "step": 41750
    },
    {
      "epoch": 4.4989775051124745,
      "grad_norm": 0.7116129994392395,
      "learning_rate": 0.0003305139532377976,
      "loss": 3.389,
      "step": 41800
    },
    {
      "epoch": 4.504359057152083,
      "grad_norm": 0.6760125756263733,
      "learning_rate": 0.0003301907122077362,
      "loss": 3.396,
      "step": 41850
    },
    {
      "epoch": 4.509740609191691,
      "grad_norm": 0.6478492617607117,
      "learning_rate": 0.0003298674711776748,
      "loss": 3.401,
      "step": 41900
    },
    {
      "epoch": 4.515122161231299,
      "grad_norm": 0.6583666205406189,
      "learning_rate": 0.0003295442301476134,
      "loss": 3.3866,
      "step": 41950
    },
    {
      "epoch": 4.520503713270907,
      "grad_norm": 0.6689998507499695,
      "learning_rate": 0.00032922098911755194,
      "loss": 3.3924,
      "step": 42000
    },
    {
      "epoch": 4.520503713270907,
      "eval_accuracy": 0.37855687451710585,
      "eval_loss": 3.431107521057129,
      "eval_runtime": 181.9242,
      "eval_samples_per_second": 99.003,
      "eval_steps_per_second": 6.189,
      "step": 42000
    },
    {
      "epoch": 4.5258852653105155,
      "grad_norm": 0.655572235584259,
      "learning_rate": 0.00032889774808749053,
      "loss": 3.407,
      "step": 42050
    },
    {
      "epoch": 4.531266817350124,
      "grad_norm": 0.6405171155929565,
      "learning_rate": 0.0003285745070574292,
      "loss": 3.4029,
      "step": 42100
    },
    {
      "epoch": 4.536648369389732,
      "grad_norm": 0.6641229391098022,
      "learning_rate": 0.0003282512660273677,
      "loss": 3.3999,
      "step": 42150
    },
    {
      "epoch": 4.54202992142934,
      "grad_norm": 0.6896253824234009,
      "learning_rate": 0.0003279280249973063,
      "loss": 3.4085,
      "step": 42200
    },
    {
      "epoch": 4.547411473468949,
      "grad_norm": 0.6799988746643066,
      "learning_rate": 0.00032760478396724486,
      "loss": 3.4027,
      "step": 42250
    },
    {
      "epoch": 4.5527930255085565,
      "grad_norm": 0.6822940111160278,
      "learning_rate": 0.00032728154293718345,
      "loss": 3.3753,
      "step": 42300
    },
    {
      "epoch": 4.558174577548165,
      "grad_norm": 0.656019389629364,
      "learning_rate": 0.00032695830190712205,
      "loss": 3.3807,
      "step": 42350
    },
    {
      "epoch": 4.563556129587774,
      "grad_norm": 0.650692880153656,
      "learning_rate": 0.00032663506087706064,
      "loss": 3.4062,
      "step": 42400
    },
    {
      "epoch": 4.568937681627381,
      "grad_norm": 0.6641072630882263,
      "learning_rate": 0.00032631181984699924,
      "loss": 3.4034,
      "step": 42450
    },
    {
      "epoch": 4.57431923366699,
      "grad_norm": 0.6998100280761719,
      "learning_rate": 0.00032598857881693783,
      "loss": 3.394,
      "step": 42500
    },
    {
      "epoch": 4.579700785706597,
      "grad_norm": 0.7412688732147217,
      "learning_rate": 0.00032566533778687637,
      "loss": 3.3865,
      "step": 42550
    },
    {
      "epoch": 4.585082337746206,
      "grad_norm": 0.6878949999809265,
      "learning_rate": 0.00032534209675681497,
      "loss": 3.3875,
      "step": 42600
    },
    {
      "epoch": 4.5904638897858145,
      "grad_norm": 0.718602180480957,
      "learning_rate": 0.0003250188557267535,
      "loss": 3.4083,
      "step": 42650
    },
    {
      "epoch": 4.595845441825422,
      "grad_norm": 0.6992116570472717,
      "learning_rate": 0.00032469561469669215,
      "loss": 3.3877,
      "step": 42700
    },
    {
      "epoch": 4.601226993865031,
      "grad_norm": 0.6840443015098572,
      "learning_rate": 0.00032437237366663075,
      "loss": 3.3814,
      "step": 42750
    },
    {
      "epoch": 4.606608545904638,
      "grad_norm": 0.6726332902908325,
      "learning_rate": 0.0003240491326365693,
      "loss": 3.4107,
      "step": 42800
    },
    {
      "epoch": 4.611990097944247,
      "grad_norm": 0.7318175435066223,
      "learning_rate": 0.0003237258916065079,
      "loss": 3.3882,
      "step": 42850
    },
    {
      "epoch": 4.6173716499838555,
      "grad_norm": 0.6592293977737427,
      "learning_rate": 0.0003234026505764465,
      "loss": 3.4084,
      "step": 42900
    },
    {
      "epoch": 4.622753202023463,
      "grad_norm": 0.70222407579422,
      "learning_rate": 0.000323079409546385,
      "loss": 3.3907,
      "step": 42950
    },
    {
      "epoch": 4.628134754063072,
      "grad_norm": 0.6887460350990295,
      "learning_rate": 0.0003227626333369248,
      "loss": 3.3912,
      "step": 43000
    },
    {
      "epoch": 4.628134754063072,
      "eval_accuracy": 0.3786048990935198,
      "eval_loss": 3.426353693008423,
      "eval_runtime": 183.3187,
      "eval_samples_per_second": 98.25,
      "eval_steps_per_second": 6.142,
      "step": 43000
    },
    {
      "epoch": 4.63351630610268,
      "grad_norm": 0.6995670795440674,
      "learning_rate": 0.0003224393923068635,
      "loss": 3.3847,
      "step": 43050
    },
    {
      "epoch": 4.638897858142288,
      "grad_norm": 0.7121273875236511,
      "learning_rate": 0.00032211615127680207,
      "loss": 3.3972,
      "step": 43100
    },
    {
      "epoch": 4.6442794101818965,
      "grad_norm": 0.6917558908462524,
      "learning_rate": 0.0003217929102467406,
      "loss": 3.4062,
      "step": 43150
    },
    {
      "epoch": 4.649660962221505,
      "grad_norm": 0.7152304649353027,
      "learning_rate": 0.0003214696692166792,
      "loss": 3.3848,
      "step": 43200
    },
    {
      "epoch": 4.655042514261113,
      "grad_norm": 0.6601977944374084,
      "learning_rate": 0.0003211464281866178,
      "loss": 3.3972,
      "step": 43250
    },
    {
      "epoch": 4.660424066300721,
      "grad_norm": 0.6530590057373047,
      "learning_rate": 0.00032082318715655634,
      "loss": 3.388,
      "step": 43300
    },
    {
      "epoch": 4.665805618340329,
      "grad_norm": 0.6723693013191223,
      "learning_rate": 0.000320499946126495,
      "loss": 3.3952,
      "step": 43350
    },
    {
      "epoch": 4.6711871703799375,
      "grad_norm": 0.6972584128379822,
      "learning_rate": 0.0003201767050964336,
      "loss": 3.4,
      "step": 43400
    },
    {
      "epoch": 4.676568722419546,
      "grad_norm": 0.692314088344574,
      "learning_rate": 0.0003198534640663721,
      "loss": 3.3883,
      "step": 43450
    },
    {
      "epoch": 4.681950274459154,
      "grad_norm": 0.6952795386314392,
      "learning_rate": 0.0003195302230363107,
      "loss": 3.3966,
      "step": 43500
    },
    {
      "epoch": 4.687331826498762,
      "grad_norm": 0.6832876205444336,
      "learning_rate": 0.00031920698200624926,
      "loss": 3.3708,
      "step": 43550
    },
    {
      "epoch": 4.692713378538371,
      "grad_norm": 0.684529185295105,
      "learning_rate": 0.00031888374097618785,
      "loss": 3.3997,
      "step": 43600
    },
    {
      "epoch": 4.6980949305779784,
      "grad_norm": 0.6592495441436768,
      "learning_rate": 0.0003185604999461265,
      "loss": 3.3893,
      "step": 43650
    },
    {
      "epoch": 4.703476482617587,
      "grad_norm": 0.7246899604797363,
      "learning_rate": 0.00031823725891606504,
      "loss": 3.3858,
      "step": 43700
    },
    {
      "epoch": 4.7088580346571955,
      "grad_norm": 0.7060039043426514,
      "learning_rate": 0.00031791401788600364,
      "loss": 3.373,
      "step": 43750
    },
    {
      "epoch": 4.714239586696803,
      "grad_norm": 0.727345883846283,
      "learning_rate": 0.00031759077685594223,
      "loss": 3.3876,
      "step": 43800
    },
    {
      "epoch": 4.719621138736412,
      "grad_norm": 0.6933861970901489,
      "learning_rate": 0.00031726753582588077,
      "loss": 3.4002,
      "step": 43850
    },
    {
      "epoch": 4.725002690776019,
      "grad_norm": 0.7261818647384644,
      "learning_rate": 0.0003169442947958194,
      "loss": 3.397,
      "step": 43900
    },
    {
      "epoch": 4.730384242815628,
      "grad_norm": 0.6434661149978638,
      "learning_rate": 0.000316621053765758,
      "loss": 3.4044,
      "step": 43950
    },
    {
      "epoch": 4.7357657948552365,
      "grad_norm": 0.6868217587471008,
      "learning_rate": 0.00031629781273569656,
      "loss": 3.4047,
      "step": 44000
    },
    {
      "epoch": 4.7357657948552365,
      "eval_accuracy": 0.3793180966491341,
      "eval_loss": 3.421250581741333,
      "eval_runtime": 187.5428,
      "eval_samples_per_second": 96.037,
      "eval_steps_per_second": 6.004,
      "step": 44000
    },
    {
      "epoch": 4.741147346894844,
      "grad_norm": 0.6731848120689392,
      "learning_rate": 0.00031597457170563515,
      "loss": 3.3793,
      "step": 44050
    },
    {
      "epoch": 4.746528898934453,
      "grad_norm": 0.6794449687004089,
      "learning_rate": 0.0003156513306755737,
      "loss": 3.39,
      "step": 44100
    },
    {
      "epoch": 4.751910450974061,
      "grad_norm": 0.7639029622077942,
      "learning_rate": 0.0003153280896455123,
      "loss": 3.402,
      "step": 44150
    },
    {
      "epoch": 4.757292003013669,
      "grad_norm": 0.686454176902771,
      "learning_rate": 0.00031500484861545094,
      "loss": 3.3893,
      "step": 44200
    },
    {
      "epoch": 4.7626735550532775,
      "grad_norm": 0.7179557085037231,
      "learning_rate": 0.0003146816075853895,
      "loss": 3.3874,
      "step": 44250
    },
    {
      "epoch": 4.768055107092886,
      "grad_norm": 0.6821876168251038,
      "learning_rate": 0.00031435836655532807,
      "loss": 3.3928,
      "step": 44300
    },
    {
      "epoch": 4.773436659132494,
      "grad_norm": 0.6775598526000977,
      "learning_rate": 0.00031403512552526667,
      "loss": 3.381,
      "step": 44350
    },
    {
      "epoch": 4.778818211172102,
      "grad_norm": 0.666349470615387,
      "learning_rate": 0.0003137118844952052,
      "loss": 3.4018,
      "step": 44400
    },
    {
      "epoch": 4.78419976321171,
      "grad_norm": 0.6841625571250916,
      "learning_rate": 0.0003133886434651438,
      "loss": 3.3952,
      "step": 44450
    },
    {
      "epoch": 4.7895813152513185,
      "grad_norm": 0.6649987697601318,
      "learning_rate": 0.00031306540243508245,
      "loss": 3.4116,
      "step": 44500
    },
    {
      "epoch": 4.794962867290927,
      "grad_norm": 0.6441112756729126,
      "learning_rate": 0.000312742161405021,
      "loss": 3.3951,
      "step": 44550
    },
    {
      "epoch": 4.800344419330535,
      "grad_norm": 0.6589279174804688,
      "learning_rate": 0.0003124189203749596,
      "loss": 3.3902,
      "step": 44600
    },
    {
      "epoch": 4.805725971370143,
      "grad_norm": 0.677620530128479,
      "learning_rate": 0.0003120956793448981,
      "loss": 3.3967,
      "step": 44650
    },
    {
      "epoch": 4.811107523409751,
      "grad_norm": 0.7282724380493164,
      "learning_rate": 0.0003117724383148367,
      "loss": 3.386,
      "step": 44700
    },
    {
      "epoch": 4.8164890754493594,
      "grad_norm": 0.6708424687385559,
      "learning_rate": 0.00031144919728477526,
      "loss": 3.3793,
      "step": 44750
    },
    {
      "epoch": 4.821870627488968,
      "grad_norm": 0.678166925907135,
      "learning_rate": 0.0003111259562547139,
      "loss": 3.4066,
      "step": 44800
    },
    {
      "epoch": 4.827252179528576,
      "grad_norm": 0.6643491387367249,
      "learning_rate": 0.0003108027152246525,
      "loss": 3.3997,
      "step": 44850
    },
    {
      "epoch": 4.832633731568184,
      "grad_norm": 0.7268407940864563,
      "learning_rate": 0.0003104794741945911,
      "loss": 3.3941,
      "step": 44900
    },
    {
      "epoch": 4.838015283607793,
      "grad_norm": 0.666113555431366,
      "learning_rate": 0.00031015623316452964,
      "loss": 3.3933,
      "step": 44950
    },
    {
      "epoch": 4.8433968356474,
      "grad_norm": 0.6685688495635986,
      "learning_rate": 0.00030983299213446823,
      "loss": 3.3776,
      "step": 45000
    },
    {
      "epoch": 4.8433968356474,
      "eval_accuracy": 0.37968338765796633,
      "eval_loss": 3.4163522720336914,
      "eval_runtime": 202.3061,
      "eval_samples_per_second": 89.028,
      "eval_steps_per_second": 5.566,
      "step": 45000
    },
    {
      "epoch": 4.848778387687009,
      "grad_norm": 0.657361626625061,
      "learning_rate": 0.0003095097511044069,
      "loss": 3.4067,
      "step": 45050
    },
    {
      "epoch": 4.8541599397266175,
      "grad_norm": 0.6610834002494812,
      "learning_rate": 0.0003091865100743454,
      "loss": 3.4013,
      "step": 45100
    },
    {
      "epoch": 4.859541491766225,
      "grad_norm": 0.6313742995262146,
      "learning_rate": 0.00030886973386488523,
      "loss": 3.4168,
      "step": 45150
    },
    {
      "epoch": 4.864923043805834,
      "grad_norm": 0.6884166598320007,
      "learning_rate": 0.0003085464928348238,
      "loss": 3.3887,
      "step": 45200
    },
    {
      "epoch": 4.870304595845441,
      "grad_norm": 0.6939536333084106,
      "learning_rate": 0.0003082232518047624,
      "loss": 3.3801,
      "step": 45250
    },
    {
      "epoch": 4.87568614788505,
      "grad_norm": 0.6692671775817871,
      "learning_rate": 0.00030790001077470096,
      "loss": 3.3979,
      "step": 45300
    },
    {
      "epoch": 4.8810676999246585,
      "grad_norm": 0.6553664803504944,
      "learning_rate": 0.00030757676974463955,
      "loss": 3.3896,
      "step": 45350
    },
    {
      "epoch": 4.886449251964266,
      "grad_norm": 0.6963851451873779,
      "learning_rate": 0.0003072535287145781,
      "loss": 3.4001,
      "step": 45400
    },
    {
      "epoch": 4.891830804003875,
      "grad_norm": 0.6625418066978455,
      "learning_rate": 0.00030693028768451674,
      "loss": 3.3867,
      "step": 45450
    },
    {
      "epoch": 4.897212356043483,
      "grad_norm": 0.6888773441314697,
      "learning_rate": 0.00030660704665445534,
      "loss": 3.3772,
      "step": 45500
    },
    {
      "epoch": 4.902593908083091,
      "grad_norm": 0.657309889793396,
      "learning_rate": 0.0003062838056243939,
      "loss": 3.3902,
      "step": 45550
    },
    {
      "epoch": 4.9079754601226995,
      "grad_norm": 0.6960524320602417,
      "learning_rate": 0.00030596056459433247,
      "loss": 3.3857,
      "step": 45600
    },
    {
      "epoch": 4.913357012162308,
      "grad_norm": 0.6952453255653381,
      "learning_rate": 0.00030563732356427107,
      "loss": 3.4173,
      "step": 45650
    },
    {
      "epoch": 4.918738564201916,
      "grad_norm": 0.6704519987106323,
      "learning_rate": 0.00030531408253420966,
      "loss": 3.3778,
      "step": 45700
    },
    {
      "epoch": 4.924120116241524,
      "grad_norm": 0.7266472578048706,
      "learning_rate": 0.00030499084150414826,
      "loss": 3.3823,
      "step": 45750
    },
    {
      "epoch": 4.929501668281132,
      "grad_norm": 0.6816993355751038,
      "learning_rate": 0.00030466760047408685,
      "loss": 3.3877,
      "step": 45800
    },
    {
      "epoch": 4.9348832203207404,
      "grad_norm": 0.6936447620391846,
      "learning_rate": 0.0003043443594440254,
      "loss": 3.3832,
      "step": 45850
    },
    {
      "epoch": 4.940264772360349,
      "grad_norm": 0.6859702467918396,
      "learning_rate": 0.000304021118413964,
      "loss": 3.4126,
      "step": 45900
    },
    {
      "epoch": 4.945646324399957,
      "grad_norm": 0.6758688688278198,
      "learning_rate": 0.0003036978773839025,
      "loss": 3.3839,
      "step": 45950
    },
    {
      "epoch": 4.951027876439565,
      "grad_norm": 0.6840824484825134,
      "learning_rate": 0.0003033746363538412,
      "loss": 3.3902,
      "step": 46000
    },
    {
      "epoch": 4.951027876439565,
      "eval_accuracy": 0.3802585960460562,
      "eval_loss": 3.4134035110473633,
      "eval_runtime": 191.2911,
      "eval_samples_per_second": 94.155,
      "eval_steps_per_second": 5.886,
      "step": 46000
    },
    {
      "epoch": 4.956409428479174,
      "grad_norm": 0.640100359916687,
      "learning_rate": 0.00030305139532377977,
      "loss": 3.4115,
      "step": 46050
    },
    {
      "epoch": 4.961790980518781,
      "grad_norm": 0.6694591641426086,
      "learning_rate": 0.0003027281542937183,
      "loss": 3.4089,
      "step": 46100
    },
    {
      "epoch": 4.96717253255839,
      "grad_norm": 0.6837192177772522,
      "learning_rate": 0.0003024049132636569,
      "loss": 3.4014,
      "step": 46150
    },
    {
      "epoch": 4.9725540845979985,
      "grad_norm": 0.6926203966140747,
      "learning_rate": 0.0003020816722335955,
      "loss": 3.3931,
      "step": 46200
    },
    {
      "epoch": 4.977935636637606,
      "grad_norm": 0.672174870967865,
      "learning_rate": 0.00030175843120353404,
      "loss": 3.3686,
      "step": 46250
    },
    {
      "epoch": 4.983317188677215,
      "grad_norm": 0.6593598127365112,
      "learning_rate": 0.0003014351901734727,
      "loss": 3.4056,
      "step": 46300
    },
    {
      "epoch": 4.988698740716822,
      "grad_norm": 0.6591873168945312,
      "learning_rate": 0.0003011119491434113,
      "loss": 3.4209,
      "step": 46350
    },
    {
      "epoch": 4.994080292756431,
      "grad_norm": 0.6413924694061279,
      "learning_rate": 0.0003007887081133498,
      "loss": 3.3968,
      "step": 46400
    },
    {
      "epoch": 4.9994618447960395,
      "grad_norm": 0.7270100116729736,
      "learning_rate": 0.0003004654670832884,
      "loss": 3.4009,
      "step": 46450
    },
    {
      "epoch": 5.004843396835647,
      "grad_norm": 0.6969584226608276,
      "learning_rate": 0.00030014222605322696,
      "loss": 3.3214,
      "step": 46500
    },
    {
      "epoch": 5.010224948875256,
      "grad_norm": 0.7092729806900024,
      "learning_rate": 0.00029981898502316555,
      "loss": 3.3004,
      "step": 46550
    },
    {
      "epoch": 5.015606500914864,
      "grad_norm": 0.7010934352874756,
      "learning_rate": 0.00029949574399310415,
      "loss": 3.3056,
      "step": 46600
    },
    {
      "epoch": 5.020988052954472,
      "grad_norm": 0.6872923970222473,
      "learning_rate": 0.00029917250296304274,
      "loss": 3.3137,
      "step": 46650
    },
    {
      "epoch": 5.0263696049940805,
      "grad_norm": 0.6743913292884827,
      "learning_rate": 0.00029884926193298134,
      "loss": 3.3004,
      "step": 46700
    },
    {
      "epoch": 5.031751157033688,
      "grad_norm": 0.7134643197059631,
      "learning_rate": 0.0002985260209029199,
      "loss": 3.2892,
      "step": 46750
    },
    {
      "epoch": 5.037132709073297,
      "grad_norm": 0.6986505389213562,
      "learning_rate": 0.00029820277987285853,
      "loss": 3.2902,
      "step": 46800
    },
    {
      "epoch": 5.042514261112905,
      "grad_norm": 0.6677493453025818,
      "learning_rate": 0.00029787953884279707,
      "loss": 3.3036,
      "step": 46850
    },
    {
      "epoch": 5.047895813152513,
      "grad_norm": 0.6945018172264099,
      "learning_rate": 0.00029755629781273566,
      "loss": 3.3008,
      "step": 46900
    },
    {
      "epoch": 5.0532773651921215,
      "grad_norm": 0.7299492955207825,
      "learning_rate": 0.00029723305678267426,
      "loss": 3.3021,
      "step": 46950
    },
    {
      "epoch": 5.05865891723173,
      "grad_norm": 0.6931959390640259,
      "learning_rate": 0.00029690981575261285,
      "loss": 3.2993,
      "step": 47000
    },
    {
      "epoch": 5.05865891723173,
      "eval_accuracy": 0.38065050701240305,
      "eval_loss": 3.4127960205078125,
      "eval_runtime": 201.6655,
      "eval_samples_per_second": 89.311,
      "eval_steps_per_second": 5.584,
      "step": 47000
    },
    {
      "epoch": 5.064040469271338,
      "grad_norm": 0.7374845147132874,
      "learning_rate": 0.0002965865747225514,
      "loss": 3.3001,
      "step": 47050
    },
    {
      "epoch": 5.069422021310946,
      "grad_norm": 0.6965045928955078,
      "learning_rate": 0.00029626333369249,
      "loss": 3.3073,
      "step": 47100
    },
    {
      "epoch": 5.074803573350554,
      "grad_norm": 0.6908303499221802,
      "learning_rate": 0.0002959400926624286,
      "loss": 3.3072,
      "step": 47150
    },
    {
      "epoch": 5.080185125390162,
      "grad_norm": 0.6920403242111206,
      "learning_rate": 0.0002956168516323672,
      "loss": 3.318,
      "step": 47200
    },
    {
      "epoch": 5.085566677429771,
      "grad_norm": 0.7072156071662903,
      "learning_rate": 0.00029529361060230577,
      "loss": 3.3103,
      "step": 47250
    },
    {
      "epoch": 5.090948229469379,
      "grad_norm": 0.691167414188385,
      "learning_rate": 0.0002949703695722443,
      "loss": 3.3072,
      "step": 47300
    },
    {
      "epoch": 5.096329781508987,
      "grad_norm": 0.6893109679222107,
      "learning_rate": 0.00029464712854218296,
      "loss": 3.3113,
      "step": 47350
    },
    {
      "epoch": 5.101711333548596,
      "grad_norm": 0.6975725889205933,
      "learning_rate": 0.0002943238875121215,
      "loss": 3.3042,
      "step": 47400
    },
    {
      "epoch": 5.107092885588203,
      "grad_norm": 0.7136406302452087,
      "learning_rate": 0.0002940006464820601,
      "loss": 3.3148,
      "step": 47450
    },
    {
      "epoch": 5.112474437627812,
      "grad_norm": 0.7137263417243958,
      "learning_rate": 0.0002936774054519987,
      "loss": 3.3261,
      "step": 47500
    },
    {
      "epoch": 5.1178559896674205,
      "grad_norm": 0.6680945754051208,
      "learning_rate": 0.0002933541644219373,
      "loss": 3.337,
      "step": 47550
    },
    {
      "epoch": 5.123237541707028,
      "grad_norm": 0.8087601661682129,
      "learning_rate": 0.0002930309233918758,
      "loss": 3.3087,
      "step": 47600
    },
    {
      "epoch": 5.128619093746637,
      "grad_norm": 0.6731941103935242,
      "learning_rate": 0.0002927076823618144,
      "loss": 3.3181,
      "step": 47650
    },
    {
      "epoch": 5.134000645786244,
      "grad_norm": 0.7058475017547607,
      "learning_rate": 0.000292384441331753,
      "loss": 3.3101,
      "step": 47700
    },
    {
      "epoch": 5.139382197825853,
      "grad_norm": 0.7096105813980103,
      "learning_rate": 0.0002920612003016916,
      "loss": 3.331,
      "step": 47750
    },
    {
      "epoch": 5.1447637498654615,
      "grad_norm": 0.7284489274024963,
      "learning_rate": 0.0002917379592716302,
      "loss": 3.3443,
      "step": 47800
    },
    {
      "epoch": 5.150145301905069,
      "grad_norm": 0.6959323883056641,
      "learning_rate": 0.00029141471824156874,
      "loss": 3.3013,
      "step": 47850
    },
    {
      "epoch": 5.155526853944678,
      "grad_norm": 0.7227029204368591,
      "learning_rate": 0.00029109147721150734,
      "loss": 3.3017,
      "step": 47900
    },
    {
      "epoch": 5.160908405984286,
      "grad_norm": 0.6779053807258606,
      "learning_rate": 0.00029076823618144593,
      "loss": 3.3166,
      "step": 47950
    },
    {
      "epoch": 5.166289958023894,
      "grad_norm": 0.7135473489761353,
      "learning_rate": 0.00029044499515138453,
      "loss": 3.3255,
      "step": 48000
    },
    {
      "epoch": 5.166289958023894,
      "eval_accuracy": 0.3803929996682827,
      "eval_loss": 3.412876844406128,
      "eval_runtime": 192.717,
      "eval_samples_per_second": 93.458,
      "eval_steps_per_second": 5.843,
      "step": 48000
    },
    {
      "epoch": 5.1716715100635025,
      "grad_norm": 0.6599064469337463,
      "learning_rate": 0.0002901217541213231,
      "loss": 3.3175,
      "step": 48050
    },
    {
      "epoch": 5.17705306210311,
      "grad_norm": 0.7011419534683228,
      "learning_rate": 0.0002897985130912617,
      "loss": 3.3167,
      "step": 48100
    },
    {
      "epoch": 5.182434614142719,
      "grad_norm": 0.7181657552719116,
      "learning_rate": 0.00028947527206120026,
      "loss": 3.3397,
      "step": 48150
    },
    {
      "epoch": 5.187816166182327,
      "grad_norm": 0.6871926784515381,
      "learning_rate": 0.00028915849585174006,
      "loss": 3.3162,
      "step": 48200
    },
    {
      "epoch": 5.193197718221935,
      "grad_norm": 0.7430890202522278,
      "learning_rate": 0.00028883525482167866,
      "loss": 3.3273,
      "step": 48250
    },
    {
      "epoch": 5.198579270261543,
      "grad_norm": 0.7614420056343079,
      "learning_rate": 0.00028851201379161725,
      "loss": 3.3204,
      "step": 48300
    },
    {
      "epoch": 5.203960822301152,
      "grad_norm": 0.6756367683410645,
      "learning_rate": 0.00028818877276155585,
      "loss": 3.327,
      "step": 48350
    },
    {
      "epoch": 5.20934237434076,
      "grad_norm": 0.7239832282066345,
      "learning_rate": 0.0002878655317314944,
      "loss": 3.3298,
      "step": 48400
    },
    {
      "epoch": 5.214723926380368,
      "grad_norm": 0.7229234576225281,
      "learning_rate": 0.00028754229070143304,
      "loss": 3.3227,
      "step": 48450
    },
    {
      "epoch": 5.220105478419977,
      "grad_norm": 0.715870201587677,
      "learning_rate": 0.0002872190496713716,
      "loss": 3.3315,
      "step": 48500
    },
    {
      "epoch": 5.225487030459584,
      "grad_norm": 0.714283287525177,
      "learning_rate": 0.0002868958086413102,
      "loss": 3.3058,
      "step": 48550
    },
    {
      "epoch": 5.230868582499193,
      "grad_norm": 0.7448010444641113,
      "learning_rate": 0.00028657256761124877,
      "loss": 3.3418,
      "step": 48600
    },
    {
      "epoch": 5.236250134538801,
      "grad_norm": 0.6719698309898376,
      "learning_rate": 0.00028624932658118736,
      "loss": 3.338,
      "step": 48650
    },
    {
      "epoch": 5.241631686578409,
      "grad_norm": 0.7021603584289551,
      "learning_rate": 0.0002859260855511259,
      "loss": 3.3204,
      "step": 48700
    },
    {
      "epoch": 5.247013238618018,
      "grad_norm": 0.7057817578315735,
      "learning_rate": 0.0002856028445210645,
      "loss": 3.3298,
      "step": 48750
    },
    {
      "epoch": 5.252394790657625,
      "grad_norm": 0.7323166131973267,
      "learning_rate": 0.0002852796034910031,
      "loss": 3.3229,
      "step": 48800
    },
    {
      "epoch": 5.257776342697234,
      "grad_norm": 0.7022436261177063,
      "learning_rate": 0.0002849563624609417,
      "loss": 3.3345,
      "step": 48850
    },
    {
      "epoch": 5.2631578947368425,
      "grad_norm": 0.7284606099128723,
      "learning_rate": 0.0002846331214308803,
      "loss": 3.3281,
      "step": 48900
    },
    {
      "epoch": 5.26853944677645,
      "grad_norm": 0.7403408288955688,
      "learning_rate": 0.0002843098804008188,
      "loss": 3.3185,
      "step": 48950
    },
    {
      "epoch": 5.273920998816059,
      "grad_norm": 0.7374369502067566,
      "learning_rate": 0.00028398663937075747,
      "loss": 3.3417,
      "step": 49000
    },
    {
      "epoch": 5.273920998816059,
      "eval_accuracy": 0.3811685639815025,
      "eval_loss": 3.4100325107574463,
      "eval_runtime": 186.625,
      "eval_samples_per_second": 96.509,
      "eval_steps_per_second": 6.033,
      "step": 49000
    },
    {
      "epoch": 5.279302550855666,
      "grad_norm": 0.7417536377906799,
      "learning_rate": 0.000283663398340696,
      "loss": 3.3255,
      "step": 49050
    },
    {
      "epoch": 5.284684102895275,
      "grad_norm": 0.7349126935005188,
      "learning_rate": 0.0002833401573106346,
      "loss": 3.3224,
      "step": 49100
    },
    {
      "epoch": 5.2900656549348835,
      "grad_norm": 0.6606712341308594,
      "learning_rate": 0.0002830169162805732,
      "loss": 3.3305,
      "step": 49150
    },
    {
      "epoch": 5.295447206974491,
      "grad_norm": 0.6845487952232361,
      "learning_rate": 0.0002826936752505118,
      "loss": 3.3289,
      "step": 49200
    },
    {
      "epoch": 5.3008287590141,
      "grad_norm": 0.7850626111030579,
      "learning_rate": 0.00028237043422045034,
      "loss": 3.338,
      "step": 49250
    },
    {
      "epoch": 5.306210311053708,
      "grad_norm": 0.7478814721107483,
      "learning_rate": 0.00028204719319038893,
      "loss": 3.3253,
      "step": 49300
    },
    {
      "epoch": 5.311591863093316,
      "grad_norm": 0.6879907250404358,
      "learning_rate": 0.0002817239521603275,
      "loss": 3.3235,
      "step": 49350
    },
    {
      "epoch": 5.316973415132924,
      "grad_norm": 0.7131553292274475,
      "learning_rate": 0.0002814007111302661,
      "loss": 3.3212,
      "step": 49400
    },
    {
      "epoch": 5.322354967172533,
      "grad_norm": 0.6914525032043457,
      "learning_rate": 0.0002810774701002047,
      "loss": 3.3049,
      "step": 49450
    },
    {
      "epoch": 5.327736519212141,
      "grad_norm": 0.7019218802452087,
      "learning_rate": 0.00028075422907014325,
      "loss": 3.3435,
      "step": 49500
    },
    {
      "epoch": 5.333118071251749,
      "grad_norm": 0.7138478755950928,
      "learning_rate": 0.00028043098804008185,
      "loss": 3.3277,
      "step": 49550
    },
    {
      "epoch": 5.338499623291357,
      "grad_norm": 0.7029947638511658,
      "learning_rate": 0.00028010774701002044,
      "loss": 3.3302,
      "step": 49600
    },
    {
      "epoch": 5.343881175330965,
      "grad_norm": 0.7279183864593506,
      "learning_rate": 0.00027978450597995904,
      "loss": 3.3124,
      "step": 49650
    },
    {
      "epoch": 5.349262727370574,
      "grad_norm": 0.7827022671699524,
      "learning_rate": 0.0002794612649498976,
      "loss": 3.324,
      "step": 49700
    },
    {
      "epoch": 5.354644279410182,
      "grad_norm": 0.7062620520591736,
      "learning_rate": 0.00027913802391983623,
      "loss": 3.3445,
      "step": 49750
    },
    {
      "epoch": 5.36002583144979,
      "grad_norm": 0.6948559284210205,
      "learning_rate": 0.00027881478288977477,
      "loss": 3.347,
      "step": 49800
    },
    {
      "epoch": 5.365407383489399,
      "grad_norm": 0.7399136424064636,
      "learning_rate": 0.00027849154185971336,
      "loss": 3.3452,
      "step": 49850
    },
    {
      "epoch": 5.370788935529006,
      "grad_norm": 0.7397146821022034,
      "learning_rate": 0.00027816830082965196,
      "loss": 3.3324,
      "step": 49900
    },
    {
      "epoch": 5.376170487568615,
      "grad_norm": 0.7319924831390381,
      "learning_rate": 0.00027784505979959055,
      "loss": 3.3462,
      "step": 49950
    },
    {
      "epoch": 5.3815520396082235,
      "grad_norm": 0.7393072843551636,
      "learning_rate": 0.00027752181876952915,
      "loss": 3.3513,
      "step": 50000
    },
    {
      "epoch": 5.3815520396082235,
      "eval_accuracy": 0.38153733188274486,
      "eval_loss": 3.4052817821502686,
      "eval_runtime": 202.9452,
      "eval_samples_per_second": 88.748,
      "eval_steps_per_second": 5.548,
      "step": 50000
    },
    {
      "epoch": 5.386933591647831,
      "grad_norm": 0.6867526173591614,
      "learning_rate": 0.0002771985777394677,
      "loss": 3.3115,
      "step": 50050
    },
    {
      "epoch": 5.39231514368744,
      "grad_norm": 0.6913804411888123,
      "learning_rate": 0.0002768753367094063,
      "loss": 3.3469,
      "step": 50100
    },
    {
      "epoch": 5.397696695727047,
      "grad_norm": 0.7089942097663879,
      "learning_rate": 0.0002765520956793449,
      "loss": 3.3267,
      "step": 50150
    },
    {
      "epoch": 5.403078247766656,
      "grad_norm": 0.6663913726806641,
      "learning_rate": 0.00027622885464928347,
      "loss": 3.3485,
      "step": 50200
    },
    {
      "epoch": 5.4084597998062645,
      "grad_norm": 0.6780911684036255,
      "learning_rate": 0.000275905613619222,
      "loss": 3.323,
      "step": 50250
    },
    {
      "epoch": 5.413841351845872,
      "grad_norm": 0.6843054294586182,
      "learning_rate": 0.00027558237258916066,
      "loss": 3.3323,
      "step": 50300
    },
    {
      "epoch": 5.419222903885481,
      "grad_norm": 0.6963738799095154,
      "learning_rate": 0.0002752591315590992,
      "loss": 3.3252,
      "step": 50350
    },
    {
      "epoch": 5.424604455925088,
      "grad_norm": 0.7334024906158447,
      "learning_rate": 0.0002749358905290378,
      "loss": 3.3618,
      "step": 50400
    },
    {
      "epoch": 5.429986007964697,
      "grad_norm": 0.7072751522064209,
      "learning_rate": 0.0002746126494989764,
      "loss": 3.3496,
      "step": 50450
    },
    {
      "epoch": 5.435367560004305,
      "grad_norm": 0.6565462350845337,
      "learning_rate": 0.000274289408468915,
      "loss": 3.3377,
      "step": 50500
    },
    {
      "epoch": 5.440749112043913,
      "grad_norm": 0.7185359597206116,
      "learning_rate": 0.0002739661674388535,
      "loss": 3.3336,
      "step": 50550
    },
    {
      "epoch": 5.446130664083522,
      "grad_norm": 0.6985934972763062,
      "learning_rate": 0.0002736429264087921,
      "loss": 3.3497,
      "step": 50600
    },
    {
      "epoch": 5.45151221612313,
      "grad_norm": 0.778202474117279,
      "learning_rate": 0.0002733196853787307,
      "loss": 3.332,
      "step": 50650
    },
    {
      "epoch": 5.456893768162738,
      "grad_norm": 0.7283360958099365,
      "learning_rate": 0.0002729964443486693,
      "loss": 3.3343,
      "step": 50700
    },
    {
      "epoch": 5.462275320202346,
      "grad_norm": 0.7053759098052979,
      "learning_rate": 0.0002726732033186079,
      "loss": 3.3314,
      "step": 50750
    },
    {
      "epoch": 5.467656872241955,
      "grad_norm": 0.7685927748680115,
      "learning_rate": 0.00027234996228854644,
      "loss": 3.3358,
      "step": 50800
    },
    {
      "epoch": 5.473038424281563,
      "grad_norm": 0.6882610321044922,
      "learning_rate": 0.0002720267212584851,
      "loss": 3.3294,
      "step": 50850
    },
    {
      "epoch": 5.478419976321171,
      "grad_norm": 0.7401322722434998,
      "learning_rate": 0.00027170348022842363,
      "loss": 3.333,
      "step": 50900
    },
    {
      "epoch": 5.483801528360779,
      "grad_norm": 0.7363923192024231,
      "learning_rate": 0.00027138023919836223,
      "loss": 3.3233,
      "step": 50950
    },
    {
      "epoch": 5.489183080400387,
      "grad_norm": 0.6889253258705139,
      "learning_rate": 0.00027106346298890204,
      "loss": 3.334,
      "step": 51000
    },
    {
      "epoch": 5.489183080400387,
      "eval_accuracy": 0.38220174429173037,
      "eval_loss": 3.4004783630371094,
      "eval_runtime": 190.6388,
      "eval_samples_per_second": 94.477,
      "eval_steps_per_second": 5.906,
      "step": 51000
    },
    {
      "epoch": 5.494564632439996,
      "grad_norm": 0.7611465454101562,
      "learning_rate": 0.00027074022195884063,
      "loss": 3.3499,
      "step": 51050
    },
    {
      "epoch": 5.499946184479604,
      "grad_norm": 0.6965069770812988,
      "learning_rate": 0.0002704169809287792,
      "loss": 3.3126,
      "step": 51100
    },
    {
      "epoch": 5.505327736519212,
      "grad_norm": 0.6993700861930847,
      "learning_rate": 0.00027009373989871776,
      "loss": 3.3497,
      "step": 51150
    },
    {
      "epoch": 5.510709288558821,
      "grad_norm": 0.7188165187835693,
      "learning_rate": 0.00026977049886865636,
      "loss": 3.3468,
      "step": 51200
    },
    {
      "epoch": 5.516090840598428,
      "grad_norm": 0.7151656746864319,
      "learning_rate": 0.00026944725783859495,
      "loss": 3.3301,
      "step": 51250
    },
    {
      "epoch": 5.521472392638037,
      "grad_norm": 0.7025049924850464,
      "learning_rate": 0.00026912401680853355,
      "loss": 3.3395,
      "step": 51300
    },
    {
      "epoch": 5.5268539446776455,
      "grad_norm": 0.7040926814079285,
      "learning_rate": 0.0002688007757784721,
      "loss": 3.3292,
      "step": 51350
    },
    {
      "epoch": 5.532235496717253,
      "grad_norm": 0.7505577802658081,
      "learning_rate": 0.00026847753474841074,
      "loss": 3.3488,
      "step": 51400
    },
    {
      "epoch": 5.537617048756862,
      "grad_norm": 0.7513281106948853,
      "learning_rate": 0.0002681542937183493,
      "loss": 3.3521,
      "step": 51450
    },
    {
      "epoch": 5.542998600796469,
      "grad_norm": 0.790105938911438,
      "learning_rate": 0.0002678310526882879,
      "loss": 3.3423,
      "step": 51500
    },
    {
      "epoch": 5.548380152836078,
      "grad_norm": 0.7354403138160706,
      "learning_rate": 0.00026750781165822647,
      "loss": 3.3304,
      "step": 51550
    },
    {
      "epoch": 5.553761704875686,
      "grad_norm": 0.7015941739082336,
      "learning_rate": 0.00026718457062816506,
      "loss": 3.3444,
      "step": 51600
    },
    {
      "epoch": 5.559143256915294,
      "grad_norm": 0.7292041182518005,
      "learning_rate": 0.00026686132959810366,
      "loss": 3.3471,
      "step": 51650
    },
    {
      "epoch": 5.564524808954903,
      "grad_norm": 0.7539200186729431,
      "learning_rate": 0.0002665380885680422,
      "loss": 3.3464,
      "step": 51700
    },
    {
      "epoch": 5.569906360994511,
      "grad_norm": 0.730979859828949,
      "learning_rate": 0.0002662148475379808,
      "loss": 3.3338,
      "step": 51750
    },
    {
      "epoch": 5.575287913034119,
      "grad_norm": 0.7592946887016296,
      "learning_rate": 0.0002658916065079194,
      "loss": 3.3366,
      "step": 51800
    },
    {
      "epoch": 5.580669465073727,
      "grad_norm": 0.7109083533287048,
      "learning_rate": 0.000265568365477858,
      "loss": 3.3529,
      "step": 51850
    },
    {
      "epoch": 5.586051017113336,
      "grad_norm": 0.7419684529304504,
      "learning_rate": 0.0002652451244477965,
      "loss": 3.333,
      "step": 51900
    },
    {
      "epoch": 5.591432569152944,
      "grad_norm": 0.7088695764541626,
      "learning_rate": 0.00026492188341773517,
      "loss": 3.3527,
      "step": 51950
    },
    {
      "epoch": 5.596814121192552,
      "grad_norm": 0.721508800983429,
      "learning_rate": 0.0002645986423876737,
      "loss": 3.334,
      "step": 52000
    },
    {
      "epoch": 5.596814121192552,
      "eval_accuracy": 0.3824253519348526,
      "eval_loss": 3.3942997455596924,
      "eval_runtime": 201.2401,
      "eval_samples_per_second": 89.5,
      "eval_steps_per_second": 5.595,
      "step": 52000
    },
    {
      "epoch": 5.60219567323216,
      "grad_norm": 0.7105787396430969,
      "learning_rate": 0.0002642754013576123,
      "loss": 3.3619,
      "step": 52050
    },
    {
      "epoch": 5.607577225271768,
      "grad_norm": 0.7615808844566345,
      "learning_rate": 0.0002639521603275509,
      "loss": 3.333,
      "step": 52100
    },
    {
      "epoch": 5.612958777311377,
      "grad_norm": 0.7386506199836731,
      "learning_rate": 0.0002636289192974895,
      "loss": 3.3353,
      "step": 52150
    },
    {
      "epoch": 5.618340329350985,
      "grad_norm": 0.7279359102249146,
      "learning_rate": 0.00026330567826742804,
      "loss": 3.3381,
      "step": 52200
    },
    {
      "epoch": 5.623721881390593,
      "grad_norm": 0.7196009755134583,
      "learning_rate": 0.00026298243723736663,
      "loss": 3.3369,
      "step": 52250
    },
    {
      "epoch": 5.629103433430201,
      "grad_norm": 0.7014727592468262,
      "learning_rate": 0.0002626591962073052,
      "loss": 3.3391,
      "step": 52300
    },
    {
      "epoch": 5.634484985469809,
      "grad_norm": 0.7047598958015442,
      "learning_rate": 0.0002623359551772438,
      "loss": 3.3345,
      "step": 52350
    },
    {
      "epoch": 5.639866537509418,
      "grad_norm": 0.764114499092102,
      "learning_rate": 0.0002620127141471824,
      "loss": 3.3531,
      "step": 52400
    },
    {
      "epoch": 5.645248089549026,
      "grad_norm": 0.7131911516189575,
      "learning_rate": 0.00026168947311712095,
      "loss": 3.3343,
      "step": 52450
    },
    {
      "epoch": 5.650629641588634,
      "grad_norm": 0.7622030973434448,
      "learning_rate": 0.00026136623208705955,
      "loss": 3.3465,
      "step": 52500
    },
    {
      "epoch": 5.656011193628243,
      "grad_norm": 0.7329249382019043,
      "learning_rate": 0.00026104299105699814,
      "loss": 3.3301,
      "step": 52550
    },
    {
      "epoch": 5.66139274566785,
      "grad_norm": 0.7030691504478455,
      "learning_rate": 0.00026071975002693674,
      "loss": 3.3294,
      "step": 52600
    },
    {
      "epoch": 5.666774297707459,
      "grad_norm": 0.7377075552940369,
      "learning_rate": 0.00026039650899687533,
      "loss": 3.3589,
      "step": 52650
    },
    {
      "epoch": 5.672155849747067,
      "grad_norm": 0.6939163208007812,
      "learning_rate": 0.00026007326796681393,
      "loss": 3.3312,
      "step": 52700
    },
    {
      "epoch": 5.677537401786675,
      "grad_norm": 0.7312877178192139,
      "learning_rate": 0.00025975002693675247,
      "loss": 3.343,
      "step": 52750
    },
    {
      "epoch": 5.682918953826284,
      "grad_norm": 0.7162089943885803,
      "learning_rate": 0.00025942678590669106,
      "loss": 3.3215,
      "step": 52800
    },
    {
      "epoch": 5.688300505865891,
      "grad_norm": 0.7658792734146118,
      "learning_rate": 0.00025910354487662966,
      "loss": 3.3147,
      "step": 52850
    },
    {
      "epoch": 5.6936820579055,
      "grad_norm": 0.7245572209358215,
      "learning_rate": 0.00025878030384656825,
      "loss": 3.3414,
      "step": 52900
    },
    {
      "epoch": 5.699063609945108,
      "grad_norm": 0.716014564037323,
      "learning_rate": 0.00025845706281650685,
      "loss": 3.3339,
      "step": 52950
    },
    {
      "epoch": 5.704445161984716,
      "grad_norm": 0.7805140018463135,
      "learning_rate": 0.0002581338217864454,
      "loss": 3.3328,
      "step": 53000
    },
    {
      "epoch": 5.704445161984716,
      "eval_accuracy": 0.3828685470642479,
      "eval_loss": 3.3905136585235596,
      "eval_runtime": 201.7044,
      "eval_samples_per_second": 89.294,
      "eval_steps_per_second": 5.582,
      "step": 53000
    },
    {
      "epoch": 5.709826714024325,
      "grad_norm": 0.748485803604126,
      "learning_rate": 0.000257810580756384,
      "loss": 3.3468,
      "step": 53050
    },
    {
      "epoch": 5.715208266063933,
      "grad_norm": 0.7458595037460327,
      "learning_rate": 0.0002574873397263226,
      "loss": 3.3035,
      "step": 53100
    },
    {
      "epoch": 5.720589818103541,
      "grad_norm": 0.6711325645446777,
      "learning_rate": 0.00025716409869626117,
      "loss": 3.3394,
      "step": 53150
    },
    {
      "epoch": 5.725971370143149,
      "grad_norm": 0.7426289319992065,
      "learning_rate": 0.0002568408576661997,
      "loss": 3.3419,
      "step": 53200
    },
    {
      "epoch": 5.731352922182758,
      "grad_norm": 0.716453492641449,
      "learning_rate": 0.00025651761663613836,
      "loss": 3.3418,
      "step": 53250
    },
    {
      "epoch": 5.736734474222366,
      "grad_norm": 0.7246510982513428,
      "learning_rate": 0.0002561943756060769,
      "loss": 3.3546,
      "step": 53300
    },
    {
      "epoch": 5.742116026261974,
      "grad_norm": 0.705548882484436,
      "learning_rate": 0.0002558711345760155,
      "loss": 3.3553,
      "step": 53350
    },
    {
      "epoch": 5.747497578301582,
      "grad_norm": 0.7819035053253174,
      "learning_rate": 0.0002555478935459541,
      "loss": 3.325,
      "step": 53400
    },
    {
      "epoch": 5.75287913034119,
      "grad_norm": 0.7534633278846741,
      "learning_rate": 0.00025522465251589263,
      "loss": 3.3426,
      "step": 53450
    },
    {
      "epoch": 5.758260682380799,
      "grad_norm": 0.7411068081855774,
      "learning_rate": 0.0002549014114858312,
      "loss": 3.3346,
      "step": 53500
    },
    {
      "epoch": 5.763642234420407,
      "grad_norm": 0.7141110897064209,
      "learning_rate": 0.0002545781704557698,
      "loss": 3.3429,
      "step": 53550
    },
    {
      "epoch": 5.769023786460015,
      "grad_norm": 0.7201070189476013,
      "learning_rate": 0.0002542549294257084,
      "loss": 3.3355,
      "step": 53600
    },
    {
      "epoch": 5.774405338499624,
      "grad_norm": 0.774978518486023,
      "learning_rate": 0.000253931688395647,
      "loss": 3.3345,
      "step": 53650
    },
    {
      "epoch": 5.779786890539231,
      "grad_norm": 0.7548540830612183,
      "learning_rate": 0.0002536084473655856,
      "loss": 3.3389,
      "step": 53700
    },
    {
      "epoch": 5.78516844257884,
      "grad_norm": 0.6902544498443604,
      "learning_rate": 0.00025328520633552415,
      "loss": 3.3277,
      "step": 53750
    },
    {
      "epoch": 5.790549994618448,
      "grad_norm": 0.7907364964485168,
      "learning_rate": 0.00025296196530546274,
      "loss": 3.3482,
      "step": 53800
    },
    {
      "epoch": 5.795931546658056,
      "grad_norm": 0.7004760503768921,
      "learning_rate": 0.00025263872427540133,
      "loss": 3.3535,
      "step": 53850
    },
    {
      "epoch": 5.801313098697665,
      "grad_norm": 0.7246608138084412,
      "learning_rate": 0.00025231548324533993,
      "loss": 3.3375,
      "step": 53900
    },
    {
      "epoch": 5.806694650737272,
      "grad_norm": 0.713356077671051,
      "learning_rate": 0.0002519922422152785,
      "loss": 3.3441,
      "step": 53950
    },
    {
      "epoch": 5.812076202776881,
      "grad_norm": 0.7337074279785156,
      "learning_rate": 0.00025166900118521706,
      "loss": 3.3229,
      "step": 54000
    },
    {
      "epoch": 5.812076202776881,
      "eval_accuracy": 0.3832519831053452,
      "eval_loss": 3.3878395557403564,
      "eval_runtime": 188.3897,
      "eval_samples_per_second": 95.605,
      "eval_steps_per_second": 5.977,
      "step": 54000
    },
    {
      "epoch": 5.817457754816489,
      "grad_norm": 0.7938897609710693,
      "learning_rate": 0.0002513522249757569,
      "loss": 3.329,
      "step": 54050
    },
    {
      "epoch": 5.822839306856097,
      "grad_norm": 0.707722544670105,
      "learning_rate": 0.00025102898394569547,
      "loss": 3.3292,
      "step": 54100
    },
    {
      "epoch": 5.828220858895706,
      "grad_norm": 0.7328303456306458,
      "learning_rate": 0.00025070574291563406,
      "loss": 3.3384,
      "step": 54150
    },
    {
      "epoch": 5.833602410935313,
      "grad_norm": 0.7096349000930786,
      "learning_rate": 0.00025038250188557265,
      "loss": 3.3243,
      "step": 54200
    },
    {
      "epoch": 5.838983962974922,
      "grad_norm": 0.7162660360336304,
      "learning_rate": 0.00025005926085551125,
      "loss": 3.3405,
      "step": 54250
    },
    {
      "epoch": 5.84436551501453,
      "grad_norm": 0.7353925108909607,
      "learning_rate": 0.0002497360198254498,
      "loss": 3.3328,
      "step": 54300
    },
    {
      "epoch": 5.849747067054138,
      "grad_norm": 0.6959530115127563,
      "learning_rate": 0.00024941277879538844,
      "loss": 3.3515,
      "step": 54350
    },
    {
      "epoch": 5.855128619093747,
      "grad_norm": 0.7663270235061646,
      "learning_rate": 0.000249089537765327,
      "loss": 3.3076,
      "step": 54400
    },
    {
      "epoch": 5.860510171133355,
      "grad_norm": 0.6977930665016174,
      "learning_rate": 0.0002487662967352656,
      "loss": 3.3244,
      "step": 54450
    },
    {
      "epoch": 5.865891723172963,
      "grad_norm": 0.7649244666099548,
      "learning_rate": 0.00024844305570520417,
      "loss": 3.3401,
      "step": 54500
    },
    {
      "epoch": 5.871273275212571,
      "grad_norm": 0.7535280585289001,
      "learning_rate": 0.00024811981467514276,
      "loss": 3.3485,
      "step": 54550
    },
    {
      "epoch": 5.87665482725218,
      "grad_norm": 0.6970841884613037,
      "learning_rate": 0.00024779657364508136,
      "loss": 3.3303,
      "step": 54600
    },
    {
      "epoch": 5.882036379291788,
      "grad_norm": 0.7283545732498169,
      "learning_rate": 0.0002474733326150199,
      "loss": 3.3424,
      "step": 54650
    },
    {
      "epoch": 5.887417931331396,
      "grad_norm": 0.721764326095581,
      "learning_rate": 0.0002471500915849585,
      "loss": 3.341,
      "step": 54700
    },
    {
      "epoch": 5.892799483371004,
      "grad_norm": 0.687520444393158,
      "learning_rate": 0.0002468268505548971,
      "loss": 3.3415,
      "step": 54750
    },
    {
      "epoch": 5.898181035410612,
      "grad_norm": 0.7138300538063049,
      "learning_rate": 0.0002465036095248357,
      "loss": 3.3447,
      "step": 54800
    },
    {
      "epoch": 5.903562587450221,
      "grad_norm": 0.6954256296157837,
      "learning_rate": 0.0002461803684947742,
      "loss": 3.3466,
      "step": 54850
    },
    {
      "epoch": 5.9089441394898286,
      "grad_norm": 0.7353760004043579,
      "learning_rate": 0.0002458571274647128,
      "loss": 3.3411,
      "step": 54900
    },
    {
      "epoch": 5.914325691529437,
      "grad_norm": 0.8013879656791687,
      "learning_rate": 0.0002455338864346514,
      "loss": 3.3581,
      "step": 54950
    },
    {
      "epoch": 5.919707243569046,
      "grad_norm": 0.7387758493423462,
      "learning_rate": 0.00024521064540459,
      "loss": 3.3521,
      "step": 55000
    },
    {
      "epoch": 5.919707243569046,
      "eval_accuracy": 0.3836273788327442,
      "eval_loss": 3.3835108280181885,
      "eval_runtime": 197.7765,
      "eval_samples_per_second": 91.067,
      "eval_steps_per_second": 5.693,
      "step": 55000
    },
    {
      "epoch": 5.925088795608653,
      "grad_norm": 0.7418875098228455,
      "learning_rate": 0.0002448874043745286,
      "loss": 3.3297,
      "step": 55050
    },
    {
      "epoch": 5.930470347648262,
      "grad_norm": 0.6959251165390015,
      "learning_rate": 0.00024456416334446714,
      "loss": 3.3306,
      "step": 55100
    },
    {
      "epoch": 5.93585189968787,
      "grad_norm": 0.6997560858726501,
      "learning_rate": 0.00024424092231440574,
      "loss": 3.3368,
      "step": 55150
    },
    {
      "epoch": 5.941233451727478,
      "grad_norm": 0.6774752736091614,
      "learning_rate": 0.00024391768128434436,
      "loss": 3.3398,
      "step": 55200
    },
    {
      "epoch": 5.946615003767087,
      "grad_norm": 0.728191614151001,
      "learning_rate": 0.00024359444025428293,
      "loss": 3.3474,
      "step": 55250
    },
    {
      "epoch": 5.951996555806694,
      "grad_norm": 0.7751122713088989,
      "learning_rate": 0.0002432711992242215,
      "loss": 3.3345,
      "step": 55300
    },
    {
      "epoch": 5.957378107846303,
      "grad_norm": 0.6916903257369995,
      "learning_rate": 0.0002429479581941601,
      "loss": 3.3525,
      "step": 55350
    },
    {
      "epoch": 5.962759659885911,
      "grad_norm": 0.704298734664917,
      "learning_rate": 0.00024262471716409868,
      "loss": 3.3457,
      "step": 55400
    },
    {
      "epoch": 5.968141211925519,
      "grad_norm": 0.7079257369041443,
      "learning_rate": 0.00024230147613403728,
      "loss": 3.3441,
      "step": 55450
    },
    {
      "epoch": 5.973522763965128,
      "grad_norm": 0.8067532777786255,
      "learning_rate": 0.00024197823510397584,
      "loss": 3.3431,
      "step": 55500
    },
    {
      "epoch": 5.978904316004736,
      "grad_norm": 0.7205045819282532,
      "learning_rate": 0.0002416549940739144,
      "loss": 3.3412,
      "step": 55550
    },
    {
      "epoch": 5.984285868044344,
      "grad_norm": 0.729873538017273,
      "learning_rate": 0.00024133175304385303,
      "loss": 3.3155,
      "step": 55600
    },
    {
      "epoch": 5.989667420083952,
      "grad_norm": 0.6925444602966309,
      "learning_rate": 0.0002410085120137916,
      "loss": 3.3504,
      "step": 55650
    },
    {
      "epoch": 5.995048972123561,
      "grad_norm": 0.7201747298240662,
      "learning_rate": 0.00024068527098373017,
      "loss": 3.3241,
      "step": 55700
    },
    {
      "epoch": 6.000430524163169,
      "grad_norm": 0.7336894273757935,
      "learning_rate": 0.0002403620299536688,
      "loss": 3.3297,
      "step": 55750
    },
    {
      "epoch": 6.005812076202777,
      "grad_norm": 0.6997312307357788,
      "learning_rate": 0.00024003878892360736,
      "loss": 3.232,
      "step": 55800
    },
    {
      "epoch": 6.011193628242385,
      "grad_norm": 0.7346082329750061,
      "learning_rate": 0.00023971554789354593,
      "loss": 3.2338,
      "step": 55850
    },
    {
      "epoch": 6.016575180281993,
      "grad_norm": 0.7229931950569153,
      "learning_rate": 0.00023939230686348452,
      "loss": 3.2316,
      "step": 55900
    },
    {
      "epoch": 6.021956732321602,
      "grad_norm": 0.7411378622055054,
      "learning_rate": 0.0002390690658334231,
      "loss": 3.2539,
      "step": 55950
    },
    {
      "epoch": 6.0273382843612096,
      "grad_norm": 0.6895947456359863,
      "learning_rate": 0.00023874582480336168,
      "loss": 3.2608,
      "step": 56000
    },
    {
      "epoch": 6.0273382843612096,
      "eval_accuracy": 0.3839755026853018,
      "eval_loss": 3.3861966133117676,
      "eval_runtime": 211.6559,
      "eval_samples_per_second": 85.096,
      "eval_steps_per_second": 5.32,
      "step": 56000
    },
    {
      "epoch": 6.032719836400818,
      "grad_norm": 0.7139065265655518,
      "learning_rate": 0.00023842258377330028,
      "loss": 3.2516,
      "step": 56050
    },
    {
      "epoch": 6.038101388440427,
      "grad_norm": 0.7533993124961853,
      "learning_rate": 0.00023810580756384006,
      "loss": 3.2595,
      "step": 56100
    },
    {
      "epoch": 6.043482940480034,
      "grad_norm": 0.7737258076667786,
      "learning_rate": 0.00023778256653377868,
      "loss": 3.2565,
      "step": 56150
    },
    {
      "epoch": 6.048864492519643,
      "grad_norm": 0.709679126739502,
      "learning_rate": 0.00023745932550371725,
      "loss": 3.2624,
      "step": 56200
    },
    {
      "epoch": 6.0542460445592505,
      "grad_norm": 0.7647435069084167,
      "learning_rate": 0.00023713608447365584,
      "loss": 3.2578,
      "step": 56250
    },
    {
      "epoch": 6.059627596598859,
      "grad_norm": 0.7607208490371704,
      "learning_rate": 0.00023681284344359444,
      "loss": 3.2503,
      "step": 56300
    },
    {
      "epoch": 6.065009148638468,
      "grad_norm": 0.76082843542099,
      "learning_rate": 0.000236489602413533,
      "loss": 3.2469,
      "step": 56350
    },
    {
      "epoch": 6.070390700678075,
      "grad_norm": 0.7211427688598633,
      "learning_rate": 0.0002361663613834716,
      "loss": 3.2735,
      "step": 56400
    },
    {
      "epoch": 6.075772252717684,
      "grad_norm": 0.7562023401260376,
      "learning_rate": 0.00023584312035341017,
      "loss": 3.2647,
      "step": 56450
    },
    {
      "epoch": 6.081153804757292,
      "grad_norm": 0.7383215427398682,
      "learning_rate": 0.00023551987932334876,
      "loss": 3.2576,
      "step": 56500
    },
    {
      "epoch": 6.0865353567969,
      "grad_norm": 0.7635093331336975,
      "learning_rate": 0.00023519663829328735,
      "loss": 3.2469,
      "step": 56550
    },
    {
      "epoch": 6.091916908836509,
      "grad_norm": 0.740917980670929,
      "learning_rate": 0.00023487339726322592,
      "loss": 3.2686,
      "step": 56600
    },
    {
      "epoch": 6.097298460876116,
      "grad_norm": 0.8048155307769775,
      "learning_rate": 0.0002345501562331645,
      "loss": 3.2547,
      "step": 56650
    },
    {
      "epoch": 6.102680012915725,
      "grad_norm": 0.7598278522491455,
      "learning_rate": 0.0002342269152031031,
      "loss": 3.2496,
      "step": 56700
    },
    {
      "epoch": 6.108061564955333,
      "grad_norm": 0.7499753832817078,
      "learning_rate": 0.00023390367417304168,
      "loss": 3.2653,
      "step": 56750
    },
    {
      "epoch": 6.113443116994941,
      "grad_norm": 0.7181298136711121,
      "learning_rate": 0.00023358043314298025,
      "loss": 3.2557,
      "step": 56800
    },
    {
      "epoch": 6.11882466903455,
      "grad_norm": 0.7646988034248352,
      "learning_rate": 0.00023325719211291887,
      "loss": 3.2724,
      "step": 56850
    },
    {
      "epoch": 6.124206221074158,
      "grad_norm": 0.7216889262199402,
      "learning_rate": 0.00023293395108285744,
      "loss": 3.2523,
      "step": 56900
    },
    {
      "epoch": 6.129587773113766,
      "grad_norm": 0.7427191734313965,
      "learning_rate": 0.000232610710052796,
      "loss": 3.2762,
      "step": 56950
    },
    {
      "epoch": 6.134969325153374,
      "grad_norm": 0.742785632610321,
      "learning_rate": 0.0002322874690227346,
      "loss": 3.2733,
      "step": 57000
    },
    {
      "epoch": 6.134969325153374,
      "eval_accuracy": 0.38427451543256724,
      "eval_loss": 3.3876760005950928,
      "eval_runtime": 199.7347,
      "eval_samples_per_second": 90.175,
      "eval_steps_per_second": 5.637,
      "step": 57000
    },
    {
      "epoch": 6.140350877192983,
      "grad_norm": 0.7947841882705688,
      "learning_rate": 0.0002319642279926732,
      "loss": 3.2932,
      "step": 57050
    },
    {
      "epoch": 6.1457324292325906,
      "grad_norm": 0.7083519101142883,
      "learning_rate": 0.00023164098696261176,
      "loss": 3.2631,
      "step": 57100
    },
    {
      "epoch": 6.151113981272199,
      "grad_norm": 0.7376063466072083,
      "learning_rate": 0.00023131774593255036,
      "loss": 3.2754,
      "step": 57150
    },
    {
      "epoch": 6.156495533311807,
      "grad_norm": 0.7607095241546631,
      "learning_rate": 0.00023099450490248892,
      "loss": 3.282,
      "step": 57200
    },
    {
      "epoch": 6.161877085351415,
      "grad_norm": 0.7469072341918945,
      "learning_rate": 0.00023067126387242754,
      "loss": 3.2629,
      "step": 57250
    },
    {
      "epoch": 6.167258637391024,
      "grad_norm": 0.7765781283378601,
      "learning_rate": 0.0002303480228423661,
      "loss": 3.2591,
      "step": 57300
    },
    {
      "epoch": 6.1726401894306315,
      "grad_norm": 0.7496424913406372,
      "learning_rate": 0.00023002478181230468,
      "loss": 3.2608,
      "step": 57350
    },
    {
      "epoch": 6.17802174147024,
      "grad_norm": 0.7514758706092834,
      "learning_rate": 0.00022970154078224327,
      "loss": 3.2634,
      "step": 57400
    },
    {
      "epoch": 6.183403293509849,
      "grad_norm": 0.7652694582939148,
      "learning_rate": 0.00022937829975218187,
      "loss": 3.2699,
      "step": 57450
    },
    {
      "epoch": 6.188784845549456,
      "grad_norm": 0.734973132610321,
      "learning_rate": 0.00022905505872212044,
      "loss": 3.2727,
      "step": 57500
    },
    {
      "epoch": 6.194166397589065,
      "grad_norm": 0.7655386328697205,
      "learning_rate": 0.00022873181769205903,
      "loss": 3.2568,
      "step": 57550
    },
    {
      "epoch": 6.1995479496286725,
      "grad_norm": 0.744789719581604,
      "learning_rate": 0.0002284085766619976,
      "loss": 3.2845,
      "step": 57600
    },
    {
      "epoch": 6.204929501668281,
      "grad_norm": 0.7866900563240051,
      "learning_rate": 0.0002280853356319362,
      "loss": 3.2839,
      "step": 57650
    },
    {
      "epoch": 6.21031105370789,
      "grad_norm": 0.753262460231781,
      "learning_rate": 0.0002277620946018748,
      "loss": 3.2868,
      "step": 57700
    },
    {
      "epoch": 6.215692605747497,
      "grad_norm": 0.7368378639221191,
      "learning_rate": 0.00022743885357181336,
      "loss": 3.2648,
      "step": 57750
    },
    {
      "epoch": 6.221074157787106,
      "grad_norm": 0.7708497643470764,
      "learning_rate": 0.00022711561254175192,
      "loss": 3.2733,
      "step": 57800
    },
    {
      "epoch": 6.226455709826714,
      "grad_norm": 0.7500898241996765,
      "learning_rate": 0.00022679237151169054,
      "loss": 3.2719,
      "step": 57850
    },
    {
      "epoch": 6.231837261866322,
      "grad_norm": 0.7447527647018433,
      "learning_rate": 0.0002264691304816291,
      "loss": 3.2606,
      "step": 57900
    },
    {
      "epoch": 6.237218813905931,
      "grad_norm": 0.7365005016326904,
      "learning_rate": 0.00022614588945156768,
      "loss": 3.2832,
      "step": 57950
    },
    {
      "epoch": 6.242600365945538,
      "grad_norm": 0.7803393006324768,
      "learning_rate": 0.0002258226484215063,
      "loss": 3.2897,
      "step": 58000
    },
    {
      "epoch": 6.242600365945538,
      "eval_accuracy": 0.38468098588588123,
      "eval_loss": 3.381647825241089,
      "eval_runtime": 201.6412,
      "eval_samples_per_second": 89.322,
      "eval_steps_per_second": 5.584,
      "step": 58000
    },
    {
      "epoch": 6.247981917985147,
      "grad_norm": 0.7322475910186768,
      "learning_rate": 0.00022549940739144487,
      "loss": 3.2811,
      "step": 58050
    },
    {
      "epoch": 6.253363470024755,
      "grad_norm": 0.7434014081954956,
      "learning_rate": 0.00022517616636138344,
      "loss": 3.2893,
      "step": 58100
    },
    {
      "epoch": 6.258745022064363,
      "grad_norm": 0.7110326290130615,
      "learning_rate": 0.00022485292533132203,
      "loss": 3.2929,
      "step": 58150
    },
    {
      "epoch": 6.264126574103972,
      "grad_norm": 0.7830908298492432,
      "learning_rate": 0.00022452968430126063,
      "loss": 3.2784,
      "step": 58200
    },
    {
      "epoch": 6.26950812614358,
      "grad_norm": 0.7722041606903076,
      "learning_rate": 0.00022420644327119922,
      "loss": 3.2607,
      "step": 58250
    },
    {
      "epoch": 6.274889678183188,
      "grad_norm": 0.7421392202377319,
      "learning_rate": 0.0002238832022411378,
      "loss": 3.2634,
      "step": 58300
    },
    {
      "epoch": 6.280271230222796,
      "grad_norm": 0.7481209635734558,
      "learning_rate": 0.00022355996121107636,
      "loss": 3.2692,
      "step": 58350
    },
    {
      "epoch": 6.285652782262405,
      "grad_norm": 0.7708833813667297,
      "learning_rate": 0.00022323672018101498,
      "loss": 3.2706,
      "step": 58400
    },
    {
      "epoch": 6.2910343343020125,
      "grad_norm": 0.8033352494239807,
      "learning_rate": 0.00022291347915095355,
      "loss": 3.2784,
      "step": 58450
    },
    {
      "epoch": 6.296415886341621,
      "grad_norm": 0.7255500555038452,
      "learning_rate": 0.0002225902381208921,
      "loss": 3.2862,
      "step": 58500
    },
    {
      "epoch": 6.301797438381229,
      "grad_norm": 0.7180657386779785,
      "learning_rate": 0.00022226699709083073,
      "loss": 3.2808,
      "step": 58550
    },
    {
      "epoch": 6.307178990420837,
      "grad_norm": 0.8310112953186035,
      "learning_rate": 0.0002219437560607693,
      "loss": 3.2954,
      "step": 58600
    },
    {
      "epoch": 6.312560542460446,
      "grad_norm": 0.7167147994041443,
      "learning_rate": 0.00022162051503070787,
      "loss": 3.2825,
      "step": 58650
    },
    {
      "epoch": 6.3179420945000535,
      "grad_norm": 0.8282204866409302,
      "learning_rate": 0.00022129727400064646,
      "loss": 3.2813,
      "step": 58700
    },
    {
      "epoch": 6.323323646539662,
      "grad_norm": 0.7656278014183044,
      "learning_rate": 0.00022097403297058506,
      "loss": 3.282,
      "step": 58750
    },
    {
      "epoch": 6.328705198579271,
      "grad_norm": 0.729318380355835,
      "learning_rate": 0.00022065079194052363,
      "loss": 3.2716,
      "step": 58800
    },
    {
      "epoch": 6.334086750618878,
      "grad_norm": 0.7712202072143555,
      "learning_rate": 0.00022032755091046222,
      "loss": 3.2681,
      "step": 58850
    },
    {
      "epoch": 6.339468302658487,
      "grad_norm": 0.7325397729873657,
      "learning_rate": 0.0002200043098804008,
      "loss": 3.2774,
      "step": 58900
    },
    {
      "epoch": 6.344849854698095,
      "grad_norm": 0.7850790619850159,
      "learning_rate": 0.00021968106885033938,
      "loss": 3.2693,
      "step": 58950
    },
    {
      "epoch": 6.350231406737703,
      "grad_norm": 0.7271026968955994,
      "learning_rate": 0.00021935782782027798,
      "loss": 3.2645,
      "step": 59000
    },
    {
      "epoch": 6.350231406737703,
      "eval_accuracy": 0.384902529124135,
      "eval_loss": 3.378819704055786,
      "eval_runtime": 223.2378,
      "eval_samples_per_second": 80.681,
      "eval_steps_per_second": 5.044,
      "step": 59000
    },
    {
      "epoch": 6.355612958777312,
      "grad_norm": 0.7542627453804016,
      "learning_rate": 0.00021903458679021655,
      "loss": 3.2872,
      "step": 59050
    },
    {
      "epoch": 6.360994510816919,
      "grad_norm": 0.7405682802200317,
      "learning_rate": 0.0002187113457601551,
      "loss": 3.266,
      "step": 59100
    },
    {
      "epoch": 6.366376062856528,
      "grad_norm": 0.7460401654243469,
      "learning_rate": 0.00021838810473009373,
      "loss": 3.2759,
      "step": 59150
    },
    {
      "epoch": 6.371757614896136,
      "grad_norm": 0.7826977372169495,
      "learning_rate": 0.0002180648637000323,
      "loss": 3.2753,
      "step": 59200
    },
    {
      "epoch": 6.377139166935744,
      "grad_norm": 0.7382274270057678,
      "learning_rate": 0.0002177416226699709,
      "loss": 3.2868,
      "step": 59250
    },
    {
      "epoch": 6.382520718975353,
      "grad_norm": 0.7684206962585449,
      "learning_rate": 0.00021741838163990946,
      "loss": 3.2841,
      "step": 59300
    },
    {
      "epoch": 6.387902271014961,
      "grad_norm": 0.7826603055000305,
      "learning_rate": 0.00021709514060984806,
      "loss": 3.2647,
      "step": 59350
    },
    {
      "epoch": 6.393283823054569,
      "grad_norm": 0.760877788066864,
      "learning_rate": 0.00021677189957978665,
      "loss": 3.2792,
      "step": 59400
    },
    {
      "epoch": 6.398665375094177,
      "grad_norm": 0.7618247866630554,
      "learning_rate": 0.00021644865854972522,
      "loss": 3.2909,
      "step": 59450
    },
    {
      "epoch": 6.404046927133785,
      "grad_norm": 0.9166077971458435,
      "learning_rate": 0.0002161254175196638,
      "loss": 3.2866,
      "step": 59500
    },
    {
      "epoch": 6.4094284791733935,
      "grad_norm": 0.764710009098053,
      "learning_rate": 0.0002158021764896024,
      "loss": 3.2599,
      "step": 59550
    },
    {
      "epoch": 6.414810031213002,
      "grad_norm": 0.7640218734741211,
      "learning_rate": 0.00021547893545954098,
      "loss": 3.2676,
      "step": 59600
    },
    {
      "epoch": 6.42019158325261,
      "grad_norm": 0.7581861615180969,
      "learning_rate": 0.00021515569442947955,
      "loss": 3.3134,
      "step": 59650
    },
    {
      "epoch": 6.425573135292218,
      "grad_norm": 0.7885519862174988,
      "learning_rate": 0.00021483245339941817,
      "loss": 3.2905,
      "step": 59700
    },
    {
      "epoch": 6.430954687331827,
      "grad_norm": 0.7909292578697205,
      "learning_rate": 0.00021450921236935674,
      "loss": 3.2853,
      "step": 59750
    },
    {
      "epoch": 6.4363362393714345,
      "grad_norm": 0.7698927521705627,
      "learning_rate": 0.0002141859713392953,
      "loss": 3.2799,
      "step": 59800
    },
    {
      "epoch": 6.441717791411043,
      "grad_norm": 0.7979402542114258,
      "learning_rate": 0.0002138627303092339,
      "loss": 3.3018,
      "step": 59850
    },
    {
      "epoch": 6.447099343450651,
      "grad_norm": 0.7837560772895813,
      "learning_rate": 0.0002135394892791725,
      "loss": 3.2739,
      "step": 59900
    },
    {
      "epoch": 6.452480895490259,
      "grad_norm": 0.7438477873802185,
      "learning_rate": 0.00021321624824911106,
      "loss": 3.2845,
      "step": 59950
    },
    {
      "epoch": 6.457862447529868,
      "grad_norm": 0.7494663596153259,
      "learning_rate": 0.00021289300721904965,
      "loss": 3.2801,
      "step": 60000
    },
    {
      "epoch": 6.457862447529868,
      "eval_accuracy": 0.38531291108141025,
      "eval_loss": 3.374072313308716,
      "eval_runtime": 209.5252,
      "eval_samples_per_second": 85.961,
      "eval_steps_per_second": 5.374,
      "step": 60000
    },
    {
      "epoch": 6.4632439995694755,
      "grad_norm": 0.7274054288864136,
      "learning_rate": 0.00021256976618898822,
      "loss": 3.2882,
      "step": 60050
    },
    {
      "epoch": 6.468625551609084,
      "grad_norm": 0.7407465577125549,
      "learning_rate": 0.00021225298997952806,
      "loss": 3.2824,
      "step": 60100
    },
    {
      "epoch": 6.474007103648693,
      "grad_norm": 0.7144964933395386,
      "learning_rate": 0.00021192974894946662,
      "loss": 3.2957,
      "step": 60150
    },
    {
      "epoch": 6.4793886556883,
      "grad_norm": 0.7461879849433899,
      "learning_rate": 0.00021160650791940524,
      "loss": 3.2838,
      "step": 60200
    },
    {
      "epoch": 6.484770207727909,
      "grad_norm": 0.755248486995697,
      "learning_rate": 0.0002112832668893438,
      "loss": 3.2862,
      "step": 60250
    },
    {
      "epoch": 6.490151759767517,
      "grad_norm": 0.7556350231170654,
      "learning_rate": 0.00021096002585928238,
      "loss": 3.2679,
      "step": 60300
    },
    {
      "epoch": 6.495533311807125,
      "grad_norm": 0.7320284843444824,
      "learning_rate": 0.00021063678482922097,
      "loss": 3.2899,
      "step": 60350
    },
    {
      "epoch": 6.500914863846734,
      "grad_norm": 0.7573622465133667,
      "learning_rate": 0.00021031354379915957,
      "loss": 3.2869,
      "step": 60400
    },
    {
      "epoch": 6.506296415886341,
      "grad_norm": 0.7960693836212158,
      "learning_rate": 0.00020999030276909814,
      "loss": 3.2784,
      "step": 60450
    },
    {
      "epoch": 6.51167796792595,
      "grad_norm": 0.8241477608680725,
      "learning_rate": 0.00020966706173903673,
      "loss": 3.2832,
      "step": 60500
    },
    {
      "epoch": 6.517059519965558,
      "grad_norm": 0.773271381855011,
      "learning_rate": 0.0002093438207089753,
      "loss": 3.2755,
      "step": 60550
    },
    {
      "epoch": 6.522441072005166,
      "grad_norm": 0.7589271664619446,
      "learning_rate": 0.00020902057967891387,
      "loss": 3.2864,
      "step": 60600
    },
    {
      "epoch": 6.5278226240447745,
      "grad_norm": 0.735211968421936,
      "learning_rate": 0.0002086973386488525,
      "loss": 3.2909,
      "step": 60650
    },
    {
      "epoch": 6.533204176084383,
      "grad_norm": 0.767329752445221,
      "learning_rate": 0.00020837409761879106,
      "loss": 3.2849,
      "step": 60700
    },
    {
      "epoch": 6.538585728123991,
      "grad_norm": 0.7533854842185974,
      "learning_rate": 0.00020805085658872962,
      "loss": 3.2829,
      "step": 60750
    },
    {
      "epoch": 6.543967280163599,
      "grad_norm": 0.8554604649543762,
      "learning_rate": 0.00020772761555866825,
      "loss": 3.284,
      "step": 60800
    },
    {
      "epoch": 6.549348832203208,
      "grad_norm": 0.7716811895370483,
      "learning_rate": 0.0002074043745286068,
      "loss": 3.2828,
      "step": 60850
    },
    {
      "epoch": 6.5547303842428155,
      "grad_norm": 0.7305426597595215,
      "learning_rate": 0.00020708113349854538,
      "loss": 3.3054,
      "step": 60900
    },
    {
      "epoch": 6.560111936282424,
      "grad_norm": 0.8045239448547363,
      "learning_rate": 0.00020675789246848397,
      "loss": 3.2869,
      "step": 60950
    },
    {
      "epoch": 6.565493488322032,
      "grad_norm": 0.7284274697303772,
      "learning_rate": 0.00020643465143842257,
      "loss": 3.2836,
      "step": 61000
    },
    {
      "epoch": 6.565493488322032,
      "eval_accuracy": 0.38564528026523476,
      "eval_loss": 3.370778799057007,
      "eval_runtime": 201.2966,
      "eval_samples_per_second": 89.475,
      "eval_steps_per_second": 5.594,
      "step": 61000
    },
    {
      "epoch": 6.57087504036164,
      "grad_norm": 0.7565922737121582,
      "learning_rate": 0.00020611141040836116,
      "loss": 3.285,
      "step": 61050
    },
    {
      "epoch": 6.576256592401249,
      "grad_norm": 0.8001964092254639,
      "learning_rate": 0.00020578816937829973,
      "loss": 3.2914,
      "step": 61100
    },
    {
      "epoch": 6.5816381444408565,
      "grad_norm": 0.7406817078590393,
      "learning_rate": 0.0002054649283482383,
      "loss": 3.2947,
      "step": 61150
    },
    {
      "epoch": 6.587019696480465,
      "grad_norm": 0.7439241409301758,
      "learning_rate": 0.00020514168731817692,
      "loss": 3.2742,
      "step": 61200
    },
    {
      "epoch": 6.592401248520073,
      "grad_norm": 0.7465095520019531,
      "learning_rate": 0.0002048184462881155,
      "loss": 3.2787,
      "step": 61250
    },
    {
      "epoch": 6.597782800559681,
      "grad_norm": 0.7967016100883484,
      "learning_rate": 0.00020449520525805406,
      "loss": 3.2923,
      "step": 61300
    },
    {
      "epoch": 6.60316435259929,
      "grad_norm": 0.7551931142807007,
      "learning_rate": 0.00020417196422799268,
      "loss": 3.2902,
      "step": 61350
    },
    {
      "epoch": 6.608545904638898,
      "grad_norm": 0.7699846029281616,
      "learning_rate": 0.00020384872319793125,
      "loss": 3.2934,
      "step": 61400
    },
    {
      "epoch": 6.613927456678506,
      "grad_norm": 0.7986684441566467,
      "learning_rate": 0.0002035254821678698,
      "loss": 3.2858,
      "step": 61450
    },
    {
      "epoch": 6.619309008718115,
      "grad_norm": 0.7962992787361145,
      "learning_rate": 0.0002032022411378084,
      "loss": 3.2827,
      "step": 61500
    },
    {
      "epoch": 6.624690560757722,
      "grad_norm": 0.7751727104187012,
      "learning_rate": 0.000202879000107747,
      "loss": 3.2885,
      "step": 61550
    },
    {
      "epoch": 6.630072112797331,
      "grad_norm": 0.7956772446632385,
      "learning_rate": 0.00020255575907768557,
      "loss": 3.3027,
      "step": 61600
    },
    {
      "epoch": 6.635453664836939,
      "grad_norm": 0.7654341459274292,
      "learning_rate": 0.00020223251804762416,
      "loss": 3.2883,
      "step": 61650
    },
    {
      "epoch": 6.640835216876547,
      "grad_norm": 0.7234878540039062,
      "learning_rate": 0.00020190927701756273,
      "loss": 3.3,
      "step": 61700
    },
    {
      "epoch": 6.6462167689161555,
      "grad_norm": 0.7403562068939209,
      "learning_rate": 0.00020158603598750133,
      "loss": 3.2763,
      "step": 61750
    },
    {
      "epoch": 6.651598320955763,
      "grad_norm": 0.7899567484855652,
      "learning_rate": 0.00020126279495743992,
      "loss": 3.2967,
      "step": 61800
    },
    {
      "epoch": 6.656979872995372,
      "grad_norm": 0.7615472674369812,
      "learning_rate": 0.0002009395539273785,
      "loss": 3.2782,
      "step": 61850
    },
    {
      "epoch": 6.66236142503498,
      "grad_norm": 0.7665838599205017,
      "learning_rate": 0.00020061631289731706,
      "loss": 3.2865,
      "step": 61900
    },
    {
      "epoch": 6.667742977074588,
      "grad_norm": 0.769633948802948,
      "learning_rate": 0.00020029307186725568,
      "loss": 3.3017,
      "step": 61950
    },
    {
      "epoch": 6.6731245291141965,
      "grad_norm": 0.7775278687477112,
      "learning_rate": 0.00019996983083719425,
      "loss": 3.2808,
      "step": 62000
    },
    {
      "epoch": 6.6731245291141965,
      "eval_accuracy": 0.3862611248333672,
      "eval_loss": 3.3647584915161133,
      "eval_runtime": 210.4271,
      "eval_samples_per_second": 85.593,
      "eval_steps_per_second": 5.351,
      "step": 62000
    },
    {
      "epoch": 6.678506081153805,
      "grad_norm": 0.7849041223526001,
      "learning_rate": 0.00019964658980713284,
      "loss": 3.2781,
      "step": 62050
    },
    {
      "epoch": 6.683887633193413,
      "grad_norm": 0.8136065602302551,
      "learning_rate": 0.00019932334877707144,
      "loss": 3.3005,
      "step": 62100
    },
    {
      "epoch": 6.689269185233021,
      "grad_norm": 0.7949984073638916,
      "learning_rate": 0.00019900657256761124,
      "loss": 3.2789,
      "step": 62150
    },
    {
      "epoch": 6.69465073727263,
      "grad_norm": 0.7674379348754883,
      "learning_rate": 0.0001986833315375498,
      "loss": 3.2832,
      "step": 62200
    },
    {
      "epoch": 6.7000322893122375,
      "grad_norm": 0.8072303533554077,
      "learning_rate": 0.00019836009050748838,
      "loss": 3.2916,
      "step": 62250
    },
    {
      "epoch": 6.705413841351846,
      "grad_norm": 0.7814190983772278,
      "learning_rate": 0.000198036849477427,
      "loss": 3.3209,
      "step": 62300
    },
    {
      "epoch": 6.710795393391454,
      "grad_norm": 0.788261890411377,
      "learning_rate": 0.00019772007326796678,
      "loss": 3.2981,
      "step": 62350
    },
    {
      "epoch": 6.716176945431062,
      "grad_norm": 0.7617114186286926,
      "learning_rate": 0.0001973968322379054,
      "loss": 3.2803,
      "step": 62400
    },
    {
      "epoch": 6.721558497470671,
      "grad_norm": 0.7262803316116333,
      "learning_rate": 0.00019707359120784397,
      "loss": 3.2805,
      "step": 62450
    },
    {
      "epoch": 6.7269400495102785,
      "grad_norm": 0.7623103260993958,
      "learning_rate": 0.00019675035017778253,
      "loss": 3.2771,
      "step": 62500
    },
    {
      "epoch": 6.732321601549887,
      "grad_norm": 0.7562897801399231,
      "learning_rate": 0.00019642710914772113,
      "loss": 3.292,
      "step": 62550
    },
    {
      "epoch": 6.737703153589496,
      "grad_norm": 0.7772451043128967,
      "learning_rate": 0.00019610386811765972,
      "loss": 3.2792,
      "step": 62600
    },
    {
      "epoch": 6.743084705629103,
      "grad_norm": 0.8128804564476013,
      "learning_rate": 0.00019578062708759832,
      "loss": 3.3054,
      "step": 62650
    },
    {
      "epoch": 6.748466257668712,
      "grad_norm": 0.8137348294258118,
      "learning_rate": 0.00019545738605753689,
      "loss": 3.2941,
      "step": 62700
    },
    {
      "epoch": 6.75384780970832,
      "grad_norm": 0.7956899404525757,
      "learning_rate": 0.00019513414502747545,
      "loss": 3.2922,
      "step": 62750
    },
    {
      "epoch": 6.759229361747928,
      "grad_norm": 0.7582367062568665,
      "learning_rate": 0.00019481090399741408,
      "loss": 3.2699,
      "step": 62800
    },
    {
      "epoch": 6.7646109137875365,
      "grad_norm": 1.1021339893341064,
      "learning_rate": 0.00019448766296735264,
      "loss": 3.2984,
      "step": 62850
    },
    {
      "epoch": 6.769992465827144,
      "grad_norm": 0.8446829319000244,
      "learning_rate": 0.0001941644219372912,
      "loss": 3.2944,
      "step": 62900
    },
    {
      "epoch": 6.775374017866753,
      "grad_norm": 0.7830290794372559,
      "learning_rate": 0.00019384118090722983,
      "loss": 3.2988,
      "step": 62950
    },
    {
      "epoch": 6.780755569906361,
      "grad_norm": 0.7751957178115845,
      "learning_rate": 0.0001935179398771684,
      "loss": 3.2776,
      "step": 63000
    },
    {
      "epoch": 6.780755569906361,
      "eval_accuracy": 0.38648049501386467,
      "eval_loss": 3.3618505001068115,
      "eval_runtime": 180.2451,
      "eval_samples_per_second": 99.925,
      "eval_steps_per_second": 6.247,
      "step": 63000
    },
    {
      "epoch": 6.786137121945969,
      "grad_norm": 0.8304117321968079,
      "learning_rate": 0.00019319469884710697,
      "loss": 3.2927,
      "step": 63050
    },
    {
      "epoch": 6.7915186739855775,
      "grad_norm": 0.7430444955825806,
      "learning_rate": 0.00019287145781704556,
      "loss": 3.2934,
      "step": 63100
    },
    {
      "epoch": 6.796900226025185,
      "grad_norm": 0.756223738193512,
      "learning_rate": 0.00019254821678698416,
      "loss": 3.2953,
      "step": 63150
    },
    {
      "epoch": 6.802281778064794,
      "grad_norm": 0.7750756144523621,
      "learning_rate": 0.00019222497575692272,
      "loss": 3.2911,
      "step": 63200
    },
    {
      "epoch": 6.807663330104402,
      "grad_norm": 0.7659454941749573,
      "learning_rate": 0.00019190173472686132,
      "loss": 3.2952,
      "step": 63250
    },
    {
      "epoch": 6.813044882144011,
      "grad_norm": 0.7663837671279907,
      "learning_rate": 0.0001915784936967999,
      "loss": 3.2953,
      "step": 63300
    },
    {
      "epoch": 6.8184264341836185,
      "grad_norm": 0.7659280300140381,
      "learning_rate": 0.00019125525266673845,
      "loss": 3.2848,
      "step": 63350
    },
    {
      "epoch": 6.823807986223227,
      "grad_norm": 0.8383036255836487,
      "learning_rate": 0.00019093201163667708,
      "loss": 3.2811,
      "step": 63400
    },
    {
      "epoch": 6.829189538262835,
      "grad_norm": 0.7594733238220215,
      "learning_rate": 0.00019060877060661564,
      "loss": 3.2928,
      "step": 63450
    },
    {
      "epoch": 6.834571090302443,
      "grad_norm": 0.7703410983085632,
      "learning_rate": 0.0001902855295765542,
      "loss": 3.2949,
      "step": 63500
    },
    {
      "epoch": 6.839952642342052,
      "grad_norm": 0.7872480154037476,
      "learning_rate": 0.00018996228854649283,
      "loss": 3.2945,
      "step": 63550
    },
    {
      "epoch": 6.8453341943816595,
      "grad_norm": 0.7895832061767578,
      "learning_rate": 0.0001896390475164314,
      "loss": 3.2881,
      "step": 63600
    },
    {
      "epoch": 6.850715746421268,
      "grad_norm": 0.7981587052345276,
      "learning_rate": 0.00018931580648637,
      "loss": 3.2887,
      "step": 63650
    },
    {
      "epoch": 6.856097298460876,
      "grad_norm": 0.7759888172149658,
      "learning_rate": 0.00018899256545630856,
      "loss": 3.3,
      "step": 63700
    },
    {
      "epoch": 6.861478850500484,
      "grad_norm": 0.7870774269104004,
      "learning_rate": 0.00018866932442624716,
      "loss": 3.2849,
      "step": 63750
    },
    {
      "epoch": 6.866860402540093,
      "grad_norm": 0.8005082011222839,
      "learning_rate": 0.00018834608339618575,
      "loss": 3.2913,
      "step": 63800
    },
    {
      "epoch": 6.8722419545797,
      "grad_norm": 0.8050322532653809,
      "learning_rate": 0.00018802284236612432,
      "loss": 3.295,
      "step": 63850
    },
    {
      "epoch": 6.877623506619309,
      "grad_norm": 0.8603694438934326,
      "learning_rate": 0.0001876996013360629,
      "loss": 3.2935,
      "step": 63900
    },
    {
      "epoch": 6.8830050586589175,
      "grad_norm": 0.8038960695266724,
      "learning_rate": 0.0001873763603060015,
      "loss": 3.2684,
      "step": 63950
    },
    {
      "epoch": 6.888386610698525,
      "grad_norm": 0.8121204376220703,
      "learning_rate": 0.00018705311927594008,
      "loss": 3.2847,
      "step": 64000
    },
    {
      "epoch": 6.888386610698525,
      "eval_accuracy": 0.3866242427844432,
      "eval_loss": 3.361416816711426,
      "eval_runtime": 180.531,
      "eval_samples_per_second": 99.767,
      "eval_steps_per_second": 6.237,
      "step": 64000
    },
    {
      "epoch": 6.893768162738134,
      "grad_norm": 0.8602204322814941,
      "learning_rate": 0.00018672987824587864,
      "loss": 3.3055,
      "step": 64050
    },
    {
      "epoch": 6.899149714777742,
      "grad_norm": 0.7669270634651184,
      "learning_rate": 0.00018640663721581727,
      "loss": 3.2954,
      "step": 64100
    },
    {
      "epoch": 6.90453126681735,
      "grad_norm": 0.7600193619728088,
      "learning_rate": 0.00018608339618575583,
      "loss": 3.2762,
      "step": 64150
    },
    {
      "epoch": 6.9099128188569585,
      "grad_norm": 0.7429041862487793,
      "learning_rate": 0.0001857601551556944,
      "loss": 3.3104,
      "step": 64200
    },
    {
      "epoch": 6.915294370896566,
      "grad_norm": 0.8116273283958435,
      "learning_rate": 0.000185436914125633,
      "loss": 3.2847,
      "step": 64250
    },
    {
      "epoch": 6.920675922936175,
      "grad_norm": 0.7676055431365967,
      "learning_rate": 0.0001851136730955716,
      "loss": 3.2907,
      "step": 64300
    },
    {
      "epoch": 6.926057474975783,
      "grad_norm": 0.7990931272506714,
      "learning_rate": 0.00018479043206551016,
      "loss": 3.2927,
      "step": 64350
    },
    {
      "epoch": 6.931439027015391,
      "grad_norm": 0.7669125199317932,
      "learning_rate": 0.00018446719103544875,
      "loss": 3.2772,
      "step": 64400
    },
    {
      "epoch": 6.9368205790549995,
      "grad_norm": 0.8254827260971069,
      "learning_rate": 0.00018414395000538732,
      "loss": 3.2951,
      "step": 64450
    },
    {
      "epoch": 6.942202131094608,
      "grad_norm": 0.7549811601638794,
      "learning_rate": 0.00018382070897532591,
      "loss": 3.29,
      "step": 64500
    },
    {
      "epoch": 6.947583683134216,
      "grad_norm": 0.8003718256950378,
      "learning_rate": 0.0001834974679452645,
      "loss": 3.2858,
      "step": 64550
    },
    {
      "epoch": 6.952965235173824,
      "grad_norm": 0.8304224014282227,
      "learning_rate": 0.00018318069173580434,
      "loss": 3.2969,
      "step": 64600
    },
    {
      "epoch": 6.958346787213433,
      "grad_norm": 0.7723833322525024,
      "learning_rate": 0.0001828574507057429,
      "loss": 3.2879,
      "step": 64650
    },
    {
      "epoch": 6.9637283392530405,
      "grad_norm": 0.7792532444000244,
      "learning_rate": 0.00018253420967568148,
      "loss": 3.2591,
      "step": 64700
    },
    {
      "epoch": 6.969109891292649,
      "grad_norm": 0.8019757270812988,
      "learning_rate": 0.00018221096864562007,
      "loss": 3.2834,
      "step": 64750
    },
    {
      "epoch": 6.974491443332257,
      "grad_norm": 0.7884219884872437,
      "learning_rate": 0.00018188772761555867,
      "loss": 3.2763,
      "step": 64800
    },
    {
      "epoch": 6.979872995371865,
      "grad_norm": 0.7742852568626404,
      "learning_rate": 0.00018156448658549723,
      "loss": 3.2954,
      "step": 64850
    },
    {
      "epoch": 6.985254547411474,
      "grad_norm": 0.7695539593696594,
      "learning_rate": 0.00018124124555543583,
      "loss": 3.2884,
      "step": 64900
    },
    {
      "epoch": 6.990636099451081,
      "grad_norm": 0.8002060055732727,
      "learning_rate": 0.0001809180045253744,
      "loss": 3.2998,
      "step": 64950
    },
    {
      "epoch": 6.99601765149069,
      "grad_norm": 0.7580960392951965,
      "learning_rate": 0.00018059476349531296,
      "loss": 3.2856,
      "step": 65000
    },
    {
      "epoch": 6.99601765149069,
      "eval_accuracy": 0.3874277308898313,
      "eval_loss": 3.353395938873291,
      "eval_runtime": 180.2586,
      "eval_samples_per_second": 99.918,
      "eval_steps_per_second": 6.247,
      "step": 65000
    },
    {
      "epoch": 7.0013992035302985,
      "grad_norm": 0.7535995244979858,
      "learning_rate": 0.00018027152246525159,
      "loss": 3.2655,
      "step": 65050
    },
    {
      "epoch": 7.006780755569906,
      "grad_norm": 0.7798009514808655,
      "learning_rate": 0.00017994828143519015,
      "loss": 3.2142,
      "step": 65100
    },
    {
      "epoch": 7.012162307609515,
      "grad_norm": 0.7834371328353882,
      "learning_rate": 0.00017962504040512872,
      "loss": 3.1959,
      "step": 65150
    },
    {
      "epoch": 7.017543859649122,
      "grad_norm": 0.7991567850112915,
      "learning_rate": 0.00017930179937506734,
      "loss": 3.1903,
      "step": 65200
    },
    {
      "epoch": 7.022925411688731,
      "grad_norm": 0.8215781450271606,
      "learning_rate": 0.0001789785583450059,
      "loss": 3.2197,
      "step": 65250
    },
    {
      "epoch": 7.0283069637283395,
      "grad_norm": 0.7673898935317993,
      "learning_rate": 0.00017865531731494448,
      "loss": 3.198,
      "step": 65300
    },
    {
      "epoch": 7.033688515767947,
      "grad_norm": 0.7570087313652039,
      "learning_rate": 0.00017833207628488307,
      "loss": 3.1956,
      "step": 65350
    },
    {
      "epoch": 7.039070067807556,
      "grad_norm": 0.7777502536773682,
      "learning_rate": 0.00017800883525482167,
      "loss": 3.2017,
      "step": 65400
    },
    {
      "epoch": 7.044451619847164,
      "grad_norm": 0.8165860176086426,
      "learning_rate": 0.00017768559422476026,
      "loss": 3.207,
      "step": 65450
    },
    {
      "epoch": 7.049833171886772,
      "grad_norm": 0.8510911464691162,
      "learning_rate": 0.00017736235319469883,
      "loss": 3.2091,
      "step": 65500
    },
    {
      "epoch": 7.0552147239263805,
      "grad_norm": 0.7689694762229919,
      "learning_rate": 0.0001770391121646374,
      "loss": 3.2019,
      "step": 65550
    },
    {
      "epoch": 7.060596275965988,
      "grad_norm": 0.7681092023849487,
      "learning_rate": 0.00017671587113457602,
      "loss": 3.2155,
      "step": 65600
    },
    {
      "epoch": 7.065977828005597,
      "grad_norm": 0.7637883424758911,
      "learning_rate": 0.0001763926301045146,
      "loss": 3.2149,
      "step": 65650
    },
    {
      "epoch": 7.071359380045205,
      "grad_norm": 0.810635507106781,
      "learning_rate": 0.00017606938907445315,
      "loss": 3.2153,
      "step": 65700
    },
    {
      "epoch": 7.076740932084813,
      "grad_norm": 0.7918748259544373,
      "learning_rate": 0.00017574614804439178,
      "loss": 3.2095,
      "step": 65750
    },
    {
      "epoch": 7.0821224841244215,
      "grad_norm": 0.8016531467437744,
      "learning_rate": 0.00017542290701433034,
      "loss": 3.211,
      "step": 65800
    },
    {
      "epoch": 7.08750403616403,
      "grad_norm": 0.8359585404396057,
      "learning_rate": 0.0001750996659842689,
      "loss": 3.2128,
      "step": 65850
    },
    {
      "epoch": 7.092885588203638,
      "grad_norm": 0.794703483581543,
      "learning_rate": 0.0001747764249542075,
      "loss": 3.1945,
      "step": 65900
    },
    {
      "epoch": 7.098267140243246,
      "grad_norm": 0.7841687202453613,
      "learning_rate": 0.0001744531839241461,
      "loss": 3.2197,
      "step": 65950
    },
    {
      "epoch": 7.103648692282855,
      "grad_norm": 0.7183166742324829,
      "learning_rate": 0.00017412994289408467,
      "loss": 3.2254,
      "step": 66000
    },
    {
      "epoch": 7.103648692282855,
      "eval_accuracy": 0.3870331036012891,
      "eval_loss": 3.3611578941345215,
      "eval_runtime": 179.9703,
      "eval_samples_per_second": 100.078,
      "eval_steps_per_second": 6.257,
      "step": 66000
    },
    {
      "epoch": 7.109030244322462,
      "grad_norm": 0.7794191837310791,
      "learning_rate": 0.00017380670186402326,
      "loss": 3.223,
      "step": 66050
    },
    {
      "epoch": 7.114411796362071,
      "grad_norm": 0.8234367966651917,
      "learning_rate": 0.00017348346083396183,
      "loss": 3.1997,
      "step": 66100
    },
    {
      "epoch": 7.119793348401679,
      "grad_norm": 0.8090280890464783,
      "learning_rate": 0.00017316021980390042,
      "loss": 3.2169,
      "step": 66150
    },
    {
      "epoch": 7.125174900441287,
      "grad_norm": 0.8259748816490173,
      "learning_rate": 0.00017283697877383902,
      "loss": 3.2235,
      "step": 66200
    },
    {
      "epoch": 7.130556452480896,
      "grad_norm": 0.8138856887817383,
      "learning_rate": 0.0001725137377437776,
      "loss": 3.2307,
      "step": 66250
    },
    {
      "epoch": 7.135938004520503,
      "grad_norm": 0.8415960669517517,
      "learning_rate": 0.00017219049671371615,
      "loss": 3.218,
      "step": 66300
    },
    {
      "epoch": 7.141319556560112,
      "grad_norm": 0.8021517992019653,
      "learning_rate": 0.00017186725568365478,
      "loss": 3.2237,
      "step": 66350
    },
    {
      "epoch": 7.1467011085997205,
      "grad_norm": 0.7997321486473083,
      "learning_rate": 0.00017154401465359334,
      "loss": 3.2073,
      "step": 66400
    },
    {
      "epoch": 7.152082660639328,
      "grad_norm": 0.7626723647117615,
      "learning_rate": 0.00017122077362353194,
      "loss": 3.2268,
      "step": 66450
    },
    {
      "epoch": 7.157464212678937,
      "grad_norm": 0.8034189939498901,
      "learning_rate": 0.00017089753259347053,
      "loss": 3.2221,
      "step": 66500
    },
    {
      "epoch": 7.162845764718545,
      "grad_norm": 0.8207605481147766,
      "learning_rate": 0.0001705742915634091,
      "loss": 3.2177,
      "step": 66550
    },
    {
      "epoch": 7.168227316758153,
      "grad_norm": 0.8149417042732239,
      "learning_rate": 0.0001702510505333477,
      "loss": 3.221,
      "step": 66600
    },
    {
      "epoch": 7.1736088687977615,
      "grad_norm": 0.8007292151451111,
      "learning_rate": 0.00016992780950328626,
      "loss": 3.2226,
      "step": 66650
    },
    {
      "epoch": 7.178990420837369,
      "grad_norm": 0.7681205868721008,
      "learning_rate": 0.00016960456847322486,
      "loss": 3.2232,
      "step": 66700
    },
    {
      "epoch": 7.184371972876978,
      "grad_norm": 0.785612940788269,
      "learning_rate": 0.00016928132744316345,
      "loss": 3.2141,
      "step": 66750
    },
    {
      "epoch": 7.189753524916586,
      "grad_norm": 0.7706198692321777,
      "learning_rate": 0.00016895808641310202,
      "loss": 3.2163,
      "step": 66800
    },
    {
      "epoch": 7.195135076956194,
      "grad_norm": 0.8889934420585632,
      "learning_rate": 0.0001686348453830406,
      "loss": 3.2103,
      "step": 66850
    },
    {
      "epoch": 7.2005166289958025,
      "grad_norm": 0.7743992209434509,
      "learning_rate": 0.0001683116043529792,
      "loss": 3.2334,
      "step": 66900
    },
    {
      "epoch": 7.205898181035411,
      "grad_norm": 0.7609925866127014,
      "learning_rate": 0.00016798836332291778,
      "loss": 3.2293,
      "step": 66950
    },
    {
      "epoch": 7.211279733075019,
      "grad_norm": 0.778520941734314,
      "learning_rate": 0.00016766512229285634,
      "loss": 3.2413,
      "step": 67000
    },
    {
      "epoch": 7.211279733075019,
      "eval_accuracy": 0.38746108732639034,
      "eval_loss": 3.359044075012207,
      "eval_runtime": 180.1842,
      "eval_samples_per_second": 99.959,
      "eval_steps_per_second": 6.249,
      "step": 67000
    },
    {
      "epoch": 7.216661285114627,
      "grad_norm": 0.7969582080841064,
      "learning_rate": 0.00016734188126279494,
      "loss": 3.2454,
      "step": 67050
    },
    {
      "epoch": 7.222042837154235,
      "grad_norm": 0.7877959609031677,
      "learning_rate": 0.00016701864023273353,
      "loss": 3.2199,
      "step": 67100
    },
    {
      "epoch": 7.2274243891938434,
      "grad_norm": 0.8033312559127808,
      "learning_rate": 0.0001666953992026721,
      "loss": 3.2453,
      "step": 67150
    },
    {
      "epoch": 7.232805941233452,
      "grad_norm": 0.775666356086731,
      "learning_rate": 0.0001663721581726107,
      "loss": 3.2301,
      "step": 67200
    },
    {
      "epoch": 7.23818749327306,
      "grad_norm": 0.851057767868042,
      "learning_rate": 0.00016604891714254926,
      "loss": 3.2248,
      "step": 67250
    },
    {
      "epoch": 7.243569045312668,
      "grad_norm": 0.8318371176719666,
      "learning_rate": 0.00016572567611248786,
      "loss": 3.2132,
      "step": 67300
    },
    {
      "epoch": 7.248950597352277,
      "grad_norm": 0.817715585231781,
      "learning_rate": 0.00016540243508242645,
      "loss": 3.238,
      "step": 67350
    },
    {
      "epoch": 7.254332149391884,
      "grad_norm": 0.7780318856239319,
      "learning_rate": 0.00016507919405236502,
      "loss": 3.2238,
      "step": 67400
    },
    {
      "epoch": 7.259713701431493,
      "grad_norm": 0.7914072275161743,
      "learning_rate": 0.00016475595302230364,
      "loss": 3.2442,
      "step": 67450
    },
    {
      "epoch": 7.265095253471101,
      "grad_norm": 0.7719249129295349,
      "learning_rate": 0.0001644327119922422,
      "loss": 3.2172,
      "step": 67500
    },
    {
      "epoch": 7.270476805510709,
      "grad_norm": 0.7739912271499634,
      "learning_rate": 0.00016410947096218078,
      "loss": 3.2197,
      "step": 67550
    },
    {
      "epoch": 7.275858357550318,
      "grad_norm": 0.8210995197296143,
      "learning_rate": 0.00016378622993211937,
      "loss": 3.228,
      "step": 67600
    },
    {
      "epoch": 7.281239909589925,
      "grad_norm": 0.8359017968177795,
      "learning_rate": 0.00016346298890205797,
      "loss": 3.2404,
      "step": 67650
    },
    {
      "epoch": 7.286621461629534,
      "grad_norm": 0.8332840204238892,
      "learning_rate": 0.00016313974787199653,
      "loss": 3.2456,
      "step": 67700
    },
    {
      "epoch": 7.2920030136691425,
      "grad_norm": 0.8179646730422974,
      "learning_rate": 0.00016281650684193513,
      "loss": 3.2327,
      "step": 67750
    },
    {
      "epoch": 7.29738456570875,
      "grad_norm": 0.7919309139251709,
      "learning_rate": 0.0001624932658118737,
      "loss": 3.2263,
      "step": 67800
    },
    {
      "epoch": 7.302766117748359,
      "grad_norm": 0.8252838850021362,
      "learning_rate": 0.0001621700247818123,
      "loss": 3.2371,
      "step": 67850
    },
    {
      "epoch": 7.308147669787967,
      "grad_norm": 0.7692036032676697,
      "learning_rate": 0.00016184678375175089,
      "loss": 3.2357,
      "step": 67900
    },
    {
      "epoch": 7.313529221827575,
      "grad_norm": 0.8547636866569519,
      "learning_rate": 0.00016153000754229067,
      "loss": 3.2201,
      "step": 67950
    },
    {
      "epoch": 7.3189107738671835,
      "grad_norm": 0.8004331588745117,
      "learning_rate": 0.0001612067665122293,
      "loss": 3.2095,
      "step": 68000
    },
    {
      "epoch": 7.3189107738671835,
      "eval_accuracy": 0.38778693733694597,
      "eval_loss": 3.356480598449707,
      "eval_runtime": 180.04,
      "eval_samples_per_second": 100.039,
      "eval_steps_per_second": 6.254,
      "step": 68000
    },
    {
      "epoch": 7.324292325906791,
      "grad_norm": 0.811409592628479,
      "learning_rate": 0.00016088352548216785,
      "loss": 3.2466,
      "step": 68050
    },
    {
      "epoch": 7.3296738779464,
      "grad_norm": 0.8523727655410767,
      "learning_rate": 0.00016056028445210642,
      "loss": 3.2269,
      "step": 68100
    },
    {
      "epoch": 7.335055429986008,
      "grad_norm": 0.809126615524292,
      "learning_rate": 0.00016023704342204504,
      "loss": 3.2503,
      "step": 68150
    },
    {
      "epoch": 7.340436982025616,
      "grad_norm": 0.8085951209068298,
      "learning_rate": 0.0001599138023919836,
      "loss": 3.2459,
      "step": 68200
    },
    {
      "epoch": 7.3458185340652244,
      "grad_norm": 0.7844497561454773,
      "learning_rate": 0.0001595905613619222,
      "loss": 3.236,
      "step": 68250
    },
    {
      "epoch": 7.351200086104833,
      "grad_norm": 0.8168286681175232,
      "learning_rate": 0.00015926732033186077,
      "loss": 3.2236,
      "step": 68300
    },
    {
      "epoch": 7.356581638144441,
      "grad_norm": 0.8217577338218689,
      "learning_rate": 0.00015894407930179934,
      "loss": 3.226,
      "step": 68350
    },
    {
      "epoch": 7.361963190184049,
      "grad_norm": 0.8433511853218079,
      "learning_rate": 0.00015862083827173796,
      "loss": 3.2294,
      "step": 68400
    },
    {
      "epoch": 7.367344742223658,
      "grad_norm": 0.8781927227973938,
      "learning_rate": 0.00015829759724167653,
      "loss": 3.2264,
      "step": 68450
    },
    {
      "epoch": 7.372726294263265,
      "grad_norm": 0.8104392290115356,
      "learning_rate": 0.0001579743562116151,
      "loss": 3.2114,
      "step": 68500
    },
    {
      "epoch": 7.378107846302874,
      "grad_norm": 0.8192629218101501,
      "learning_rate": 0.00015765111518155372,
      "loss": 3.2297,
      "step": 68550
    },
    {
      "epoch": 7.383489398342482,
      "grad_norm": 0.7943025231361389,
      "learning_rate": 0.0001573278741514923,
      "loss": 3.2313,
      "step": 68600
    },
    {
      "epoch": 7.38887095038209,
      "grad_norm": 0.858521580696106,
      "learning_rate": 0.00015700463312143085,
      "loss": 3.2268,
      "step": 68650
    },
    {
      "epoch": 7.394252502421699,
      "grad_norm": 0.804550290107727,
      "learning_rate": 0.00015668139209136945,
      "loss": 3.2131,
      "step": 68700
    },
    {
      "epoch": 7.399634054461306,
      "grad_norm": 0.8339864611625671,
      "learning_rate": 0.00015635815106130804,
      "loss": 3.2387,
      "step": 68750
    },
    {
      "epoch": 7.405015606500915,
      "grad_norm": 0.7839356064796448,
      "learning_rate": 0.0001560349100312466,
      "loss": 3.2269,
      "step": 68800
    },
    {
      "epoch": 7.4103971585405235,
      "grad_norm": 0.8349343538284302,
      "learning_rate": 0.0001557116690011852,
      "loss": 3.2374,
      "step": 68850
    },
    {
      "epoch": 7.415778710580131,
      "grad_norm": 0.8796477317810059,
      "learning_rate": 0.00015538842797112377,
      "loss": 3.2449,
      "step": 68900
    },
    {
      "epoch": 7.42116026261974,
      "grad_norm": 0.8090130686759949,
      "learning_rate": 0.00015506518694106237,
      "loss": 3.2412,
      "step": 68950
    },
    {
      "epoch": 7.426541814659347,
      "grad_norm": 0.8464242219924927,
      "learning_rate": 0.00015474194591100096,
      "loss": 3.2314,
      "step": 69000
    },
    {
      "epoch": 7.426541814659347,
      "eval_accuracy": 0.38814625243694845,
      "eval_loss": 3.3514676094055176,
      "eval_runtime": 180.0963,
      "eval_samples_per_second": 100.008,
      "eval_steps_per_second": 6.252,
      "step": 69000
    },
    {
      "epoch": 7.431923366698956,
      "grad_norm": 0.848106324672699,
      "learning_rate": 0.00015441870488093953,
      "loss": 3.2432,
      "step": 69050
    },
    {
      "epoch": 7.4373049187385645,
      "grad_norm": 0.8055359125137329,
      "learning_rate": 0.0001540954638508781,
      "loss": 3.2373,
      "step": 69100
    },
    {
      "epoch": 7.442686470778172,
      "grad_norm": 0.8940873742103577,
      "learning_rate": 0.00015377222282081672,
      "loss": 3.2389,
      "step": 69150
    },
    {
      "epoch": 7.448068022817781,
      "grad_norm": 0.7894365787506104,
      "learning_rate": 0.0001534489817907553,
      "loss": 3.237,
      "step": 69200
    },
    {
      "epoch": 7.453449574857389,
      "grad_norm": 0.8256045579910278,
      "learning_rate": 0.00015312574076069388,
      "loss": 3.2281,
      "step": 69250
    },
    {
      "epoch": 7.458831126896997,
      "grad_norm": 0.7994895577430725,
      "learning_rate": 0.00015280249973063248,
      "loss": 3.2525,
      "step": 69300
    },
    {
      "epoch": 7.4642126789366054,
      "grad_norm": 0.8331586718559265,
      "learning_rate": 0.00015247925870057104,
      "loss": 3.2337,
      "step": 69350
    },
    {
      "epoch": 7.469594230976213,
      "grad_norm": 0.7736141085624695,
      "learning_rate": 0.00015215601767050964,
      "loss": 3.2536,
      "step": 69400
    },
    {
      "epoch": 7.474975783015822,
      "grad_norm": 0.8336132764816284,
      "learning_rate": 0.0001518327766404482,
      "loss": 3.2434,
      "step": 69450
    },
    {
      "epoch": 7.48035733505543,
      "grad_norm": 0.8452787399291992,
      "learning_rate": 0.0001515095356103868,
      "loss": 3.2308,
      "step": 69500
    },
    {
      "epoch": 7.485738887095038,
      "grad_norm": 0.7975885272026062,
      "learning_rate": 0.0001511862945803254,
      "loss": 3.2414,
      "step": 69550
    },
    {
      "epoch": 7.491120439134646,
      "grad_norm": 0.7722816467285156,
      "learning_rate": 0.00015086305355026396,
      "loss": 3.2329,
      "step": 69600
    },
    {
      "epoch": 7.496501991174255,
      "grad_norm": 0.7675669193267822,
      "learning_rate": 0.00015053981252020253,
      "loss": 3.2375,
      "step": 69650
    },
    {
      "epoch": 7.501883543213863,
      "grad_norm": 0.7885386347770691,
      "learning_rate": 0.00015021657149014115,
      "loss": 3.2363,
      "step": 69700
    },
    {
      "epoch": 7.507265095253471,
      "grad_norm": 0.792849063873291,
      "learning_rate": 0.00014989333046007972,
      "loss": 3.2484,
      "step": 69750
    },
    {
      "epoch": 7.51264664729308,
      "grad_norm": 0.837902843952179,
      "learning_rate": 0.00014957008943001832,
      "loss": 3.2335,
      "step": 69800
    },
    {
      "epoch": 7.518028199332687,
      "grad_norm": 0.8501192331314087,
      "learning_rate": 0.00014924684839995688,
      "loss": 3.2218,
      "step": 69850
    },
    {
      "epoch": 7.523409751372296,
      "grad_norm": 0.8147408962249756,
      "learning_rate": 0.00014892360736989548,
      "loss": 3.2178,
      "step": 69900
    },
    {
      "epoch": 7.528791303411904,
      "grad_norm": 0.7939187288284302,
      "learning_rate": 0.00014860036633983407,
      "loss": 3.2485,
      "step": 69950
    },
    {
      "epoch": 7.534172855451512,
      "grad_norm": 0.846481204032898,
      "learning_rate": 0.00014827712530977264,
      "loss": 3.2351,
      "step": 70000
    },
    {
      "epoch": 7.534172855451512,
      "eval_accuracy": 0.38860226860710584,
      "eval_loss": 3.34824800491333,
      "eval_runtime": 180.1854,
      "eval_samples_per_second": 99.958,
      "eval_steps_per_second": 6.249,
      "step": 70000
    },
    {
      "epoch": 7.539554407491121,
      "grad_norm": 0.822512686252594,
      "learning_rate": 0.00014795388427971123,
      "loss": 3.2308,
      "step": 70050
    },
    {
      "epoch": 7.544935959530728,
      "grad_norm": 0.8041254878044128,
      "learning_rate": 0.0001476306432496498,
      "loss": 3.2348,
      "step": 70100
    },
    {
      "epoch": 7.550317511570337,
      "grad_norm": 0.8251069784164429,
      "learning_rate": 0.0001473074022195884,
      "loss": 3.2205,
      "step": 70150
    },
    {
      "epoch": 7.5556990636099455,
      "grad_norm": 0.7714356184005737,
      "learning_rate": 0.00014698416118952696,
      "loss": 3.2205,
      "step": 70200
    },
    {
      "epoch": 7.561080615649553,
      "grad_norm": 0.8164318203926086,
      "learning_rate": 0.00014666092015946556,
      "loss": 3.2301,
      "step": 70250
    },
    {
      "epoch": 7.566462167689162,
      "grad_norm": 0.8093104958534241,
      "learning_rate": 0.00014633767912940415,
      "loss": 3.2324,
      "step": 70300
    },
    {
      "epoch": 7.57184371972877,
      "grad_norm": 0.822935938835144,
      "learning_rate": 0.00014601443809934272,
      "loss": 3.2378,
      "step": 70350
    },
    {
      "epoch": 7.577225271768378,
      "grad_norm": 0.8421322703361511,
      "learning_rate": 0.00014569119706928132,
      "loss": 3.2322,
      "step": 70400
    },
    {
      "epoch": 7.5826068238079865,
      "grad_norm": 0.7766196727752686,
      "learning_rate": 0.0001453679560392199,
      "loss": 3.245,
      "step": 70450
    },
    {
      "epoch": 7.587988375847594,
      "grad_norm": 0.8198171854019165,
      "learning_rate": 0.00014504471500915848,
      "loss": 3.2239,
      "step": 70500
    },
    {
      "epoch": 7.593369927887203,
      "grad_norm": 0.7970831394195557,
      "learning_rate": 0.00014472147397909707,
      "loss": 3.2244,
      "step": 70550
    },
    {
      "epoch": 7.598751479926811,
      "grad_norm": 0.8099634051322937,
      "learning_rate": 0.00014439823294903564,
      "loss": 3.239,
      "step": 70600
    },
    {
      "epoch": 7.604133031966419,
      "grad_norm": 0.7905451655387878,
      "learning_rate": 0.00014407499191897423,
      "loss": 3.2474,
      "step": 70650
    },
    {
      "epoch": 7.609514584006027,
      "grad_norm": 0.8306140303611755,
      "learning_rate": 0.0001437517508889128,
      "loss": 3.2549,
      "step": 70700
    },
    {
      "epoch": 7.614896136045635,
      "grad_norm": 0.8547305464744568,
      "learning_rate": 0.0001434285098588514,
      "loss": 3.245,
      "step": 70750
    },
    {
      "epoch": 7.620277688085244,
      "grad_norm": 0.8246287107467651,
      "learning_rate": 0.00014311173364939123,
      "loss": 3.236,
      "step": 70800
    },
    {
      "epoch": 7.625659240124852,
      "grad_norm": 0.7652173638343811,
      "learning_rate": 0.0001427884926193298,
      "loss": 3.2216,
      "step": 70850
    },
    {
      "epoch": 7.63104079216446,
      "grad_norm": 0.8844686150550842,
      "learning_rate": 0.0001424652515892684,
      "loss": 3.2323,
      "step": 70900
    },
    {
      "epoch": 7.636422344204068,
      "grad_norm": 0.8048098087310791,
      "learning_rate": 0.000142142010559207,
      "loss": 3.2566,
      "step": 70950
    },
    {
      "epoch": 7.641803896243677,
      "grad_norm": 0.7921103835105896,
      "learning_rate": 0.00014181876952914555,
      "loss": 3.2345,
      "step": 71000
    },
    {
      "epoch": 7.641803896243677,
      "eval_accuracy": 0.3890942488831298,
      "eval_loss": 3.343825101852417,
      "eval_runtime": 180.1847,
      "eval_samples_per_second": 99.959,
      "eval_steps_per_second": 6.249,
      "step": 71000
    },
    {
      "epoch": 7.647185448283285,
      "grad_norm": 0.799433171749115,
      "learning_rate": 0.00014149552849908415,
      "loss": 3.2373,
      "step": 71050
    },
    {
      "epoch": 7.652567000322893,
      "grad_norm": 0.8794057369232178,
      "learning_rate": 0.00014117228746902272,
      "loss": 3.2458,
      "step": 71100
    },
    {
      "epoch": 7.657948552362502,
      "grad_norm": 0.8924712538719177,
      "learning_rate": 0.0001408490464389613,
      "loss": 3.2278,
      "step": 71150
    },
    {
      "epoch": 7.663330104402109,
      "grad_norm": 0.8285935521125793,
      "learning_rate": 0.00014052580540889988,
      "loss": 3.2317,
      "step": 71200
    },
    {
      "epoch": 7.668711656441718,
      "grad_norm": 0.8792213201522827,
      "learning_rate": 0.00014020256437883847,
      "loss": 3.2425,
      "step": 71250
    },
    {
      "epoch": 7.674093208481326,
      "grad_norm": 0.8292412161827087,
      "learning_rate": 0.00013987932334877707,
      "loss": 3.2234,
      "step": 71300
    },
    {
      "epoch": 7.679474760520934,
      "grad_norm": 0.7959591150283813,
      "learning_rate": 0.00013955608231871564,
      "loss": 3.2474,
      "step": 71350
    },
    {
      "epoch": 7.684856312560543,
      "grad_norm": 0.821670651435852,
      "learning_rate": 0.00013923284128865423,
      "loss": 3.2452,
      "step": 71400
    },
    {
      "epoch": 7.69023786460015,
      "grad_norm": 0.8048251271247864,
      "learning_rate": 0.00013890960025859283,
      "loss": 3.2417,
      "step": 71450
    },
    {
      "epoch": 7.695619416639759,
      "grad_norm": 0.846571147441864,
      "learning_rate": 0.0001385863592285314,
      "loss": 3.2356,
      "step": 71500
    },
    {
      "epoch": 7.7010009686793675,
      "grad_norm": 0.8968475461006165,
      "learning_rate": 0.00013826311819847,
      "loss": 3.2225,
      "step": 71550
    },
    {
      "epoch": 7.706382520718975,
      "grad_norm": 0.8345704674720764,
      "learning_rate": 0.00013793987716840858,
      "loss": 3.2244,
      "step": 71600
    },
    {
      "epoch": 7.711764072758584,
      "grad_norm": 0.7907124757766724,
      "learning_rate": 0.00013761663613834715,
      "loss": 3.2141,
      "step": 71650
    },
    {
      "epoch": 7.717145624798192,
      "grad_norm": 0.8455408811569214,
      "learning_rate": 0.00013729339510828572,
      "loss": 3.2401,
      "step": 71700
    },
    {
      "epoch": 7.7225271768378,
      "grad_norm": 0.8723199963569641,
      "learning_rate": 0.0001369701540782243,
      "loss": 3.241,
      "step": 71750
    },
    {
      "epoch": 7.727908728877408,
      "grad_norm": 0.8454627990722656,
      "learning_rate": 0.0001366469130481629,
      "loss": 3.2365,
      "step": 71800
    },
    {
      "epoch": 7.733290280917016,
      "grad_norm": 0.8864803314208984,
      "learning_rate": 0.00013632367201810147,
      "loss": 3.2511,
      "step": 71850
    },
    {
      "epoch": 7.738671832956625,
      "grad_norm": 0.8611536026000977,
      "learning_rate": 0.00013600043098804007,
      "loss": 3.2308,
      "step": 71900
    },
    {
      "epoch": 7.744053384996233,
      "grad_norm": 0.8165392875671387,
      "learning_rate": 0.00013567718995797866,
      "loss": 3.2312,
      "step": 71950
    },
    {
      "epoch": 7.749434937035841,
      "grad_norm": 0.8212934732437134,
      "learning_rate": 0.00013535394892791723,
      "loss": 3.2414,
      "step": 72000
    },
    {
      "epoch": 7.749434937035841,
      "eval_accuracy": 0.38907805960284547,
      "eval_loss": 3.3423426151275635,
      "eval_runtime": 180.2997,
      "eval_samples_per_second": 99.895,
      "eval_steps_per_second": 6.245,
      "step": 72000
    },
    {
      "epoch": 7.754816489075449,
      "grad_norm": 0.7960253953933716,
      "learning_rate": 0.00013503070789785583,
      "loss": 3.2214,
      "step": 72050
    },
    {
      "epoch": 7.760198041115058,
      "grad_norm": 0.8525424003601074,
      "learning_rate": 0.00013470746686779442,
      "loss": 3.2315,
      "step": 72100
    },
    {
      "epoch": 7.765579593154666,
      "grad_norm": 0.7953471541404724,
      "learning_rate": 0.000134384225837733,
      "loss": 3.2473,
      "step": 72150
    },
    {
      "epoch": 7.770961145194274,
      "grad_norm": 0.8736206293106079,
      "learning_rate": 0.00013406098480767158,
      "loss": 3.241,
      "step": 72200
    },
    {
      "epoch": 7.776342697233883,
      "grad_norm": 0.794816255569458,
      "learning_rate": 0.00013373774377761015,
      "loss": 3.2199,
      "step": 72250
    },
    {
      "epoch": 7.78172424927349,
      "grad_norm": 0.8048193454742432,
      "learning_rate": 0.00013341450274754874,
      "loss": 3.2403,
      "step": 72300
    },
    {
      "epoch": 7.787105801313099,
      "grad_norm": 0.8669823408126831,
      "learning_rate": 0.0001330912617174873,
      "loss": 3.2492,
      "step": 72350
    },
    {
      "epoch": 7.792487353352707,
      "grad_norm": 0.8714507818222046,
      "learning_rate": 0.0001327680206874259,
      "loss": 3.2362,
      "step": 72400
    },
    {
      "epoch": 7.797868905392315,
      "grad_norm": 0.7854097485542297,
      "learning_rate": 0.0001324447796573645,
      "loss": 3.2506,
      "step": 72450
    },
    {
      "epoch": 7.803250457431924,
      "grad_norm": 0.8088889122009277,
      "learning_rate": 0.00013212153862730307,
      "loss": 3.2229,
      "step": 72500
    },
    {
      "epoch": 7.808632009471531,
      "grad_norm": 0.8250226378440857,
      "learning_rate": 0.00013179829759724166,
      "loss": 3.2271,
      "step": 72550
    },
    {
      "epoch": 7.81401356151114,
      "grad_norm": 0.7839990854263306,
      "learning_rate": 0.00013147505656718026,
      "loss": 3.2476,
      "step": 72600
    },
    {
      "epoch": 7.819395113550748,
      "grad_norm": 0.8592643141746521,
      "learning_rate": 0.00013115181553711883,
      "loss": 3.2563,
      "step": 72650
    },
    {
      "epoch": 7.824776665590356,
      "grad_norm": 0.834892749786377,
      "learning_rate": 0.00013082857450705742,
      "loss": 3.2486,
      "step": 72700
    },
    {
      "epoch": 7.830158217629965,
      "grad_norm": 0.8037497401237488,
      "learning_rate": 0.00013050533347699602,
      "loss": 3.2666,
      "step": 72750
    },
    {
      "epoch": 7.835539769669572,
      "grad_norm": 0.7950841188430786,
      "learning_rate": 0.00013018209244693458,
      "loss": 3.2504,
      "step": 72800
    },
    {
      "epoch": 7.840921321709181,
      "grad_norm": 0.868694007396698,
      "learning_rate": 0.00012985885141687318,
      "loss": 3.2259,
      "step": 72850
    },
    {
      "epoch": 7.846302873748789,
      "grad_norm": 0.8432836532592773,
      "learning_rate": 0.00012953561038681175,
      "loss": 3.2525,
      "step": 72900
    },
    {
      "epoch": 7.851684425788397,
      "grad_norm": 0.8124203681945801,
      "learning_rate": 0.00012921236935675034,
      "loss": 3.2448,
      "step": 72950
    },
    {
      "epoch": 7.857065977828006,
      "grad_norm": 0.8090877532958984,
      "learning_rate": 0.0001288891283266889,
      "loss": 3.2356,
      "step": 73000
    },
    {
      "epoch": 7.857065977828006,
      "eval_accuracy": 0.38977637171282914,
      "eval_loss": 3.336695671081543,
      "eval_runtime": 180.1915,
      "eval_samples_per_second": 99.955,
      "eval_steps_per_second": 6.249,
      "step": 73000
    },
    {
      "epoch": 7.862447529867614,
      "grad_norm": 0.8222204446792603,
      "learning_rate": 0.0001285658872966275,
      "loss": 3.2259,
      "step": 73050
    },
    {
      "epoch": 7.867829081907222,
      "grad_norm": 0.8175197839736938,
      "learning_rate": 0.0001282426462665661,
      "loss": 3.2454,
      "step": 73100
    },
    {
      "epoch": 7.87321063394683,
      "grad_norm": 0.7853533625602722,
      "learning_rate": 0.00012791940523650466,
      "loss": 3.2267,
      "step": 73150
    },
    {
      "epoch": 7.878592185986438,
      "grad_norm": 0.8725240230560303,
      "learning_rate": 0.00012759616420644326,
      "loss": 3.2317,
      "step": 73200
    },
    {
      "epoch": 7.883973738026047,
      "grad_norm": 0.8071550726890564,
      "learning_rate": 0.00012727292317638185,
      "loss": 3.2218,
      "step": 73250
    },
    {
      "epoch": 7.889355290065655,
      "grad_norm": 0.8475930094718933,
      "learning_rate": 0.00012694968214632045,
      "loss": 3.2259,
      "step": 73300
    },
    {
      "epoch": 7.894736842105263,
      "grad_norm": 0.8381208181381226,
      "learning_rate": 0.00012662644111625902,
      "loss": 3.2366,
      "step": 73350
    },
    {
      "epoch": 7.900118394144871,
      "grad_norm": 0.809328019618988,
      "learning_rate": 0.0001263032000861976,
      "loss": 3.2301,
      "step": 73400
    },
    {
      "epoch": 7.90549994618448,
      "grad_norm": 0.8766582608222961,
      "learning_rate": 0.00012597995905613618,
      "loss": 3.2601,
      "step": 73450
    },
    {
      "epoch": 7.910881498224088,
      "grad_norm": 0.8222476840019226,
      "learning_rate": 0.00012565671802607477,
      "loss": 3.2429,
      "step": 73500
    },
    {
      "epoch": 7.916263050263696,
      "grad_norm": 0.8451269268989563,
      "learning_rate": 0.00012533347699601334,
      "loss": 3.2484,
      "step": 73550
    },
    {
      "epoch": 7.921644602303305,
      "grad_norm": 0.8155948519706726,
      "learning_rate": 0.00012501023596595193,
      "loss": 3.2388,
      "step": 73600
    },
    {
      "epoch": 7.927026154342912,
      "grad_norm": 0.8491314053535461,
      "learning_rate": 0.0001246869949358905,
      "loss": 3.2469,
      "step": 73650
    },
    {
      "epoch": 7.932407706382521,
      "grad_norm": 0.8263399600982666,
      "learning_rate": 0.0001243637539058291,
      "loss": 3.2575,
      "step": 73700
    },
    {
      "epoch": 7.937789258422129,
      "grad_norm": 0.8386456966400146,
      "learning_rate": 0.0001240405128757677,
      "loss": 3.2509,
      "step": 73750
    },
    {
      "epoch": 7.943170810461737,
      "grad_norm": 0.7784790396690369,
      "learning_rate": 0.00012371727184570629,
      "loss": 3.2274,
      "step": 73800
    },
    {
      "epoch": 7.948552362501346,
      "grad_norm": 0.8531802892684937,
      "learning_rate": 0.00012339403081564485,
      "loss": 3.2205,
      "step": 73850
    },
    {
      "epoch": 7.953933914540953,
      "grad_norm": 0.8842747211456299,
      "learning_rate": 0.00012307078978558345,
      "loss": 3.2316,
      "step": 73900
    },
    {
      "epoch": 7.959315466580562,
      "grad_norm": 0.8715245723724365,
      "learning_rate": 0.00012274754875552202,
      "loss": 3.2406,
      "step": 73950
    },
    {
      "epoch": 7.96469701862017,
      "grad_norm": 0.8610970973968506,
      "learning_rate": 0.0001224243077254606,
      "loss": 3.2443,
      "step": 74000
    },
    {
      "epoch": 7.96469701862017,
      "eval_accuracy": 0.39020728906590135,
      "eval_loss": 3.335205554962158,
      "eval_runtime": 180.1463,
      "eval_samples_per_second": 99.98,
      "eval_steps_per_second": 6.25,
      "step": 74000
    },
    {
      "epoch": 7.970078570659778,
      "grad_norm": 0.8783667087554932,
      "learning_rate": 0.00012210106669539918,
      "loss": 3.239,
      "step": 74050
    },
    {
      "epoch": 7.975460122699387,
      "grad_norm": 0.8203068971633911,
      "learning_rate": 0.00012177782566533779,
      "loss": 3.2352,
      "step": 74100
    },
    {
      "epoch": 7.980841674738995,
      "grad_norm": 0.8619286417961121,
      "learning_rate": 0.00012145458463527635,
      "loss": 3.2383,
      "step": 74150
    },
    {
      "epoch": 7.986223226778603,
      "grad_norm": 0.8609179854393005,
      "learning_rate": 0.00012113134360521495,
      "loss": 3.248,
      "step": 74200
    },
    {
      "epoch": 7.991604778818211,
      "grad_norm": 0.8277957439422607,
      "learning_rate": 0.00012080810257515353,
      "loss": 3.2689,
      "step": 74250
    },
    {
      "epoch": 7.996986330857819,
      "grad_norm": 0.8138071894645691,
      "learning_rate": 0.00012048486154509212,
      "loss": 3.2372,
      "step": 74300
    },
    {
      "epoch": 8.002367882897428,
      "grad_norm": 0.8171045184135437,
      "learning_rate": 0.00012016162051503069,
      "loss": 3.2023,
      "step": 74350
    },
    {
      "epoch": 8.007749434937036,
      "grad_norm": 0.884177029132843,
      "learning_rate": 0.00011983837948496929,
      "loss": 3.1379,
      "step": 74400
    },
    {
      "epoch": 8.013130986976645,
      "grad_norm": 0.8275094032287598,
      "learning_rate": 0.0001195216032755091,
      "loss": 3.1535,
      "step": 74450
    },
    {
      "epoch": 8.018512539016251,
      "grad_norm": 0.8046960830688477,
      "learning_rate": 0.00011919836224544767,
      "loss": 3.1526,
      "step": 74500
    },
    {
      "epoch": 8.02389409105586,
      "grad_norm": 0.8428935408592224,
      "learning_rate": 0.00011887512121538627,
      "loss": 3.1598,
      "step": 74550
    },
    {
      "epoch": 8.029275643095469,
      "grad_norm": 0.8413782715797424,
      "learning_rate": 0.00011855188018532485,
      "loss": 3.1582,
      "step": 74600
    },
    {
      "epoch": 8.034657195135077,
      "grad_norm": 0.8058221340179443,
      "learning_rate": 0.00011822863915526343,
      "loss": 3.1563,
      "step": 74650
    },
    {
      "epoch": 8.040038747174686,
      "grad_norm": 0.8850978016853333,
      "learning_rate": 0.00011790539812520201,
      "loss": 3.1662,
      "step": 74700
    },
    {
      "epoch": 8.045420299214294,
      "grad_norm": 0.8446381688117981,
      "learning_rate": 0.00011758215709514061,
      "loss": 3.1713,
      "step": 74750
    },
    {
      "epoch": 8.050801851253901,
      "grad_norm": 0.8259122967720032,
      "learning_rate": 0.00011725891606507917,
      "loss": 3.1745,
      "step": 74800
    },
    {
      "epoch": 8.05618340329351,
      "grad_norm": 0.8464698195457458,
      "learning_rate": 0.00011693567503501777,
      "loss": 3.1503,
      "step": 74850
    },
    {
      "epoch": 8.061564955333118,
      "grad_norm": 0.841551661491394,
      "learning_rate": 0.00011661243400495635,
      "loss": 3.1633,
      "step": 74900
    },
    {
      "epoch": 8.066946507372727,
      "grad_norm": 0.8218307495117188,
      "learning_rate": 0.00011628919297489493,
      "loss": 3.1706,
      "step": 74950
    },
    {
      "epoch": 8.072328059412335,
      "grad_norm": 0.8135493993759155,
      "learning_rate": 0.00011596595194483351,
      "loss": 3.1627,
      "step": 75000
    },
    {
      "epoch": 8.072328059412335,
      "eval_accuracy": 0.3899024090626939,
      "eval_loss": 3.3387045860290527,
      "eval_runtime": 180.3237,
      "eval_samples_per_second": 99.881,
      "eval_steps_per_second": 6.244,
      "step": 75000
    },
    {
      "epoch": 8.077709611451942,
      "grad_norm": 0.8384975790977478,
      "learning_rate": 0.00011564271091477211,
      "loss": 3.1987,
      "step": 75050
    },
    {
      "epoch": 8.08309116349155,
      "grad_norm": 0.8627749681472778,
      "learning_rate": 0.0001153194698847107,
      "loss": 3.1645,
      "step": 75100
    },
    {
      "epoch": 8.088472715531159,
      "grad_norm": 0.8553587794303894,
      "learning_rate": 0.00011499622885464927,
      "loss": 3.1827,
      "step": 75150
    },
    {
      "epoch": 8.093854267570768,
      "grad_norm": 0.8322249054908752,
      "learning_rate": 0.00011467298782458786,
      "loss": 3.1613,
      "step": 75200
    },
    {
      "epoch": 8.099235819610376,
      "grad_norm": 0.858957827091217,
      "learning_rate": 0.00011434974679452645,
      "loss": 3.156,
      "step": 75250
    },
    {
      "epoch": 8.104617371649983,
      "grad_norm": 0.8273110389709473,
      "learning_rate": 0.00011402650576446503,
      "loss": 3.1704,
      "step": 75300
    },
    {
      "epoch": 8.109998923689592,
      "grad_norm": 0.829047441482544,
      "learning_rate": 0.00011370326473440361,
      "loss": 3.1814,
      "step": 75350
    },
    {
      "epoch": 8.1153804757292,
      "grad_norm": 0.8540376424789429,
      "learning_rate": 0.0001133800237043422,
      "loss": 3.172,
      "step": 75400
    },
    {
      "epoch": 8.120762027768809,
      "grad_norm": 0.8101136684417725,
      "learning_rate": 0.00011305678267428077,
      "loss": 3.17,
      "step": 75450
    },
    {
      "epoch": 8.126143579808417,
      "grad_norm": 0.8184477686882019,
      "learning_rate": 0.00011273354164421936,
      "loss": 3.1818,
      "step": 75500
    },
    {
      "epoch": 8.131525131848026,
      "grad_norm": 0.8428223133087158,
      "learning_rate": 0.00011241030061415795,
      "loss": 3.1718,
      "step": 75550
    },
    {
      "epoch": 8.136906683887632,
      "grad_norm": 0.8273821473121643,
      "learning_rate": 0.00011208705958409654,
      "loss": 3.1799,
      "step": 75600
    },
    {
      "epoch": 8.142288235927241,
      "grad_norm": 0.8538662195205688,
      "learning_rate": 0.00011176381855403511,
      "loss": 3.1665,
      "step": 75650
    },
    {
      "epoch": 8.14766978796685,
      "grad_norm": 0.8688125014305115,
      "learning_rate": 0.0001114405775239737,
      "loss": 3.17,
      "step": 75700
    },
    {
      "epoch": 8.153051340006458,
      "grad_norm": 0.8661782145500183,
      "learning_rate": 0.00011111733649391228,
      "loss": 3.174,
      "step": 75750
    },
    {
      "epoch": 8.158432892046067,
      "grad_norm": 0.8403778672218323,
      "learning_rate": 0.00011079409546385086,
      "loss": 3.1963,
      "step": 75800
    },
    {
      "epoch": 8.163814444085673,
      "grad_norm": 0.8399979472160339,
      "learning_rate": 0.00011047085443378945,
      "loss": 3.1708,
      "step": 75850
    },
    {
      "epoch": 8.169195996125282,
      "grad_norm": 0.8141344785690308,
      "learning_rate": 0.00011014761340372804,
      "loss": 3.1854,
      "step": 75900
    },
    {
      "epoch": 8.17457754816489,
      "grad_norm": 0.8441236019134521,
      "learning_rate": 0.00010982437237366661,
      "loss": 3.1724,
      "step": 75950
    },
    {
      "epoch": 8.1799591002045,
      "grad_norm": 0.8193464875221252,
      "learning_rate": 0.0001095011313436052,
      "loss": 3.1778,
      "step": 76000
    },
    {
      "epoch": 8.1799591002045,
      "eval_accuracy": 0.3904291582628185,
      "eval_loss": 3.335801839828491,
      "eval_runtime": 180.1521,
      "eval_samples_per_second": 99.977,
      "eval_steps_per_second": 6.25,
      "step": 76000
    },
    {
      "epoch": 8.185340652244108,
      "grad_norm": 0.8355197906494141,
      "learning_rate": 0.0001091778903135438,
      "loss": 3.1686,
      "step": 76050
    },
    {
      "epoch": 8.190722204283716,
      "grad_norm": 0.8153267502784729,
      "learning_rate": 0.00010885464928348238,
      "loss": 3.1838,
      "step": 76100
    },
    {
      "epoch": 8.196103756323323,
      "grad_norm": 0.8740131258964539,
      "learning_rate": 0.00010853140825342096,
      "loss": 3.1696,
      "step": 76150
    },
    {
      "epoch": 8.201485308362932,
      "grad_norm": 0.8934776782989502,
      "learning_rate": 0.00010820816722335954,
      "loss": 3.1742,
      "step": 76200
    },
    {
      "epoch": 8.20686686040254,
      "grad_norm": 0.9179877042770386,
      "learning_rate": 0.00010788492619329814,
      "loss": 3.1695,
      "step": 76250
    },
    {
      "epoch": 8.212248412442149,
      "grad_norm": 0.8545275926589966,
      "learning_rate": 0.0001075616851632367,
      "loss": 3.172,
      "step": 76300
    },
    {
      "epoch": 8.217629964481757,
      "grad_norm": 0.8500077724456787,
      "learning_rate": 0.0001072384441331753,
      "loss": 3.1938,
      "step": 76350
    },
    {
      "epoch": 8.223011516521364,
      "grad_norm": 0.8435236811637878,
      "learning_rate": 0.00010691520310311388,
      "loss": 3.1834,
      "step": 76400
    },
    {
      "epoch": 8.228393068560973,
      "grad_norm": 0.8567854762077332,
      "learning_rate": 0.00010659196207305246,
      "loss": 3.169,
      "step": 76450
    },
    {
      "epoch": 8.233774620600581,
      "grad_norm": 0.8797134757041931,
      "learning_rate": 0.00010626872104299104,
      "loss": 3.1859,
      "step": 76500
    },
    {
      "epoch": 8.23915617264019,
      "grad_norm": 0.8667067289352417,
      "learning_rate": 0.00010594548001292964,
      "loss": 3.1973,
      "step": 76550
    },
    {
      "epoch": 8.244537724679798,
      "grad_norm": 0.8613109588623047,
      "learning_rate": 0.00010562223898286823,
      "loss": 3.1941,
      "step": 76600
    },
    {
      "epoch": 8.249919276719407,
      "grad_norm": 0.8120648264884949,
      "learning_rate": 0.0001052989979528068,
      "loss": 3.1899,
      "step": 76650
    },
    {
      "epoch": 8.255300828759013,
      "grad_norm": 0.8621121048927307,
      "learning_rate": 0.00010497575692274539,
      "loss": 3.1836,
      "step": 76700
    },
    {
      "epoch": 8.260682380798622,
      "grad_norm": 0.8764901757240295,
      "learning_rate": 0.00010465251589268397,
      "loss": 3.2011,
      "step": 76750
    },
    {
      "epoch": 8.26606393283823,
      "grad_norm": 0.8889468312263489,
      "learning_rate": 0.00010432927486262254,
      "loss": 3.1871,
      "step": 76800
    },
    {
      "epoch": 8.27144548487784,
      "grad_norm": 0.8703168630599976,
      "learning_rate": 0.00010400603383256114,
      "loss": 3.1714,
      "step": 76850
    },
    {
      "epoch": 8.276827036917448,
      "grad_norm": 0.8331544399261475,
      "learning_rate": 0.00010368279280249973,
      "loss": 3.1679,
      "step": 76900
    },
    {
      "epoch": 8.282208588957054,
      "grad_norm": 0.8356614708900452,
      "learning_rate": 0.0001033595517724383,
      "loss": 3.187,
      "step": 76950
    },
    {
      "epoch": 8.287590140996663,
      "grad_norm": 0.9109562039375305,
      "learning_rate": 0.00010303631074237689,
      "loss": 3.1902,
      "step": 77000
    },
    {
      "epoch": 8.287590140996663,
      "eval_accuracy": 0.3904268765521744,
      "eval_loss": 3.3348872661590576,
      "eval_runtime": 180.23,
      "eval_samples_per_second": 99.933,
      "eval_steps_per_second": 6.248,
      "step": 77000
    },
    {
      "epoch": 8.292971693036272,
      "grad_norm": 0.857496440410614,
      "learning_rate": 0.00010271306971231547,
      "loss": 3.1796,
      "step": 77050
    },
    {
      "epoch": 8.29835324507588,
      "grad_norm": 0.8897091746330261,
      "learning_rate": 0.00010238982868225407,
      "loss": 3.195,
      "step": 77100
    },
    {
      "epoch": 8.303734797115489,
      "grad_norm": 0.8627259135246277,
      "learning_rate": 0.00010206658765219264,
      "loss": 3.1633,
      "step": 77150
    },
    {
      "epoch": 8.309116349155097,
      "grad_norm": 0.8576809167861938,
      "learning_rate": 0.00010174334662213123,
      "loss": 3.178,
      "step": 77200
    },
    {
      "epoch": 8.314497901194704,
      "grad_norm": 0.913425862789154,
      "learning_rate": 0.00010142010559206981,
      "loss": 3.2039,
      "step": 77250
    },
    {
      "epoch": 8.319879453234313,
      "grad_norm": 0.8628541231155396,
      "learning_rate": 0.00010109686456200839,
      "loss": 3.1867,
      "step": 77300
    },
    {
      "epoch": 8.325261005273921,
      "grad_norm": 0.8340932726860046,
      "learning_rate": 0.00010077362353194697,
      "loss": 3.1896,
      "step": 77350
    },
    {
      "epoch": 8.33064255731353,
      "grad_norm": 0.8404185175895691,
      "learning_rate": 0.00010045038250188557,
      "loss": 3.1841,
      "step": 77400
    },
    {
      "epoch": 8.336024109353138,
      "grad_norm": 0.863701343536377,
      "learning_rate": 0.00010012714147182414,
      "loss": 3.1791,
      "step": 77450
    },
    {
      "epoch": 8.341405661392745,
      "grad_norm": 0.859533965587616,
      "learning_rate": 9.980390044176273e-05,
      "loss": 3.1825,
      "step": 77500
    },
    {
      "epoch": 8.346787213432354,
      "grad_norm": 0.8829686641693115,
      "learning_rate": 9.948065941170133e-05,
      "loss": 3.1839,
      "step": 77550
    },
    {
      "epoch": 8.352168765471962,
      "grad_norm": 0.8447333574295044,
      "learning_rate": 9.91574183816399e-05,
      "loss": 3.1852,
      "step": 77600
    },
    {
      "epoch": 8.35755031751157,
      "grad_norm": 0.8658083081245422,
      "learning_rate": 9.883417735157849e-05,
      "loss": 3.1787,
      "step": 77650
    },
    {
      "epoch": 8.36293186955118,
      "grad_norm": 0.8201994895935059,
      "learning_rate": 9.851093632151707e-05,
      "loss": 3.187,
      "step": 77700
    },
    {
      "epoch": 8.368313421590786,
      "grad_norm": 0.8592942953109741,
      "learning_rate": 9.818769529145566e-05,
      "loss": 3.1745,
      "step": 77750
    },
    {
      "epoch": 8.373694973630395,
      "grad_norm": 0.8522955775260925,
      "learning_rate": 9.786445426139423e-05,
      "loss": 3.1748,
      "step": 77800
    },
    {
      "epoch": 8.379076525670003,
      "grad_norm": 0.8451487421989441,
      "learning_rate": 9.754121323133283e-05,
      "loss": 3.1835,
      "step": 77850
    },
    {
      "epoch": 8.384458077709612,
      "grad_norm": 0.8497421145439148,
      "learning_rate": 9.72179722012714e-05,
      "loss": 3.186,
      "step": 77900
    },
    {
      "epoch": 8.38983962974922,
      "grad_norm": 0.8562092781066895,
      "learning_rate": 9.689473117120999e-05,
      "loss": 3.1989,
      "step": 77950
    },
    {
      "epoch": 8.395221181788829,
      "grad_norm": 0.8583953380584717,
      "learning_rate": 9.657149014114857e-05,
      "loss": 3.1921,
      "step": 78000
    },
    {
      "epoch": 8.395221181788829,
      "eval_accuracy": 0.3905158632672945,
      "eval_loss": 3.331059694290161,
      "eval_runtime": 180.1209,
      "eval_samples_per_second": 99.994,
      "eval_steps_per_second": 6.251,
      "step": 78000
    },
    {
      "epoch": 8.400602733828435,
      "grad_norm": 0.860522449016571,
      "learning_rate": 9.624824911108716e-05,
      "loss": 3.1781,
      "step": 78050
    },
    {
      "epoch": 8.405984285868044,
      "grad_norm": 0.875423014163971,
      "learning_rate": 9.592500808102574e-05,
      "loss": 3.172,
      "step": 78100
    },
    {
      "epoch": 8.411365837907653,
      "grad_norm": 0.8359448313713074,
      "learning_rate": 9.560176705096433e-05,
      "loss": 3.1911,
      "step": 78150
    },
    {
      "epoch": 8.416747389947261,
      "grad_norm": 0.8645302653312683,
      "learning_rate": 9.52785260209029e-05,
      "loss": 3.1773,
      "step": 78200
    },
    {
      "epoch": 8.42212894198687,
      "grad_norm": 0.8695240616798401,
      "learning_rate": 9.49552849908415e-05,
      "loss": 3.1858,
      "step": 78250
    },
    {
      "epoch": 8.427510494026476,
      "grad_norm": 0.8338247537612915,
      "learning_rate": 9.463204396078007e-05,
      "loss": 3.1981,
      "step": 78300
    },
    {
      "epoch": 8.432892046066085,
      "grad_norm": 0.9113908410072327,
      "learning_rate": 9.430880293071866e-05,
      "loss": 3.1838,
      "step": 78350
    },
    {
      "epoch": 8.438273598105694,
      "grad_norm": 0.853624165058136,
      "learning_rate": 9.398556190065726e-05,
      "loss": 3.1826,
      "step": 78400
    },
    {
      "epoch": 8.443655150145302,
      "grad_norm": 0.8597407341003418,
      "learning_rate": 9.366878569119705e-05,
      "loss": 3.1836,
      "step": 78450
    },
    {
      "epoch": 8.44903670218491,
      "grad_norm": 0.8677853941917419,
      "learning_rate": 9.334554466113565e-05,
      "loss": 3.1936,
      "step": 78500
    },
    {
      "epoch": 8.45441825422452,
      "grad_norm": 0.938667893409729,
      "learning_rate": 9.302230363107424e-05,
      "loss": 3.194,
      "step": 78550
    },
    {
      "epoch": 8.459799806264126,
      "grad_norm": 0.8398536443710327,
      "learning_rate": 9.269906260101281e-05,
      "loss": 3.1795,
      "step": 78600
    },
    {
      "epoch": 8.465181358303735,
      "grad_norm": 0.8896383047103882,
      "learning_rate": 9.23758215709514e-05,
      "loss": 3.2054,
      "step": 78650
    },
    {
      "epoch": 8.470562910343343,
      "grad_norm": 0.8669857382774353,
      "learning_rate": 9.205258054088998e-05,
      "loss": 3.1779,
      "step": 78700
    },
    {
      "epoch": 8.475944462382952,
      "grad_norm": 0.8286911249160767,
      "learning_rate": 9.172933951082856e-05,
      "loss": 3.1926,
      "step": 78750
    },
    {
      "epoch": 8.48132601442256,
      "grad_norm": 0.9176493883132935,
      "learning_rate": 9.140609848076715e-05,
      "loss": 3.174,
      "step": 78800
    },
    {
      "epoch": 8.486707566462167,
      "grad_norm": 0.8651835322380066,
      "learning_rate": 9.108285745070574e-05,
      "loss": 3.1803,
      "step": 78850
    },
    {
      "epoch": 8.492089118501776,
      "grad_norm": 0.8289731740951538,
      "learning_rate": 9.075961642064432e-05,
      "loss": 3.1996,
      "step": 78900
    },
    {
      "epoch": 8.497470670541384,
      "grad_norm": 0.9024319052696228,
      "learning_rate": 9.04363753905829e-05,
      "loss": 3.1928,
      "step": 78950
    },
    {
      "epoch": 8.502852222580993,
      "grad_norm": 0.8506482243537903,
      "learning_rate": 9.011313436052148e-05,
      "loss": 3.1977,
      "step": 79000
    },
    {
      "epoch": 8.502852222580993,
      "eval_accuracy": 0.3912251493189474,
      "eval_loss": 3.3283579349517822,
      "eval_runtime": 180.3098,
      "eval_samples_per_second": 99.889,
      "eval_steps_per_second": 6.245,
      "step": 79000
    },
    {
      "epoch": 8.508233774620601,
      "grad_norm": 0.8655647039413452,
      "learning_rate": 8.978989333046008e-05,
      "loss": 3.1843,
      "step": 79050
    },
    {
      "epoch": 8.513615326660208,
      "grad_norm": 0.8432492613792419,
      "learning_rate": 8.946665230039865e-05,
      "loss": 3.1584,
      "step": 79100
    },
    {
      "epoch": 8.518996878699816,
      "grad_norm": 0.817113995552063,
      "learning_rate": 8.914341127033724e-05,
      "loss": 3.2022,
      "step": 79150
    },
    {
      "epoch": 8.524378430739425,
      "grad_norm": 0.8752052783966064,
      "learning_rate": 8.882017024027582e-05,
      "loss": 3.1791,
      "step": 79200
    },
    {
      "epoch": 8.529759982779034,
      "grad_norm": 0.8376759886741638,
      "learning_rate": 8.84969292102144e-05,
      "loss": 3.1689,
      "step": 79250
    },
    {
      "epoch": 8.535141534818642,
      "grad_norm": 0.8467958569526672,
      "learning_rate": 8.817368818015298e-05,
      "loss": 3.1819,
      "step": 79300
    },
    {
      "epoch": 8.54052308685825,
      "grad_norm": 0.842562735080719,
      "learning_rate": 8.785044715009158e-05,
      "loss": 3.1974,
      "step": 79350
    },
    {
      "epoch": 8.545904638897857,
      "grad_norm": 0.9210129976272583,
      "learning_rate": 8.752720612003017e-05,
      "loss": 3.191,
      "step": 79400
    },
    {
      "epoch": 8.551286190937466,
      "grad_norm": 0.8453426361083984,
      "learning_rate": 8.720396508996874e-05,
      "loss": 3.1839,
      "step": 79450
    },
    {
      "epoch": 8.556667742977075,
      "grad_norm": 0.8634918928146362,
      "learning_rate": 8.688072405990734e-05,
      "loss": 3.1739,
      "step": 79500
    },
    {
      "epoch": 8.562049295016683,
      "grad_norm": 0.8293809294700623,
      "learning_rate": 8.655748302984592e-05,
      "loss": 3.169,
      "step": 79550
    },
    {
      "epoch": 8.567430847056292,
      "grad_norm": 0.8583179116249084,
      "learning_rate": 8.62342419997845e-05,
      "loss": 3.1982,
      "step": 79600
    },
    {
      "epoch": 8.572812399095898,
      "grad_norm": 0.8895404934883118,
      "learning_rate": 8.591100096972308e-05,
      "loss": 3.1812,
      "step": 79650
    },
    {
      "epoch": 8.578193951135507,
      "grad_norm": 0.820783257484436,
      "learning_rate": 8.558775993966167e-05,
      "loss": 3.1779,
      "step": 79700
    },
    {
      "epoch": 8.583575503175116,
      "grad_norm": 0.9021645188331604,
      "learning_rate": 8.526451890960024e-05,
      "loss": 3.1975,
      "step": 79750
    },
    {
      "epoch": 8.588957055214724,
      "grad_norm": 0.8516561985015869,
      "learning_rate": 8.494127787953884e-05,
      "loss": 3.1713,
      "step": 79800
    },
    {
      "epoch": 8.594338607254333,
      "grad_norm": 0.8891592621803284,
      "learning_rate": 8.461803684947742e-05,
      "loss": 3.2062,
      "step": 79850
    },
    {
      "epoch": 8.599720159293941,
      "grad_norm": 0.891211211681366,
      "learning_rate": 8.429479581941601e-05,
      "loss": 3.1899,
      "step": 79900
    },
    {
      "epoch": 8.605101711333548,
      "grad_norm": 0.8360291719436646,
      "learning_rate": 8.397155478935458e-05,
      "loss": 3.1905,
      "step": 79950
    },
    {
      "epoch": 8.610483263373157,
      "grad_norm": 0.8511280417442322,
      "learning_rate": 8.364831375929317e-05,
      "loss": 3.1876,
      "step": 80000
    },
    {
      "epoch": 8.610483263373157,
      "eval_accuracy": 0.39159695950104856,
      "eval_loss": 3.3249058723449707,
      "eval_runtime": 180.4737,
      "eval_samples_per_second": 99.798,
      "eval_steps_per_second": 6.239,
      "step": 80000
    },
    {
      "epoch": 8.615864815412765,
      "grad_norm": 0.9268209934234619,
      "learning_rate": 8.332507272923177e-05,
      "loss": 3.2012,
      "step": 80050
    },
    {
      "epoch": 8.621246367452374,
      "grad_norm": 0.865637481212616,
      "learning_rate": 8.300183169917034e-05,
      "loss": 3.1838,
      "step": 80100
    },
    {
      "epoch": 8.626627919491982,
      "grad_norm": 0.8277395367622375,
      "learning_rate": 8.267859066910892e-05,
      "loss": 3.1959,
      "step": 80150
    },
    {
      "epoch": 8.632009471531589,
      "grad_norm": 0.8716737627983093,
      "learning_rate": 8.235534963904751e-05,
      "loss": 3.1944,
      "step": 80200
    },
    {
      "epoch": 8.637391023571197,
      "grad_norm": 0.8669558763504028,
      "learning_rate": 8.203210860898608e-05,
      "loss": 3.2017,
      "step": 80250
    },
    {
      "epoch": 8.642772575610806,
      "grad_norm": 0.842424213886261,
      "learning_rate": 8.170886757892467e-05,
      "loss": 3.158,
      "step": 80300
    },
    {
      "epoch": 8.648154127650415,
      "grad_norm": 0.8678565621376038,
      "learning_rate": 8.138562654886327e-05,
      "loss": 3.1759,
      "step": 80350
    },
    {
      "epoch": 8.653535679690023,
      "grad_norm": 0.8407251834869385,
      "learning_rate": 8.106238551880185e-05,
      "loss": 3.1856,
      "step": 80400
    },
    {
      "epoch": 8.658917231729632,
      "grad_norm": 0.8746572732925415,
      "learning_rate": 8.073914448874043e-05,
      "loss": 3.1802,
      "step": 80450
    },
    {
      "epoch": 8.664298783769238,
      "grad_norm": 0.827447235584259,
      "learning_rate": 8.042236827928025e-05,
      "loss": 3.1879,
      "step": 80500
    },
    {
      "epoch": 8.669680335808847,
      "grad_norm": 0.8395963907241821,
      "learning_rate": 8.009912724921882e-05,
      "loss": 3.1832,
      "step": 80550
    },
    {
      "epoch": 8.675061887848456,
      "grad_norm": 0.8884485363960266,
      "learning_rate": 7.977588621915741e-05,
      "loss": 3.1769,
      "step": 80600
    },
    {
      "epoch": 8.680443439888064,
      "grad_norm": 0.8452661633491516,
      "learning_rate": 7.9452645189096e-05,
      "loss": 3.21,
      "step": 80650
    },
    {
      "epoch": 8.685824991927673,
      "grad_norm": 0.9012867212295532,
      "learning_rate": 7.912940415903459e-05,
      "loss": 3.1787,
      "step": 80700
    },
    {
      "epoch": 8.69120654396728,
      "grad_norm": 0.8785222172737122,
      "learning_rate": 7.880616312897316e-05,
      "loss": 3.1909,
      "step": 80750
    },
    {
      "epoch": 8.696588096006888,
      "grad_norm": 0.9251953959465027,
      "learning_rate": 7.848292209891175e-05,
      "loss": 3.1933,
      "step": 80800
    },
    {
      "epoch": 8.701969648046497,
      "grad_norm": 0.8459815382957458,
      "learning_rate": 7.815968106885033e-05,
      "loss": 3.1805,
      "step": 80850
    },
    {
      "epoch": 8.707351200086105,
      "grad_norm": 0.8944352865219116,
      "learning_rate": 7.783644003878891e-05,
      "loss": 3.1875,
      "step": 80900
    },
    {
      "epoch": 8.712732752125714,
      "grad_norm": 0.8620814681053162,
      "learning_rate": 7.75131990087275e-05,
      "loss": 3.1919,
      "step": 80950
    },
    {
      "epoch": 8.718114304165322,
      "grad_norm": 0.8819995522499084,
      "learning_rate": 7.718995797866609e-05,
      "loss": 3.1825,
      "step": 81000
    },
    {
      "epoch": 8.718114304165322,
      "eval_accuracy": 0.3917101758101512,
      "eval_loss": 3.322167158126831,
      "eval_runtime": 180.0608,
      "eval_samples_per_second": 100.027,
      "eval_steps_per_second": 6.253,
      "step": 81000
    },
    {
      "epoch": 8.723495856204929,
      "grad_norm": 0.8823017477989197,
      "learning_rate": 7.686671694860466e-05,
      "loss": 3.1884,
      "step": 81050
    },
    {
      "epoch": 8.728877408244538,
      "grad_norm": 0.8862807750701904,
      "learning_rate": 7.654347591854325e-05,
      "loss": 3.1889,
      "step": 81100
    },
    {
      "epoch": 8.734258960284146,
      "grad_norm": 0.9089580774307251,
      "learning_rate": 7.622023488848185e-05,
      "loss": 3.1832,
      "step": 81150
    },
    {
      "epoch": 8.739640512323755,
      "grad_norm": 0.8399238586425781,
      "learning_rate": 7.589699385842043e-05,
      "loss": 3.1812,
      "step": 81200
    },
    {
      "epoch": 8.745022064363363,
      "grad_norm": 0.868499755859375,
      "learning_rate": 7.557375282835901e-05,
      "loss": 3.1851,
      "step": 81250
    },
    {
      "epoch": 8.75040361640297,
      "grad_norm": 0.8629152774810791,
      "learning_rate": 7.525051179829759e-05,
      "loss": 3.18,
      "step": 81300
    },
    {
      "epoch": 8.755785168442578,
      "grad_norm": 0.8513661623001099,
      "learning_rate": 7.492727076823617e-05,
      "loss": 3.1762,
      "step": 81350
    },
    {
      "epoch": 8.761166720482187,
      "grad_norm": 0.9037156105041504,
      "learning_rate": 7.460402973817477e-05,
      "loss": 3.2112,
      "step": 81400
    },
    {
      "epoch": 8.766548272521796,
      "grad_norm": 0.8453684449195862,
      "learning_rate": 7.428078870811335e-05,
      "loss": 3.182,
      "step": 81450
    },
    {
      "epoch": 8.771929824561404,
      "grad_norm": 0.831931471824646,
      "learning_rate": 7.395754767805193e-05,
      "loss": 3.1843,
      "step": 81500
    },
    {
      "epoch": 8.777311376601011,
      "grad_norm": 0.8770779967308044,
      "learning_rate": 7.363430664799051e-05,
      "loss": 3.1861,
      "step": 81550
    },
    {
      "epoch": 8.78269292864062,
      "grad_norm": 0.8554553985595703,
      "learning_rate": 7.331753043853033e-05,
      "loss": 3.1837,
      "step": 81600
    },
    {
      "epoch": 8.788074480680228,
      "grad_norm": 0.8347432613372803,
      "learning_rate": 7.299428940846891e-05,
      "loss": 3.1902,
      "step": 81650
    },
    {
      "epoch": 8.793456032719837,
      "grad_norm": 0.8451250195503235,
      "learning_rate": 7.267104837840749e-05,
      "loss": 3.1677,
      "step": 81700
    },
    {
      "epoch": 8.798837584759445,
      "grad_norm": 0.8608439564704895,
      "learning_rate": 7.234780734834607e-05,
      "loss": 3.1794,
      "step": 81750
    },
    {
      "epoch": 8.804219136799054,
      "grad_norm": 0.8560290336608887,
      "learning_rate": 7.202456631828465e-05,
      "loss": 3.1846,
      "step": 81800
    },
    {
      "epoch": 8.80960068883866,
      "grad_norm": 0.8419299721717834,
      "learning_rate": 7.170132528822325e-05,
      "loss": 3.1821,
      "step": 81850
    },
    {
      "epoch": 8.814982240878269,
      "grad_norm": 0.8612204194068909,
      "learning_rate": 7.137808425816183e-05,
      "loss": 3.1991,
      "step": 81900
    },
    {
      "epoch": 8.820363792917878,
      "grad_norm": 0.8462232947349548,
      "learning_rate": 7.105484322810041e-05,
      "loss": 3.1852,
      "step": 81950
    },
    {
      "epoch": 8.825745344957486,
      "grad_norm": 0.8963456153869629,
      "learning_rate": 7.073160219803899e-05,
      "loss": 3.2005,
      "step": 82000
    },
    {
      "epoch": 8.825745344957486,
      "eval_accuracy": 0.3922394240266956,
      "eval_loss": 3.3186590671539307,
      "eval_runtime": 180.0835,
      "eval_samples_per_second": 100.015,
      "eval_steps_per_second": 6.253,
      "step": 82000
    },
    {
      "epoch": 8.831126896997095,
      "grad_norm": 0.8609824776649475,
      "learning_rate": 7.040836116797757e-05,
      "loss": 3.1799,
      "step": 82050
    },
    {
      "epoch": 8.836508449036701,
      "grad_norm": 0.839068591594696,
      "learning_rate": 7.008512013791617e-05,
      "loss": 3.1796,
      "step": 82100
    },
    {
      "epoch": 8.84189000107631,
      "grad_norm": 0.850020170211792,
      "learning_rate": 6.976187910785475e-05,
      "loss": 3.19,
      "step": 82150
    },
    {
      "epoch": 8.847271553115919,
      "grad_norm": 0.8698260188102722,
      "learning_rate": 6.943863807779334e-05,
      "loss": 3.1946,
      "step": 82200
    },
    {
      "epoch": 8.852653105155527,
      "grad_norm": 0.8751007914543152,
      "learning_rate": 6.911539704773192e-05,
      "loss": 3.1885,
      "step": 82250
    },
    {
      "epoch": 8.858034657195136,
      "grad_norm": 0.8869954943656921,
      "learning_rate": 6.87921560176705e-05,
      "loss": 3.1783,
      "step": 82300
    },
    {
      "epoch": 8.863416209234742,
      "grad_norm": 0.8722847104072571,
      "learning_rate": 6.846891498760909e-05,
      "loss": 3.1916,
      "step": 82350
    },
    {
      "epoch": 8.868797761274351,
      "grad_norm": 0.8729155659675598,
      "learning_rate": 6.814567395754767e-05,
      "loss": 3.1889,
      "step": 82400
    },
    {
      "epoch": 8.87417931331396,
      "grad_norm": 0.8469229936599731,
      "learning_rate": 6.782243292748626e-05,
      "loss": 3.188,
      "step": 82450
    },
    {
      "epoch": 8.879560865353568,
      "grad_norm": 0.8862374424934387,
      "learning_rate": 6.749919189742484e-05,
      "loss": 3.1697,
      "step": 82500
    },
    {
      "epoch": 8.884942417393177,
      "grad_norm": 0.8705187439918518,
      "learning_rate": 6.717595086736342e-05,
      "loss": 3.1786,
      "step": 82550
    },
    {
      "epoch": 8.890323969432785,
      "grad_norm": 0.8562043309211731,
      "learning_rate": 6.6852709837302e-05,
      "loss": 3.1752,
      "step": 82600
    },
    {
      "epoch": 8.895705521472392,
      "grad_norm": 0.9573428630828857,
      "learning_rate": 6.652946880724059e-05,
      "loss": 3.1849,
      "step": 82650
    },
    {
      "epoch": 8.901087073512,
      "grad_norm": 0.8421146869659424,
      "learning_rate": 6.620622777717918e-05,
      "loss": 3.1855,
      "step": 82700
    },
    {
      "epoch": 8.906468625551609,
      "grad_norm": 0.8492905497550964,
      "learning_rate": 6.588298674711776e-05,
      "loss": 3.1856,
      "step": 82750
    },
    {
      "epoch": 8.911850177591218,
      "grad_norm": 0.8871941566467285,
      "learning_rate": 6.555974571705636e-05,
      "loss": 3.2015,
      "step": 82800
    },
    {
      "epoch": 8.917231729630826,
      "grad_norm": 0.8534443378448486,
      "learning_rate": 6.523650468699494e-05,
      "loss": 3.168,
      "step": 82850
    },
    {
      "epoch": 8.922613281670433,
      "grad_norm": 0.8974738717079163,
      "learning_rate": 6.491326365693352e-05,
      "loss": 3.1728,
      "step": 82900
    },
    {
      "epoch": 8.927994833710041,
      "grad_norm": 0.8600006103515625,
      "learning_rate": 6.45900226268721e-05,
      "loss": 3.1972,
      "step": 82950
    },
    {
      "epoch": 8.93337638574965,
      "grad_norm": 0.8655703663825989,
      "learning_rate": 6.426678159681068e-05,
      "loss": 3.1977,
      "step": 83000
    },
    {
      "epoch": 8.93337638574965,
      "eval_accuracy": 0.39260102083734216,
      "eval_loss": 3.314582586288452,
      "eval_runtime": 180.2494,
      "eval_samples_per_second": 99.923,
      "eval_steps_per_second": 6.247,
      "step": 83000
    },
    {
      "epoch": 8.938757937789259,
      "grad_norm": 0.8925529718399048,
      "learning_rate": 6.394354056674928e-05,
      "loss": 3.1958,
      "step": 83050
    },
    {
      "epoch": 8.944139489828867,
      "grad_norm": 0.8607410788536072,
      "learning_rate": 6.362029953668786e-05,
      "loss": 3.1825,
      "step": 83100
    },
    {
      "epoch": 8.949521041868476,
      "grad_norm": 0.8853652477264404,
      "learning_rate": 6.329705850662644e-05,
      "loss": 3.1842,
      "step": 83150
    },
    {
      "epoch": 8.954902593908082,
      "grad_norm": 0.8844691514968872,
      "learning_rate": 6.297381747656502e-05,
      "loss": 3.1719,
      "step": 83200
    },
    {
      "epoch": 8.960284145947691,
      "grad_norm": 0.8798843622207642,
      "learning_rate": 6.26505764465036e-05,
      "loss": 3.1837,
      "step": 83250
    },
    {
      "epoch": 8.9656656979873,
      "grad_norm": 0.8720901608467102,
      "learning_rate": 6.23273354164422e-05,
      "loss": 3.1988,
      "step": 83300
    },
    {
      "epoch": 8.971047250026908,
      "grad_norm": 0.8754738569259644,
      "learning_rate": 6.200409438638078e-05,
      "loss": 3.1926,
      "step": 83350
    },
    {
      "epoch": 8.976428802066517,
      "grad_norm": 0.8542868494987488,
      "learning_rate": 6.168085335631936e-05,
      "loss": 3.1732,
      "step": 83400
    },
    {
      "epoch": 8.981810354106123,
      "grad_norm": 0.8611506223678589,
      "learning_rate": 6.135761232625794e-05,
      "loss": 3.1916,
      "step": 83450
    },
    {
      "epoch": 8.987191906145732,
      "grad_norm": 0.8679481744766235,
      "learning_rate": 6.1034371296196526e-05,
      "loss": 3.1845,
      "step": 83500
    },
    {
      "epoch": 8.99257345818534,
      "grad_norm": 0.8732113838195801,
      "learning_rate": 6.0711130266135114e-05,
      "loss": 3.1882,
      "step": 83550
    },
    {
      "epoch": 8.997955010224949,
      "grad_norm": 0.8252281546592712,
      "learning_rate": 6.0387889236073695e-05,
      "loss": 3.2006,
      "step": 83600
    },
    {
      "epoch": 9.003336562264558,
      "grad_norm": 0.8319926857948303,
      "learning_rate": 6.0064648206012276e-05,
      "loss": 3.145,
      "step": 83650
    },
    {
      "epoch": 9.008718114304166,
      "grad_norm": 0.8742890357971191,
      "learning_rate": 5.9741407175950864e-05,
      "loss": 3.1252,
      "step": 83700
    },
    {
      "epoch": 9.014099666343773,
      "grad_norm": 0.8541421294212341,
      "learning_rate": 5.9418166145889445e-05,
      "loss": 3.1408,
      "step": 83750
    },
    {
      "epoch": 9.019481218383381,
      "grad_norm": 0.9140190482139587,
      "learning_rate": 5.910138993642926e-05,
      "loss": 3.1268,
      "step": 83800
    },
    {
      "epoch": 9.02486277042299,
      "grad_norm": 0.8530510067939758,
      "learning_rate": 5.877814890636784e-05,
      "loss": 3.1121,
      "step": 83850
    },
    {
      "epoch": 9.030244322462599,
      "grad_norm": 0.8723753094673157,
      "learning_rate": 5.845490787630643e-05,
      "loss": 3.1191,
      "step": 83900
    },
    {
      "epoch": 9.035625874502207,
      "grad_norm": 0.8698916435241699,
      "learning_rate": 5.813166684624501e-05,
      "loss": 3.1148,
      "step": 83950
    },
    {
      "epoch": 9.041007426541814,
      "grad_norm": 0.8469932675361633,
      "learning_rate": 5.7808425816183596e-05,
      "loss": 3.1246,
      "step": 84000
    },
    {
      "epoch": 9.041007426541814,
      "eval_accuracy": 0.392553104913816,
      "eval_loss": 3.317844867706299,
      "eval_runtime": 180.533,
      "eval_samples_per_second": 99.766,
      "eval_steps_per_second": 6.237,
      "step": 84000
    },
    {
      "epoch": 9.046388978581422,
      "grad_norm": 0.8725888133049011,
      "learning_rate": 5.748518478612218e-05,
      "loss": 3.1223,
      "step": 84050
    },
    {
      "epoch": 9.051770530621031,
      "grad_norm": 0.8994928598403931,
      "learning_rate": 5.716194375606076e-05,
      "loss": 3.1452,
      "step": 84100
    },
    {
      "epoch": 9.05715208266064,
      "grad_norm": 0.8656386137008667,
      "learning_rate": 5.6838702725999346e-05,
      "loss": 3.1392,
      "step": 84150
    },
    {
      "epoch": 9.062533634700248,
      "grad_norm": 0.878341794013977,
      "learning_rate": 5.651546169593793e-05,
      "loss": 3.1356,
      "step": 84200
    },
    {
      "epoch": 9.067915186739857,
      "grad_norm": 0.8708506226539612,
      "learning_rate": 5.619222066587652e-05,
      "loss": 3.1184,
      "step": 84250
    },
    {
      "epoch": 9.073296738779463,
      "grad_norm": 0.8545727133750916,
      "learning_rate": 5.58689796358151e-05,
      "loss": 3.1142,
      "step": 84300
    },
    {
      "epoch": 9.078678290819072,
      "grad_norm": 0.8976442217826843,
      "learning_rate": 5.554573860575369e-05,
      "loss": 3.107,
      "step": 84350
    },
    {
      "epoch": 9.08405984285868,
      "grad_norm": 0.8379390239715576,
      "learning_rate": 5.522249757569227e-05,
      "loss": 3.1199,
      "step": 84400
    },
    {
      "epoch": 9.089441394898289,
      "grad_norm": 0.8370739221572876,
      "learning_rate": 5.489925654563085e-05,
      "loss": 3.1285,
      "step": 84450
    },
    {
      "epoch": 9.094822946937898,
      "grad_norm": 0.8469734191894531,
      "learning_rate": 5.457601551556944e-05,
      "loss": 3.1455,
      "step": 84500
    },
    {
      "epoch": 9.100204498977504,
      "grad_norm": 0.8684940934181213,
      "learning_rate": 5.425277448550802e-05,
      "loss": 3.1088,
      "step": 84550
    },
    {
      "epoch": 9.105586051017113,
      "grad_norm": 0.8705862760543823,
      "learning_rate": 5.392953345544661e-05,
      "loss": 3.1245,
      "step": 84600
    },
    {
      "epoch": 9.110967603056721,
      "grad_norm": 0.8357287645339966,
      "learning_rate": 5.360629242538519e-05,
      "loss": 3.1275,
      "step": 84650
    },
    {
      "epoch": 9.11634915509633,
      "grad_norm": 0.8421687483787537,
      "learning_rate": 5.328305139532377e-05,
      "loss": 3.1243,
      "step": 84700
    },
    {
      "epoch": 9.121730707135939,
      "grad_norm": 0.8550165891647339,
      "learning_rate": 5.295981036526236e-05,
      "loss": 3.1221,
      "step": 84750
    },
    {
      "epoch": 9.127112259175545,
      "grad_norm": 0.8496278524398804,
      "learning_rate": 5.263656933520094e-05,
      "loss": 3.1369,
      "step": 84800
    },
    {
      "epoch": 9.132493811215154,
      "grad_norm": 0.8498865365982056,
      "learning_rate": 5.231332830513953e-05,
      "loss": 3.1421,
      "step": 84850
    },
    {
      "epoch": 9.137875363254762,
      "grad_norm": 0.905013382434845,
      "learning_rate": 5.199008727507811e-05,
      "loss": 3.1319,
      "step": 84900
    },
    {
      "epoch": 9.143256915294371,
      "grad_norm": 0.9200946092605591,
      "learning_rate": 5.166684624501669e-05,
      "loss": 3.112,
      "step": 84950
    },
    {
      "epoch": 9.14863846733398,
      "grad_norm": 0.847834050655365,
      "learning_rate": 5.1343605214955286e-05,
      "loss": 3.1312,
      "step": 85000
    },
    {
      "epoch": 9.14863846733398,
      "eval_accuracy": 0.39271803999751836,
      "eval_loss": 3.315976142883301,
      "eval_runtime": 179.964,
      "eval_samples_per_second": 100.081,
      "eval_steps_per_second": 6.257,
      "step": 85000
    },
    {
      "epoch": 9.154020019373588,
      "grad_norm": 0.8495249152183533,
      "learning_rate": 5.102036418489387e-05,
      "loss": 3.1226,
      "step": 85050
    },
    {
      "epoch": 9.159401571413195,
      "grad_norm": 0.9137347936630249,
      "learning_rate": 5.0697123154832455e-05,
      "loss": 3.1358,
      "step": 85100
    },
    {
      "epoch": 9.164783123452803,
      "grad_norm": 0.8848801851272583,
      "learning_rate": 5.0373882124771036e-05,
      "loss": 3.1319,
      "step": 85150
    },
    {
      "epoch": 9.170164675492412,
      "grad_norm": 0.8463972806930542,
      "learning_rate": 5.005064109470962e-05,
      "loss": 3.1284,
      "step": 85200
    },
    {
      "epoch": 9.17554622753202,
      "grad_norm": 0.8625044822692871,
      "learning_rate": 4.9727400064648205e-05,
      "loss": 3.1187,
      "step": 85250
    },
    {
      "epoch": 9.180927779571629,
      "grad_norm": 0.8781099319458008,
      "learning_rate": 4.9404159034586786e-05,
      "loss": 3.1446,
      "step": 85300
    },
    {
      "epoch": 9.186309331611236,
      "grad_norm": 0.877238392829895,
      "learning_rate": 4.9080918004525374e-05,
      "loss": 3.1407,
      "step": 85350
    },
    {
      "epoch": 9.191690883650844,
      "grad_norm": 0.9211892485618591,
      "learning_rate": 4.8757676974463955e-05,
      "loss": 3.1393,
      "step": 85400
    },
    {
      "epoch": 9.197072435690453,
      "grad_norm": 0.8986982107162476,
      "learning_rate": 4.8434435944402536e-05,
      "loss": 3.1425,
      "step": 85450
    },
    {
      "epoch": 9.202453987730062,
      "grad_norm": 0.8680441975593567,
      "learning_rate": 4.8111194914341124e-05,
      "loss": 3.1271,
      "step": 85500
    },
    {
      "epoch": 9.20783553976967,
      "grad_norm": 0.8860616683959961,
      "learning_rate": 4.7787953884279705e-05,
      "loss": 3.1492,
      "step": 85550
    },
    {
      "epoch": 9.213217091809279,
      "grad_norm": 0.895710289478302,
      "learning_rate": 4.746471285421829e-05,
      "loss": 3.1334,
      "step": 85600
    },
    {
      "epoch": 9.218598643848885,
      "grad_norm": 0.8937450051307678,
      "learning_rate": 4.7141471824156874e-05,
      "loss": 3.1275,
      "step": 85650
    },
    {
      "epoch": 9.223980195888494,
      "grad_norm": 0.8729709386825562,
      "learning_rate": 4.6818230794095455e-05,
      "loss": 3.1199,
      "step": 85700
    },
    {
      "epoch": 9.229361747928102,
      "grad_norm": 0.878231942653656,
      "learning_rate": 4.649498976403404e-05,
      "loss": 3.1359,
      "step": 85750
    },
    {
      "epoch": 9.234743299967711,
      "grad_norm": 0.8926569223403931,
      "learning_rate": 4.6171748733972624e-05,
      "loss": 3.1282,
      "step": 85800
    },
    {
      "epoch": 9.24012485200732,
      "grad_norm": 0.945385754108429,
      "learning_rate": 4.584850770391122e-05,
      "loss": 3.1277,
      "step": 85850
    },
    {
      "epoch": 9.245506404046926,
      "grad_norm": 0.8847855925559998,
      "learning_rate": 4.55252666738498e-05,
      "loss": 3.1344,
      "step": 85900
    },
    {
      "epoch": 9.250887956086535,
      "grad_norm": 0.8700735569000244,
      "learning_rate": 4.520202564378838e-05,
      "loss": 3.1481,
      "step": 85950
    },
    {
      "epoch": 9.256269508126143,
      "grad_norm": 0.8868721127510071,
      "learning_rate": 4.487878461372697e-05,
      "loss": 3.1217,
      "step": 86000
    },
    {
      "epoch": 9.256269508126143,
      "eval_accuracy": 0.39288764715539676,
      "eval_loss": 3.315528392791748,
      "eval_runtime": 180.0796,
      "eval_samples_per_second": 100.017,
      "eval_steps_per_second": 6.253,
      "step": 86000
    },
    {
      "epoch": 9.261651060165752,
      "grad_norm": 0.8628836274147034,
      "learning_rate": 4.455554358366555e-05,
      "loss": 3.14,
      "step": 86050
    },
    {
      "epoch": 9.26703261220536,
      "grad_norm": 0.8455495834350586,
      "learning_rate": 4.423230255360414e-05,
      "loss": 3.1434,
      "step": 86100
    },
    {
      "epoch": 9.272414164244967,
      "grad_norm": 0.8857659101486206,
      "learning_rate": 4.390906152354272e-05,
      "loss": 3.143,
      "step": 86150
    },
    {
      "epoch": 9.277795716284576,
      "grad_norm": 0.8970041275024414,
      "learning_rate": 4.35858204934813e-05,
      "loss": 3.1309,
      "step": 86200
    },
    {
      "epoch": 9.283177268324184,
      "grad_norm": 0.8499646186828613,
      "learning_rate": 4.326257946341989e-05,
      "loss": 3.1318,
      "step": 86250
    },
    {
      "epoch": 9.288558820363793,
      "grad_norm": 0.9052905440330505,
      "learning_rate": 4.293933843335847e-05,
      "loss": 3.1497,
      "step": 86300
    },
    {
      "epoch": 9.293940372403402,
      "grad_norm": 0.8457171320915222,
      "learning_rate": 4.261609740329706e-05,
      "loss": 3.1347,
      "step": 86350
    },
    {
      "epoch": 9.29932192444301,
      "grad_norm": 0.9543759226799011,
      "learning_rate": 4.229285637323564e-05,
      "loss": 3.1198,
      "step": 86400
    },
    {
      "epoch": 9.304703476482617,
      "grad_norm": 0.8748459219932556,
      "learning_rate": 4.196961534317422e-05,
      "loss": 3.1473,
      "step": 86450
    },
    {
      "epoch": 9.310085028522225,
      "grad_norm": 0.8755471110343933,
      "learning_rate": 4.164637431311281e-05,
      "loss": 3.1387,
      "step": 86500
    },
    {
      "epoch": 9.315466580561834,
      "grad_norm": 0.8886383175849915,
      "learning_rate": 4.132313328305139e-05,
      "loss": 3.1433,
      "step": 86550
    },
    {
      "epoch": 9.320848132601443,
      "grad_norm": 0.9134498238563538,
      "learning_rate": 4.099989225298998e-05,
      "loss": 3.1315,
      "step": 86600
    },
    {
      "epoch": 9.326229684641051,
      "grad_norm": 0.8910537958145142,
      "learning_rate": 4.0676651222928564e-05,
      "loss": 3.1212,
      "step": 86650
    },
    {
      "epoch": 9.331611236680658,
      "grad_norm": 0.8579962849617004,
      "learning_rate": 4.035341019286714e-05,
      "loss": 3.1215,
      "step": 86700
    },
    {
      "epoch": 9.336992788720266,
      "grad_norm": 0.9000516533851624,
      "learning_rate": 4.003016916280573e-05,
      "loss": 3.144,
      "step": 86750
    },
    {
      "epoch": 9.342374340759875,
      "grad_norm": 0.8634475469589233,
      "learning_rate": 3.9706928132744314e-05,
      "loss": 3.1411,
      "step": 86800
    },
    {
      "epoch": 9.347755892799483,
      "grad_norm": 0.8831536769866943,
      "learning_rate": 3.93836871026829e-05,
      "loss": 3.1445,
      "step": 86850
    },
    {
      "epoch": 9.353137444839092,
      "grad_norm": 0.8995107412338257,
      "learning_rate": 3.906044607262148e-05,
      "loss": 3.1412,
      "step": 86900
    },
    {
      "epoch": 9.3585189968787,
      "grad_norm": 0.8746486306190491,
      "learning_rate": 3.8737205042560064e-05,
      "loss": 3.1454,
      "step": 86950
    },
    {
      "epoch": 9.363900548918307,
      "grad_norm": 0.8589686155319214,
      "learning_rate": 3.841396401249865e-05,
      "loss": 3.1471,
      "step": 87000
    },
    {
      "epoch": 9.363900548918307,
      "eval_accuracy": 0.39334594503619824,
      "eval_loss": 3.312450885772705,
      "eval_runtime": 180.3192,
      "eval_samples_per_second": 99.884,
      "eval_steps_per_second": 6.244,
      "step": 87000
    },
    {
      "epoch": 9.369282100957916,
      "grad_norm": 0.8290409445762634,
      "learning_rate": 3.809072298243723e-05,
      "loss": 3.1332,
      "step": 87050
    },
    {
      "epoch": 9.374663652997524,
      "grad_norm": 0.8649762272834778,
      "learning_rate": 3.776748195237582e-05,
      "loss": 3.1382,
      "step": 87100
    },
    {
      "epoch": 9.380045205037133,
      "grad_norm": 0.9703359603881836,
      "learning_rate": 3.74442409223144e-05,
      "loss": 3.1272,
      "step": 87150
    },
    {
      "epoch": 9.385426757076742,
      "grad_norm": 0.8370105028152466,
      "learning_rate": 3.712099989225299e-05,
      "loss": 3.1192,
      "step": 87200
    },
    {
      "epoch": 9.390808309116348,
      "grad_norm": 0.8700204491615295,
      "learning_rate": 3.679775886219157e-05,
      "loss": 3.144,
      "step": 87250
    },
    {
      "epoch": 9.396189861155957,
      "grad_norm": 0.9093730449676514,
      "learning_rate": 3.647451783213015e-05,
      "loss": 3.1191,
      "step": 87300
    },
    {
      "epoch": 9.401571413195565,
      "grad_norm": 0.8856383562088013,
      "learning_rate": 3.615127680206874e-05,
      "loss": 3.1242,
      "step": 87350
    },
    {
      "epoch": 9.406952965235174,
      "grad_norm": 0.8729093670845032,
      "learning_rate": 3.582803577200732e-05,
      "loss": 3.1429,
      "step": 87400
    },
    {
      "epoch": 9.412334517274783,
      "grad_norm": 0.8521737456321716,
      "learning_rate": 3.550479474194591e-05,
      "loss": 3.1177,
      "step": 87450
    },
    {
      "epoch": 9.417716069314391,
      "grad_norm": 0.8820858597755432,
      "learning_rate": 3.518155371188449e-05,
      "loss": 3.1373,
      "step": 87500
    },
    {
      "epoch": 9.423097621353998,
      "grad_norm": 0.8602292537689209,
      "learning_rate": 3.485831268182308e-05,
      "loss": 3.1505,
      "step": 87550
    },
    {
      "epoch": 9.428479173393606,
      "grad_norm": 0.8709748387336731,
      "learning_rate": 3.453507165176166e-05,
      "loss": 3.1302,
      "step": 87600
    },
    {
      "epoch": 9.433860725433215,
      "grad_norm": 0.8324317932128906,
      "learning_rate": 3.421183062170025e-05,
      "loss": 3.1146,
      "step": 87650
    },
    {
      "epoch": 9.439242277472824,
      "grad_norm": 0.8815144300460815,
      "learning_rate": 3.388858959163883e-05,
      "loss": 3.1346,
      "step": 87700
    },
    {
      "epoch": 9.444623829512432,
      "grad_norm": 0.866257905960083,
      "learning_rate": 3.3565348561577416e-05,
      "loss": 3.1255,
      "step": 87750
    },
    {
      "epoch": 9.450005381552039,
      "grad_norm": 0.8640872240066528,
      "learning_rate": 3.3242107531516e-05,
      "loss": 3.1326,
      "step": 87800
    },
    {
      "epoch": 9.455386933591647,
      "grad_norm": 0.876515805721283,
      "learning_rate": 3.291886650145458e-05,
      "loss": 3.1376,
      "step": 87850
    },
    {
      "epoch": 9.460768485631256,
      "grad_norm": 0.879612147808075,
      "learning_rate": 3.260209029199439e-05,
      "loss": 3.1468,
      "step": 87900
    },
    {
      "epoch": 9.466150037670864,
      "grad_norm": 0.8799201846122742,
      "learning_rate": 3.227884926193298e-05,
      "loss": 3.1336,
      "step": 87950
    },
    {
      "epoch": 9.471531589710473,
      "grad_norm": 0.8598930239677429,
      "learning_rate": 3.195560823187157e-05,
      "loss": 3.1102,
      "step": 88000
    },
    {
      "epoch": 9.471531589710473,
      "eval_accuracy": 0.3935834602489607,
      "eval_loss": 3.309974431991577,
      "eval_runtime": 179.9534,
      "eval_samples_per_second": 100.087,
      "eval_steps_per_second": 6.257,
      "step": 88000
    },
    {
      "epoch": 9.476913141750082,
      "grad_norm": 0.8778337836265564,
      "learning_rate": 3.163236720181014e-05,
      "loss": 3.1379,
      "step": 88050
    },
    {
      "epoch": 9.482294693789688,
      "grad_norm": 0.8822025656700134,
      "learning_rate": 3.130912617174873e-05,
      "loss": 3.1384,
      "step": 88100
    },
    {
      "epoch": 9.487676245829297,
      "grad_norm": 0.927155077457428,
      "learning_rate": 3.098588514168732e-05,
      "loss": 3.1588,
      "step": 88150
    },
    {
      "epoch": 9.493057797868905,
      "grad_norm": 0.8520760536193848,
      "learning_rate": 3.06626441116259e-05,
      "loss": 3.1472,
      "step": 88200
    },
    {
      "epoch": 9.498439349908514,
      "grad_norm": 0.8926303386688232,
      "learning_rate": 3.0339403081564487e-05,
      "loss": 3.1344,
      "step": 88250
    },
    {
      "epoch": 9.503820901948123,
      "grad_norm": 0.8950697779655457,
      "learning_rate": 3.0016162051503068e-05,
      "loss": 3.135,
      "step": 88300
    },
    {
      "epoch": 9.50920245398773,
      "grad_norm": 0.8820440173149109,
      "learning_rate": 2.9692921021441652e-05,
      "loss": 3.1264,
      "step": 88350
    },
    {
      "epoch": 9.514584006027338,
      "grad_norm": 0.8579733371734619,
      "learning_rate": 2.9369679991380237e-05,
      "loss": 3.1423,
      "step": 88400
    },
    {
      "epoch": 9.519965558066946,
      "grad_norm": 0.8843384385108948,
      "learning_rate": 2.904643896131882e-05,
      "loss": 3.1311,
      "step": 88450
    },
    {
      "epoch": 9.525347110106555,
      "grad_norm": 0.8826112747192383,
      "learning_rate": 2.8723197931257406e-05,
      "loss": 3.1382,
      "step": 88500
    },
    {
      "epoch": 9.530728662146164,
      "grad_norm": 0.8869061470031738,
      "learning_rate": 2.8399956901195987e-05,
      "loss": 3.1228,
      "step": 88550
    },
    {
      "epoch": 9.536110214185772,
      "grad_norm": 0.8704985976219177,
      "learning_rate": 2.807671587113457e-05,
      "loss": 3.1219,
      "step": 88600
    },
    {
      "epoch": 9.541491766225379,
      "grad_norm": 0.9074740409851074,
      "learning_rate": 2.775347484107316e-05,
      "loss": 3.1322,
      "step": 88650
    },
    {
      "epoch": 9.546873318264987,
      "grad_norm": 0.8920992016792297,
      "learning_rate": 2.7430233811011744e-05,
      "loss": 3.1456,
      "step": 88700
    },
    {
      "epoch": 9.552254870304596,
      "grad_norm": 0.9204345941543579,
      "learning_rate": 2.7106992780950328e-05,
      "loss": 3.1497,
      "step": 88750
    },
    {
      "epoch": 9.557636422344205,
      "grad_norm": 0.8399198055267334,
      "learning_rate": 2.678375175088891e-05,
      "loss": 3.1334,
      "step": 88800
    },
    {
      "epoch": 9.563017974383813,
      "grad_norm": 0.8751394748687744,
      "learning_rate": 2.6460510720827494e-05,
      "loss": 3.1389,
      "step": 88850
    },
    {
      "epoch": 9.56839952642342,
      "grad_norm": 0.8489452600479126,
      "learning_rate": 2.6137269690766078e-05,
      "loss": 3.1277,
      "step": 88900
    },
    {
      "epoch": 9.573781078463028,
      "grad_norm": 0.8947514295578003,
      "learning_rate": 2.5814028660704663e-05,
      "loss": 3.1329,
      "step": 88950
    },
    {
      "epoch": 9.579162630502637,
      "grad_norm": 0.8745483160018921,
      "learning_rate": 2.5490787630643247e-05,
      "loss": 3.1177,
      "step": 89000
    },
    {
      "epoch": 9.579162630502637,
      "eval_accuracy": 0.3937311195235006,
      "eval_loss": 3.307589054107666,
      "eval_runtime": 180.1555,
      "eval_samples_per_second": 99.975,
      "eval_steps_per_second": 6.25,
      "step": 89000
    },
    {
      "epoch": 9.584544182542245,
      "grad_norm": 0.9029579162597656,
      "learning_rate": 2.5167546600581828e-05,
      "loss": 3.1263,
      "step": 89050
    },
    {
      "epoch": 9.589925734581854,
      "grad_norm": 0.8661422729492188,
      "learning_rate": 2.4844305570520416e-05,
      "loss": 3.1395,
      "step": 89100
    },
    {
      "epoch": 9.59530728662146,
      "grad_norm": 0.8325968980789185,
      "learning_rate": 2.4521064540459e-05,
      "loss": 3.1621,
      "step": 89150
    },
    {
      "epoch": 9.60068883866107,
      "grad_norm": 0.8808440566062927,
      "learning_rate": 2.4197823510397585e-05,
      "loss": 3.1332,
      "step": 89200
    },
    {
      "epoch": 9.606070390700678,
      "grad_norm": 0.8744810223579407,
      "learning_rate": 2.387458248033617e-05,
      "loss": 3.1255,
      "step": 89250
    },
    {
      "epoch": 9.611451942740286,
      "grad_norm": 0.8575035333633423,
      "learning_rate": 2.355134145027475e-05,
      "loss": 3.122,
      "step": 89300
    },
    {
      "epoch": 9.616833494779895,
      "grad_norm": 0.9315617084503174,
      "learning_rate": 2.3228100420213335e-05,
      "loss": 3.1325,
      "step": 89350
    },
    {
      "epoch": 9.622215046819504,
      "grad_norm": 0.8789585828781128,
      "learning_rate": 2.291132421075315e-05,
      "loss": 3.1261,
      "step": 89400
    },
    {
      "epoch": 9.62759659885911,
      "grad_norm": 0.8857566714286804,
      "learning_rate": 2.2588083180691733e-05,
      "loss": 3.1192,
      "step": 89450
    },
    {
      "epoch": 9.632978150898719,
      "grad_norm": 0.911800742149353,
      "learning_rate": 2.2264842150630318e-05,
      "loss": 3.1527,
      "step": 89500
    },
    {
      "epoch": 9.638359702938327,
      "grad_norm": 0.8703020811080933,
      "learning_rate": 2.1941601120568905e-05,
      "loss": 3.1369,
      "step": 89550
    },
    {
      "epoch": 9.643741254977936,
      "grad_norm": 0.878156840801239,
      "learning_rate": 2.1618360090507483e-05,
      "loss": 3.1499,
      "step": 89600
    },
    {
      "epoch": 9.649122807017545,
      "grad_norm": 0.950110912322998,
      "learning_rate": 2.129511906044607e-05,
      "loss": 3.1427,
      "step": 89650
    },
    {
      "epoch": 9.654504359057151,
      "grad_norm": 0.8555633425712585,
      "learning_rate": 2.0971878030384655e-05,
      "loss": 3.1397,
      "step": 89700
    },
    {
      "epoch": 9.65988591109676,
      "grad_norm": 0.8625248074531555,
      "learning_rate": 2.064863700032324e-05,
      "loss": 3.1312,
      "step": 89750
    },
    {
      "epoch": 9.665267463136368,
      "grad_norm": 0.8741613626480103,
      "learning_rate": 2.0325395970261824e-05,
      "loss": 3.1403,
      "step": 89800
    },
    {
      "epoch": 9.670649015175977,
      "grad_norm": 0.8547689914703369,
      "learning_rate": 2.0002154940200406e-05,
      "loss": 3.1416,
      "step": 89850
    },
    {
      "epoch": 9.676030567215586,
      "grad_norm": 0.8787466883659363,
      "learning_rate": 1.967891391013899e-05,
      "loss": 3.1432,
      "step": 89900
    },
    {
      "epoch": 9.681412119255192,
      "grad_norm": 0.9100810289382935,
      "learning_rate": 1.9355672880077575e-05,
      "loss": 3.1335,
      "step": 89950
    },
    {
      "epoch": 9.6867936712948,
      "grad_norm": 0.8840546607971191,
      "learning_rate": 1.9032431850016162e-05,
      "loss": 3.1193,
      "step": 90000
    },
    {
      "epoch": 9.6867936712948,
      "eval_accuracy": 0.3940206794695262,
      "eval_loss": 3.3056087493896484,
      "eval_runtime": 180.2428,
      "eval_samples_per_second": 99.926,
      "eval_steps_per_second": 6.247,
      "step": 90000
    },
    {
      "epoch": 9.69217522333441,
      "grad_norm": 0.8934172987937927,
      "learning_rate": 1.8709190819954744e-05,
      "loss": 3.1394,
      "step": 90050
    },
    {
      "epoch": 9.697556775374018,
      "grad_norm": 0.8928881287574768,
      "learning_rate": 1.8385949789893328e-05,
      "loss": 3.1375,
      "step": 90100
    },
    {
      "epoch": 9.702938327413626,
      "grad_norm": 0.8657145500183105,
      "learning_rate": 1.8062708759831913e-05,
      "loss": 3.1228,
      "step": 90150
    },
    {
      "epoch": 9.708319879453235,
      "grad_norm": 0.8868884444236755,
      "learning_rate": 1.7739467729770497e-05,
      "loss": 3.1286,
      "step": 90200
    },
    {
      "epoch": 9.713701431492842,
      "grad_norm": 0.8712571859359741,
      "learning_rate": 1.741622669970908e-05,
      "loss": 3.1284,
      "step": 90250
    },
    {
      "epoch": 9.71908298353245,
      "grad_norm": 0.9264222979545593,
      "learning_rate": 1.7092985669647666e-05,
      "loss": 3.1212,
      "step": 90300
    },
    {
      "epoch": 9.724464535572059,
      "grad_norm": 0.8676959276199341,
      "learning_rate": 1.676974463958625e-05,
      "loss": 3.1413,
      "step": 90350
    },
    {
      "epoch": 9.729846087611667,
      "grad_norm": 0.8672192692756653,
      "learning_rate": 1.6446503609524835e-05,
      "loss": 3.153,
      "step": 90400
    },
    {
      "epoch": 9.735227639651276,
      "grad_norm": 0.8709021210670471,
      "learning_rate": 1.612326257946342e-05,
      "loss": 3.143,
      "step": 90450
    },
    {
      "epoch": 9.740609191690883,
      "grad_norm": 0.8936007618904114,
      "learning_rate": 1.5800021549402004e-05,
      "loss": 3.1406,
      "step": 90500
    },
    {
      "epoch": 9.745990743730491,
      "grad_norm": 0.9075430035591125,
      "learning_rate": 1.5476780519340585e-05,
      "loss": 3.1166,
      "step": 90550
    },
    {
      "epoch": 9.7513722957701,
      "grad_norm": 0.881115198135376,
      "learning_rate": 1.5153539489279171e-05,
      "loss": 3.1441,
      "step": 90600
    },
    {
      "epoch": 9.756753847809708,
      "grad_norm": 0.9102500081062317,
      "learning_rate": 1.4830298459217756e-05,
      "loss": 3.1275,
      "step": 90650
    },
    {
      "epoch": 9.762135399849317,
      "grad_norm": 0.9028261303901672,
      "learning_rate": 1.4507057429156339e-05,
      "loss": 3.1364,
      "step": 90700
    },
    {
      "epoch": 9.767516951888926,
      "grad_norm": 0.8845318555831909,
      "learning_rate": 1.4183816399094925e-05,
      "loss": 3.1406,
      "step": 90750
    },
    {
      "epoch": 9.772898503928532,
      "grad_norm": 0.861149787902832,
      "learning_rate": 1.3860575369033508e-05,
      "loss": 3.1309,
      "step": 90800
    },
    {
      "epoch": 9.77828005596814,
      "grad_norm": 0.8670579195022583,
      "learning_rate": 1.3537334338972092e-05,
      "loss": 3.143,
      "step": 90850
    },
    {
      "epoch": 9.78366160800775,
      "grad_norm": 0.8900049924850464,
      "learning_rate": 1.3214093308910678e-05,
      "loss": 3.1461,
      "step": 90900
    },
    {
      "epoch": 9.789043160047358,
      "grad_norm": 0.851944625377655,
      "learning_rate": 1.2890852278849261e-05,
      "loss": 3.1187,
      "step": 90950
    },
    {
      "epoch": 9.794424712086967,
      "grad_norm": 0.8535908460617065,
      "learning_rate": 1.2567611248787846e-05,
      "loss": 3.1506,
      "step": 91000
    },
    {
      "epoch": 9.794424712086967,
      "eval_accuracy": 0.3942323352949888,
      "eval_loss": 3.3036699295043945,
      "eval_runtime": 180.5405,
      "eval_samples_per_second": 99.762,
      "eval_steps_per_second": 6.237,
      "step": 91000
    },
    {
      "epoch": 9.799806264126573,
      "grad_norm": 0.8900703191757202,
      "learning_rate": 1.2244370218726428e-05,
      "loss": 3.1451,
      "step": 91050
    },
    {
      "epoch": 9.805187816166182,
      "grad_norm": 0.8927872776985168,
      "learning_rate": 1.1921129188665013e-05,
      "loss": 3.1323,
      "step": 91100
    },
    {
      "epoch": 9.81056936820579,
      "grad_norm": 0.8900262713432312,
      "learning_rate": 1.1597888158603599e-05,
      "loss": 3.1302,
      "step": 91150
    },
    {
      "epoch": 9.815950920245399,
      "grad_norm": 0.8994975090026855,
      "learning_rate": 1.1274647128542182e-05,
      "loss": 3.1444,
      "step": 91200
    },
    {
      "epoch": 9.821332472285007,
      "grad_norm": 0.8664494752883911,
      "learning_rate": 1.0951406098480766e-05,
      "loss": 3.1298,
      "step": 91250
    },
    {
      "epoch": 9.826714024324616,
      "grad_norm": 0.8540681004524231,
      "learning_rate": 1.0628165068419349e-05,
      "loss": 3.1188,
      "step": 91300
    },
    {
      "epoch": 9.832095576364223,
      "grad_norm": 0.8558421730995178,
      "learning_rate": 1.0304924038357935e-05,
      "loss": 3.133,
      "step": 91350
    },
    {
      "epoch": 9.837477128403831,
      "grad_norm": 0.8718653917312622,
      "learning_rate": 9.98168300829652e-06,
      "loss": 3.1366,
      "step": 91400
    },
    {
      "epoch": 9.84285868044344,
      "grad_norm": 0.8789852857589722,
      "learning_rate": 9.658441978235103e-06,
      "loss": 3.1252,
      "step": 91450
    },
    {
      "epoch": 9.848240232483048,
      "grad_norm": 0.9523253440856934,
      "learning_rate": 9.335200948173687e-06,
      "loss": 3.1188,
      "step": 91500
    },
    {
      "epoch": 9.853621784522657,
      "grad_norm": 0.9213250279426575,
      "learning_rate": 9.011959918112272e-06,
      "loss": 3.1519,
      "step": 91550
    },
    {
      "epoch": 9.859003336562264,
      "grad_norm": 0.9170287251472473,
      "learning_rate": 8.688718888050856e-06,
      "loss": 3.1392,
      "step": 91600
    },
    {
      "epoch": 9.864384888601872,
      "grad_norm": 0.9603930115699768,
      "learning_rate": 8.36547785798944e-06,
      "loss": 3.1508,
      "step": 91650
    },
    {
      "epoch": 9.869766440641481,
      "grad_norm": 0.8868028521537781,
      "learning_rate": 8.042236827928023e-06,
      "loss": 3.1331,
      "step": 91700
    },
    {
      "epoch": 9.87514799268109,
      "grad_norm": 0.8630801439285278,
      "learning_rate": 7.71899579786661e-06,
      "loss": 3.1354,
      "step": 91750
    },
    {
      "epoch": 9.880529544720698,
      "grad_norm": 0.9191954135894775,
      "learning_rate": 7.395754767805193e-06,
      "loss": 3.1435,
      "step": 91800
    },
    {
      "epoch": 9.885911096760307,
      "grad_norm": 0.8929259181022644,
      "learning_rate": 7.072513737743777e-06,
      "loss": 3.1199,
      "step": 91850
    },
    {
      "epoch": 9.891292648799913,
      "grad_norm": 0.8772810101509094,
      "learning_rate": 6.749272707682361e-06,
      "loss": 3.1623,
      "step": 91900
    },
    {
      "epoch": 9.896674200839522,
      "grad_norm": 0.8684717416763306,
      "learning_rate": 6.432496498222174e-06,
      "loss": 3.1247,
      "step": 91950
    },
    {
      "epoch": 9.90205575287913,
      "grad_norm": 0.8589189052581787,
      "learning_rate": 6.1092554681607575e-06,
      "loss": 3.1476,
      "step": 92000
    },
    {
      "epoch": 9.90205575287913,
      "eval_accuracy": 0.3944637659460336,
      "eval_loss": 3.3020360469818115,
      "eval_runtime": 180.13,
      "eval_samples_per_second": 99.989,
      "eval_steps_per_second": 6.251,
      "step": 92000
    },
    {
      "epoch": 9.907437304918739,
      "grad_norm": 0.8615749478340149,
      "learning_rate": 5.792479258700571e-06,
      "loss": 3.1224,
      "step": 92050
    },
    {
      "epoch": 9.912818856958348,
      "grad_norm": 0.8723222017288208,
      "learning_rate": 5.469238228639154e-06,
      "loss": 3.1229,
      "step": 92100
    },
    {
      "epoch": 9.918200408997954,
      "grad_norm": 0.8633239269256592,
      "learning_rate": 5.145997198577739e-06,
      "loss": 3.1426,
      "step": 92150
    },
    {
      "epoch": 9.923581961037563,
      "grad_norm": 0.8642317652702332,
      "learning_rate": 4.822756168516323e-06,
      "loss": 3.1477,
      "step": 92200
    },
    {
      "epoch": 9.928963513077171,
      "grad_norm": 0.8761366605758667,
      "learning_rate": 4.499515138454908e-06,
      "loss": 3.1307,
      "step": 92250
    },
    {
      "epoch": 9.93434506511678,
      "grad_norm": 0.880110502243042,
      "learning_rate": 4.1762741083934915e-06,
      "loss": 3.1246,
      "step": 92300
    },
    {
      "epoch": 9.939726617156388,
      "grad_norm": 0.89411860704422,
      "learning_rate": 3.853033078332076e-06,
      "loss": 3.1239,
      "step": 92350
    },
    {
      "epoch": 9.945108169195997,
      "grad_norm": 0.8907015323638916,
      "learning_rate": 3.52979204827066e-06,
      "loss": 3.1234,
      "step": 92400
    },
    {
      "epoch": 9.950489721235604,
      "grad_norm": 0.8394994735717773,
      "learning_rate": 3.206551018209244e-06,
      "loss": 3.1339,
      "step": 92450
    },
    {
      "epoch": 9.955871273275212,
      "grad_norm": 0.8820265531539917,
      "learning_rate": 2.8833099881478286e-06,
      "loss": 3.1329,
      "step": 92500
    },
    {
      "epoch": 9.961252825314821,
      "grad_norm": 0.8925215005874634,
      "learning_rate": 2.5600689580864127e-06,
      "loss": 3.1347,
      "step": 92550
    },
    {
      "epoch": 9.96663437735443,
      "grad_norm": 0.9452439546585083,
      "learning_rate": 2.2368279280249972e-06,
      "loss": 3.1504,
      "step": 92600
    },
    {
      "epoch": 9.972015929394038,
      "grad_norm": 0.873985767364502,
      "learning_rate": 1.9135868979635813e-06,
      "loss": 3.1342,
      "step": 92650
    },
    {
      "epoch": 9.977397481433645,
      "grad_norm": 0.8428253531455994,
      "learning_rate": 1.5903458679021656e-06,
      "loss": 3.1225,
      "step": 92700
    },
    {
      "epoch": 9.982779033473253,
      "grad_norm": 0.9114113450050354,
      "learning_rate": 1.2671048378407496e-06,
      "loss": 3.1229,
      "step": 92750
    },
    {
      "epoch": 9.988160585512862,
      "grad_norm": 0.8490803837776184,
      "learning_rate": 9.438638077793341e-07,
      "loss": 3.1276,
      "step": 92800
    },
    {
      "epoch": 9.99354213755247,
      "grad_norm": 0.8661682605743408,
      "learning_rate": 6.206227777179182e-07,
      "loss": 3.1459,
      "step": 92850
    },
    {
      "epoch": 9.998923689592079,
      "grad_norm": 0.8826032876968384,
      "learning_rate": 2.9738174765650254e-07,
      "loss": 3.1464,
      "step": 92900
    },
    {
      "epoch": 10.0,
      "step": 92910,
      "total_flos": 7.7681598529536e+17,
      "train_loss": 3.46011078748004,
      "train_runtime": 80911.2202,
      "train_samples_per_second": 36.744,
      "train_steps_per_second": 1.148
    }
  ],
  "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": 7.7681598529536e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}