{
  "best_metric": 3.3091650009155273,
  "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_6910/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": 1.012463927268982,
      "learning_rate": 0.00028199999999999997,
      "loss": 9.0273,
      "step": 50
    },
    {
      "epoch": 0.010763104079216447,
      "grad_norm": 2.3169212341308594,
      "learning_rate": 0.0005819999999999999,
      "loss": 6.9892,
      "step": 100
    },
    {
      "epoch": 0.01614465611882467,
      "grad_norm": 1.2467113733291626,
      "learning_rate": 0.0005996961534317422,
      "loss": 6.5352,
      "step": 150
    },
    {
      "epoch": 0.021526208158432893,
      "grad_norm": 0.9622871279716492,
      "learning_rate": 0.0005993729124016807,
      "loss": 6.2488,
      "step": 200
    },
    {
      "epoch": 0.026907760198041114,
      "grad_norm": 2.9747376441955566,
      "learning_rate": 0.0005990496713716194,
      "loss": 6.0737,
      "step": 250
    },
    {
      "epoch": 0.03228931223764934,
      "grad_norm": 1.333069086074829,
      "learning_rate": 0.000598726430341558,
      "loss": 5.9643,
      "step": 300
    },
    {
      "epoch": 0.03767086427725756,
      "grad_norm": 2.613489866256714,
      "learning_rate": 0.0005984031893114966,
      "loss": 5.884,
      "step": 350
    },
    {
      "epoch": 0.04305241631686579,
      "grad_norm": 3.0947482585906982,
      "learning_rate": 0.0005980799482814351,
      "loss": 5.7964,
      "step": 400
    },
    {
      "epoch": 0.048433968356474004,
      "grad_norm": 1.2225009202957153,
      "learning_rate": 0.0005977567072513737,
      "loss": 5.729,
      "step": 450
    },
    {
      "epoch": 0.05381552039608223,
      "grad_norm": 1.5363976955413818,
      "learning_rate": 0.0005974334662213123,
      "loss": 5.6387,
      "step": 500
    },
    {
      "epoch": 0.05919707243569045,
      "grad_norm": 1.005996584892273,
      "learning_rate": 0.0005971102251912509,
      "loss": 5.5698,
      "step": 550
    },
    {
      "epoch": 0.06457862447529868,
      "grad_norm": 1.4028828144073486,
      "learning_rate": 0.0005967869841611895,
      "loss": 5.5079,
      "step": 600
    },
    {
      "epoch": 0.0699601765149069,
      "grad_norm": 1.3071937561035156,
      "learning_rate": 0.000596463743131128,
      "loss": 5.4329,
      "step": 650
    },
    {
      "epoch": 0.07534172855451512,
      "grad_norm": 1.0713516473770142,
      "learning_rate": 0.0005961405021010667,
      "loss": 5.3669,
      "step": 700
    },
    {
      "epoch": 0.08072328059412334,
      "grad_norm": 1.1204853057861328,
      "learning_rate": 0.0005958172610710052,
      "loss": 5.3316,
      "step": 750
    },
    {
      "epoch": 0.08610483263373157,
      "grad_norm": 1.0640522241592407,
      "learning_rate": 0.0005954940200409439,
      "loss": 5.2546,
      "step": 800
    },
    {
      "epoch": 0.09148638467333979,
      "grad_norm": 1.1928809881210327,
      "learning_rate": 0.0005951707790108824,
      "loss": 5.2056,
      "step": 850
    },
    {
      "epoch": 0.09686793671294801,
      "grad_norm": 1.4153245687484741,
      "learning_rate": 0.0005948475379808209,
      "loss": 5.1773,
      "step": 900
    },
    {
      "epoch": 0.10224948875255624,
      "grad_norm": 1.0611209869384766,
      "learning_rate": 0.0005945242969507596,
      "loss": 5.119,
      "step": 950
    },
    {
      "epoch": 0.10763104079216446,
      "grad_norm": 1.3633168935775757,
      "learning_rate": 0.0005942010559206981,
      "loss": 5.0801,
      "step": 1000
    },
    {
      "epoch": 0.10763104079216446,
      "eval_accuracy": 0.22714755420719737,
      "eval_loss": 5.0232367515563965,
      "eval_runtime": 185.2692,
      "eval_samples_per_second": 97.215,
      "eval_steps_per_second": 6.078,
      "step": 1000
    },
    {
      "epoch": 0.11301259283177269,
      "grad_norm": 0.7290101647377014,
      "learning_rate": 0.0005938778148906367,
      "loss": 5.0289,
      "step": 1050
    },
    {
      "epoch": 0.1183941448713809,
      "grad_norm": 1.2157477140426636,
      "learning_rate": 0.0005935545738605753,
      "loss": 5.0208,
      "step": 1100
    },
    {
      "epoch": 0.12377569691098914,
      "grad_norm": 1.1402803659439087,
      "learning_rate": 0.000593231332830514,
      "loss": 4.9719,
      "step": 1150
    },
    {
      "epoch": 0.12915724895059735,
      "grad_norm": 0.9659836292266846,
      "learning_rate": 0.0005929080918004525,
      "loss": 4.9851,
      "step": 1200
    },
    {
      "epoch": 0.13453880099020557,
      "grad_norm": 0.924443781375885,
      "learning_rate": 0.000592584850770391,
      "loss": 4.9411,
      "step": 1250
    },
    {
      "epoch": 0.1399203530298138,
      "grad_norm": 1.008213996887207,
      "learning_rate": 0.0005922616097403296,
      "loss": 4.9117,
      "step": 1300
    },
    {
      "epoch": 0.14530190506942203,
      "grad_norm": 1.2060256004333496,
      "learning_rate": 0.0005919383687102682,
      "loss": 4.8645,
      "step": 1350
    },
    {
      "epoch": 0.15068345710903025,
      "grad_norm": 0.8767279386520386,
      "learning_rate": 0.0005916151276802069,
      "loss": 4.827,
      "step": 1400
    },
    {
      "epoch": 0.15606500914863847,
      "grad_norm": 0.9375826120376587,
      "learning_rate": 0.0005912918866501454,
      "loss": 4.8417,
      "step": 1450
    },
    {
      "epoch": 0.16144656118824668,
      "grad_norm": 1.3111040592193604,
      "learning_rate": 0.000590968645620084,
      "loss": 4.7997,
      "step": 1500
    },
    {
      "epoch": 0.1668281132278549,
      "grad_norm": 1.154634714126587,
      "learning_rate": 0.0005906454045900226,
      "loss": 4.8048,
      "step": 1550
    },
    {
      "epoch": 0.17220966526746315,
      "grad_norm": 1.1702287197113037,
      "learning_rate": 0.0005903221635599612,
      "loss": 4.7529,
      "step": 1600
    },
    {
      "epoch": 0.17759121730707136,
      "grad_norm": 0.890550434589386,
      "learning_rate": 0.0005899989225298998,
      "loss": 4.7481,
      "step": 1650
    },
    {
      "epoch": 0.18297276934667958,
      "grad_norm": 0.9153488278388977,
      "learning_rate": 0.0005896756814998383,
      "loss": 4.7375,
      "step": 1700
    },
    {
      "epoch": 0.1883543213862878,
      "grad_norm": 0.869870126247406,
      "learning_rate": 0.0005893524404697769,
      "loss": 4.6931,
      "step": 1750
    },
    {
      "epoch": 0.19373587342589602,
      "grad_norm": 0.912928581237793,
      "learning_rate": 0.0005890291994397155,
      "loss": 4.6906,
      "step": 1800
    },
    {
      "epoch": 0.19911742546550426,
      "grad_norm": 1.2039134502410889,
      "learning_rate": 0.0005887059584096541,
      "loss": 4.6566,
      "step": 1850
    },
    {
      "epoch": 0.20449897750511248,
      "grad_norm": 0.8101952075958252,
      "learning_rate": 0.0005883827173795926,
      "loss": 4.6463,
      "step": 1900
    },
    {
      "epoch": 0.2098805295447207,
      "grad_norm": 0.8931390047073364,
      "learning_rate": 0.0005880594763495313,
      "loss": 4.6135,
      "step": 1950
    },
    {
      "epoch": 0.2152620815843289,
      "grad_norm": 0.8494347929954529,
      "learning_rate": 0.0005877362353194698,
      "loss": 4.596,
      "step": 2000
    },
    {
      "epoch": 0.2152620815843289,
      "eval_accuracy": 0.26789064869142265,
      "eval_loss": 4.530566692352295,
      "eval_runtime": 185.2933,
      "eval_samples_per_second": 97.203,
      "eval_steps_per_second": 6.077,
      "step": 2000
    },
    {
      "epoch": 0.22064363362393713,
      "grad_norm": 0.9353415966033936,
      "learning_rate": 0.0005874129942894084,
      "loss": 4.5881,
      "step": 2050
    },
    {
      "epoch": 0.22602518566354537,
      "grad_norm": 0.8548711538314819,
      "learning_rate": 0.000587089753259347,
      "loss": 4.5562,
      "step": 2100
    },
    {
      "epoch": 0.2314067377031536,
      "grad_norm": 0.7093924283981323,
      "learning_rate": 0.0005867665122292855,
      "loss": 4.5411,
      "step": 2150
    },
    {
      "epoch": 0.2367882897427618,
      "grad_norm": 0.8327104449272156,
      "learning_rate": 0.0005864432711992242,
      "loss": 4.5189,
      "step": 2200
    },
    {
      "epoch": 0.24216984178237003,
      "grad_norm": 1.3014960289001465,
      "learning_rate": 0.0005861200301691628,
      "loss": 4.5392,
      "step": 2250
    },
    {
      "epoch": 0.24755139382197827,
      "grad_norm": 3.5569815635681152,
      "learning_rate": 0.0005857967891391014,
      "loss": 4.5137,
      "step": 2300
    },
    {
      "epoch": 0.2529329458615865,
      "grad_norm": 0.847800076007843,
      "learning_rate": 0.0005854735481090399,
      "loss": 4.4963,
      "step": 2350
    },
    {
      "epoch": 0.2583144979011947,
      "grad_norm": 0.7744940519332886,
      "learning_rate": 0.0005851503070789784,
      "loss": 4.4674,
      "step": 2400
    },
    {
      "epoch": 0.2636960499408029,
      "grad_norm": 0.884650468826294,
      "learning_rate": 0.0005848270660489171,
      "loss": 4.4627,
      "step": 2450
    },
    {
      "epoch": 0.26907760198041114,
      "grad_norm": 0.7891039252281189,
      "learning_rate": 0.0005845038250188556,
      "loss": 4.434,
      "step": 2500
    },
    {
      "epoch": 0.27445915402001936,
      "grad_norm": 0.7952529788017273,
      "learning_rate": 0.0005841805839887943,
      "loss": 4.4319,
      "step": 2550
    },
    {
      "epoch": 0.2798407060596276,
      "grad_norm": 0.910103440284729,
      "learning_rate": 0.0005838573429587328,
      "loss": 4.4102,
      "step": 2600
    },
    {
      "epoch": 0.2852222580992358,
      "grad_norm": 1.2309143543243408,
      "learning_rate": 0.0005835341019286715,
      "loss": 4.4165,
      "step": 2650
    },
    {
      "epoch": 0.29060381013884407,
      "grad_norm": 0.7446426749229431,
      "learning_rate": 0.00058321086089861,
      "loss": 4.4104,
      "step": 2700
    },
    {
      "epoch": 0.2959853621784523,
      "grad_norm": 0.7638971209526062,
      "learning_rate": 0.0005828876198685486,
      "loss": 4.3802,
      "step": 2750
    },
    {
      "epoch": 0.3013669142180605,
      "grad_norm": 0.663977324962616,
      "learning_rate": 0.0005825643788384872,
      "loss": 4.3825,
      "step": 2800
    },
    {
      "epoch": 0.3067484662576687,
      "grad_norm": 0.9235962629318237,
      "learning_rate": 0.0005822411378084257,
      "loss": 4.3486,
      "step": 2850
    },
    {
      "epoch": 0.31213001829727693,
      "grad_norm": 0.9214714765548706,
      "learning_rate": 0.0005819178967783644,
      "loss": 4.3427,
      "step": 2900
    },
    {
      "epoch": 0.31751157033688515,
      "grad_norm": 0.7709562182426453,
      "learning_rate": 0.0005815946557483029,
      "loss": 4.3149,
      "step": 2950
    },
    {
      "epoch": 0.32289312237649337,
      "grad_norm": 0.9406850337982178,
      "learning_rate": 0.0005812714147182415,
      "loss": 4.3232,
      "step": 3000
    },
    {
      "epoch": 0.32289312237649337,
      "eval_accuracy": 0.2974132681039343,
      "eval_loss": 4.245625972747803,
      "eval_runtime": 185.0258,
      "eval_samples_per_second": 97.343,
      "eval_steps_per_second": 6.086,
      "step": 3000
    },
    {
      "epoch": 0.3282746744161016,
      "grad_norm": 0.776454746723175,
      "learning_rate": 0.0005809481736881801,
      "loss": 4.3005,
      "step": 3050
    },
    {
      "epoch": 0.3336562264557098,
      "grad_norm": 0.8669567108154297,
      "learning_rate": 0.0005806249326581187,
      "loss": 4.2941,
      "step": 3100
    },
    {
      "epoch": 0.3390377784953181,
      "grad_norm": 0.7872880101203918,
      "learning_rate": 0.0005803016916280573,
      "loss": 4.3118,
      "step": 3150
    },
    {
      "epoch": 0.3444193305349263,
      "grad_norm": 0.804801344871521,
      "learning_rate": 0.0005799784505979959,
      "loss": 4.2907,
      "step": 3200
    },
    {
      "epoch": 0.3498008825745345,
      "grad_norm": 0.8986364603042603,
      "learning_rate": 0.0005796552095679344,
      "loss": 4.276,
      "step": 3250
    },
    {
      "epoch": 0.35518243461414273,
      "grad_norm": 0.9777130484580994,
      "learning_rate": 0.000579331968537873,
      "loss": 4.2523,
      "step": 3300
    },
    {
      "epoch": 0.36056398665375095,
      "grad_norm": 0.6971152424812317,
      "learning_rate": 0.0005790087275078116,
      "loss": 4.252,
      "step": 3350
    },
    {
      "epoch": 0.36594553869335916,
      "grad_norm": 0.7938759326934814,
      "learning_rate": 0.0005786854864777502,
      "loss": 4.2613,
      "step": 3400
    },
    {
      "epoch": 0.3713270907329674,
      "grad_norm": 0.8243275284767151,
      "learning_rate": 0.0005783622454476888,
      "loss": 4.2493,
      "step": 3450
    },
    {
      "epoch": 0.3767086427725756,
      "grad_norm": 0.7867190837860107,
      "learning_rate": 0.0005780390044176273,
      "loss": 4.2182,
      "step": 3500
    },
    {
      "epoch": 0.3820901948121838,
      "grad_norm": 0.7070367932319641,
      "learning_rate": 0.000577715763387566,
      "loss": 4.1995,
      "step": 3550
    },
    {
      "epoch": 0.38747174685179203,
      "grad_norm": 0.7603887319564819,
      "learning_rate": 0.0005773925223575045,
      "loss": 4.2193,
      "step": 3600
    },
    {
      "epoch": 0.3928532988914003,
      "grad_norm": 0.7130142450332642,
      "learning_rate": 0.0005770692813274432,
      "loss": 4.2243,
      "step": 3650
    },
    {
      "epoch": 0.3982348509310085,
      "grad_norm": 0.735981822013855,
      "learning_rate": 0.0005767460402973817,
      "loss": 4.1956,
      "step": 3700
    },
    {
      "epoch": 0.40361640297061674,
      "grad_norm": 0.8146218657493591,
      "learning_rate": 0.0005764227992673203,
      "loss": 4.1842,
      "step": 3750
    },
    {
      "epoch": 0.40899795501022496,
      "grad_norm": 0.9765662550926208,
      "learning_rate": 0.0005760995582372589,
      "loss": 4.2111,
      "step": 3800
    },
    {
      "epoch": 0.4143795070498332,
      "grad_norm": 0.7952522039413452,
      "learning_rate": 0.0005757763172071974,
      "loss": 4.1702,
      "step": 3850
    },
    {
      "epoch": 0.4197610590894414,
      "grad_norm": 0.7104473114013672,
      "learning_rate": 0.000575453076177136,
      "loss": 4.1731,
      "step": 3900
    },
    {
      "epoch": 0.4251426111290496,
      "grad_norm": 0.6354160904884338,
      "learning_rate": 0.0005751298351470746,
      "loss": 4.1884,
      "step": 3950
    },
    {
      "epoch": 0.4305241631686578,
      "grad_norm": 0.8940901756286621,
      "learning_rate": 0.0005748065941170133,
      "loss": 4.1553,
      "step": 4000
    },
    {
      "epoch": 0.4305241631686578,
      "eval_accuracy": 0.3118288986476084,
      "eval_loss": 4.100198268890381,
      "eval_runtime": 186.202,
      "eval_samples_per_second": 96.728,
      "eval_steps_per_second": 6.047,
      "step": 4000
    },
    {
      "epoch": 0.43590571520826604,
      "grad_norm": 0.6561173796653748,
      "learning_rate": 0.0005744833530869518,
      "loss": 4.1551,
      "step": 4050
    },
    {
      "epoch": 0.44128726724787426,
      "grad_norm": 0.6744108200073242,
      "learning_rate": 0.0005741601120568903,
      "loss": 4.16,
      "step": 4100
    },
    {
      "epoch": 0.44666881928748253,
      "grad_norm": 0.6778728365898132,
      "learning_rate": 0.0005738368710268289,
      "loss": 4.1374,
      "step": 4150
    },
    {
      "epoch": 0.45205037132709075,
      "grad_norm": 0.6743355989456177,
      "learning_rate": 0.0005735136299967675,
      "loss": 4.1451,
      "step": 4200
    },
    {
      "epoch": 0.45743192336669897,
      "grad_norm": 0.7129095196723938,
      "learning_rate": 0.0005731903889667062,
      "loss": 4.136,
      "step": 4250
    },
    {
      "epoch": 0.4628134754063072,
      "grad_norm": 0.5477722883224487,
      "learning_rate": 0.0005728671479366447,
      "loss": 4.1317,
      "step": 4300
    },
    {
      "epoch": 0.4681950274459154,
      "grad_norm": 0.7229121327400208,
      "learning_rate": 0.0005725439069065833,
      "loss": 4.1304,
      "step": 4350
    },
    {
      "epoch": 0.4735765794855236,
      "grad_norm": 0.7431455254554749,
      "learning_rate": 0.0005722206658765219,
      "loss": 4.1227,
      "step": 4400
    },
    {
      "epoch": 0.47895813152513184,
      "grad_norm": 0.689704418182373,
      "learning_rate": 0.0005718974248464605,
      "loss": 4.1182,
      "step": 4450
    },
    {
      "epoch": 0.48433968356474005,
      "grad_norm": 0.63039231300354,
      "learning_rate": 0.000571574183816399,
      "loss": 4.1083,
      "step": 4500
    },
    {
      "epoch": 0.48972123560434827,
      "grad_norm": 0.6493752598762512,
      "learning_rate": 0.0005712509427863376,
      "loss": 4.1024,
      "step": 4550
    },
    {
      "epoch": 0.49510278764395654,
      "grad_norm": 0.6460608839988708,
      "learning_rate": 0.0005709277017562762,
      "loss": 4.1095,
      "step": 4600
    },
    {
      "epoch": 0.5004843396835648,
      "grad_norm": 0.7702430486679077,
      "learning_rate": 0.0005706044607262148,
      "loss": 4.1043,
      "step": 4650
    },
    {
      "epoch": 0.505865891723173,
      "grad_norm": 0.7992980480194092,
      "learning_rate": 0.0005702812196961534,
      "loss": 4.0978,
      "step": 4700
    },
    {
      "epoch": 0.5112474437627812,
      "grad_norm": 0.6669775247573853,
      "learning_rate": 0.0005699579786660919,
      "loss": 4.1093,
      "step": 4750
    },
    {
      "epoch": 0.5166289958023894,
      "grad_norm": 0.7825795412063599,
      "learning_rate": 0.0005696347376360306,
      "loss": 4.0798,
      "step": 4800
    },
    {
      "epoch": 0.5220105478419976,
      "grad_norm": 0.6550343036651611,
      "learning_rate": 0.0005693114966059691,
      "loss": 4.0798,
      "step": 4850
    },
    {
      "epoch": 0.5273920998816058,
      "grad_norm": 0.6412573456764221,
      "learning_rate": 0.0005689882555759077,
      "loss": 4.0601,
      "step": 4900
    },
    {
      "epoch": 0.5327736519212141,
      "grad_norm": 0.5895215272903442,
      "learning_rate": 0.0005686650145458463,
      "loss": 4.0578,
      "step": 4950
    },
    {
      "epoch": 0.5381552039608223,
      "grad_norm": 0.8030828237533569,
      "learning_rate": 0.0005683417735157848,
      "loss": 4.0665,
      "step": 5000
    },
    {
      "epoch": 0.5381552039608223,
      "eval_accuracy": 0.31988942178301344,
      "eval_loss": 4.002148628234863,
      "eval_runtime": 185.4564,
      "eval_samples_per_second": 97.117,
      "eval_steps_per_second": 6.072,
      "step": 5000
    },
    {
      "epoch": 0.5435367560004305,
      "grad_norm": 0.6629536151885986,
      "learning_rate": 0.0005680185324857235,
      "loss": 4.0312,
      "step": 5050
    },
    {
      "epoch": 0.5489183080400387,
      "grad_norm": 0.7006453275680542,
      "learning_rate": 0.0005676952914556621,
      "loss": 4.0681,
      "step": 5100
    },
    {
      "epoch": 0.5542998600796469,
      "grad_norm": 0.6232185959815979,
      "learning_rate": 0.0005673720504256007,
      "loss": 4.0486,
      "step": 5150
    },
    {
      "epoch": 0.5596814121192552,
      "grad_norm": 0.6412311792373657,
      "learning_rate": 0.0005670488093955392,
      "loss": 4.0512,
      "step": 5200
    },
    {
      "epoch": 0.5650629641588634,
      "grad_norm": 0.5949878692626953,
      "learning_rate": 0.0005667255683654777,
      "loss": 4.0554,
      "step": 5250
    },
    {
      "epoch": 0.5704445161984716,
      "grad_norm": 0.6533347964286804,
      "learning_rate": 0.0005664023273354164,
      "loss": 4.0577,
      "step": 5300
    },
    {
      "epoch": 0.5758260682380799,
      "grad_norm": 0.8103878498077393,
      "learning_rate": 0.0005660790863053549,
      "loss": 4.0541,
      "step": 5350
    },
    {
      "epoch": 0.5812076202776881,
      "grad_norm": 0.9533542990684509,
      "learning_rate": 0.0005657558452752936,
      "loss": 4.0466,
      "step": 5400
    },
    {
      "epoch": 0.5865891723172963,
      "grad_norm": 0.6021135449409485,
      "learning_rate": 0.0005654326042452321,
      "loss": 4.0231,
      "step": 5450
    },
    {
      "epoch": 0.5919707243569046,
      "grad_norm": 0.7371340394020081,
      "learning_rate": 0.0005651093632151708,
      "loss": 4.0395,
      "step": 5500
    },
    {
      "epoch": 0.5973522763965128,
      "grad_norm": 0.597688615322113,
      "learning_rate": 0.0005647861221851093,
      "loss": 4.0398,
      "step": 5550
    },
    {
      "epoch": 0.602733828436121,
      "grad_norm": 0.6549320816993713,
      "learning_rate": 0.0005644628811550479,
      "loss": 4.0115,
      "step": 5600
    },
    {
      "epoch": 0.6081153804757292,
      "grad_norm": 0.7144424319267273,
      "learning_rate": 0.0005641396401249865,
      "loss": 4.0202,
      "step": 5650
    },
    {
      "epoch": 0.6134969325153374,
      "grad_norm": 0.6588114500045776,
      "learning_rate": 0.000563816399094925,
      "loss": 4.0153,
      "step": 5700
    },
    {
      "epoch": 0.6188784845549457,
      "grad_norm": 0.6274276971817017,
      "learning_rate": 0.0005634931580648637,
      "loss": 4.0137,
      "step": 5750
    },
    {
      "epoch": 0.6242600365945539,
      "grad_norm": 0.580619752407074,
      "learning_rate": 0.0005631699170348022,
      "loss": 4.0198,
      "step": 5800
    },
    {
      "epoch": 0.6296415886341621,
      "grad_norm": 0.7373949885368347,
      "learning_rate": 0.0005628466760047408,
      "loss": 4.0187,
      "step": 5850
    },
    {
      "epoch": 0.6350231406737703,
      "grad_norm": 0.657120943069458,
      "learning_rate": 0.0005625234349746794,
      "loss": 4.0095,
      "step": 5900
    },
    {
      "epoch": 0.6404046927133785,
      "grad_norm": 0.6227992177009583,
      "learning_rate": 0.000562200193944618,
      "loss": 3.9925,
      "step": 5950
    },
    {
      "epoch": 0.6457862447529867,
      "grad_norm": 0.6100704669952393,
      "learning_rate": 0.0005618769529145566,
      "loss": 3.9882,
      "step": 6000
    },
    {
      "epoch": 0.6457862447529867,
      "eval_accuracy": 0.3271768796216185,
      "eval_loss": 3.931014060974121,
      "eval_runtime": 185.0012,
      "eval_samples_per_second": 97.356,
      "eval_steps_per_second": 6.086,
      "step": 6000
    },
    {
      "epoch": 0.651167796792595,
      "grad_norm": 0.6571896076202393,
      "learning_rate": 0.0005615537118844952,
      "loss": 3.9853,
      "step": 6050
    },
    {
      "epoch": 0.6565493488322032,
      "grad_norm": 0.7249941825866699,
      "learning_rate": 0.0005612304708544337,
      "loss": 3.9897,
      "step": 6100
    },
    {
      "epoch": 0.6619309008718114,
      "grad_norm": 0.6451249122619629,
      "learning_rate": 0.0005609072298243723,
      "loss": 3.9789,
      "step": 6150
    },
    {
      "epoch": 0.6673124529114196,
      "grad_norm": 0.6867013573646545,
      "learning_rate": 0.0005605839887943109,
      "loss": 3.9974,
      "step": 6200
    },
    {
      "epoch": 0.6726940049510278,
      "grad_norm": 0.6106132864952087,
      "learning_rate": 0.0005602607477642495,
      "loss": 3.9823,
      "step": 6250
    },
    {
      "epoch": 0.6780755569906362,
      "grad_norm": 0.65805584192276,
      "learning_rate": 0.0005599375067341881,
      "loss": 4.0005,
      "step": 6300
    },
    {
      "epoch": 0.6834571090302444,
      "grad_norm": 0.6924935579299927,
      "learning_rate": 0.0005596142657041266,
      "loss": 3.9689,
      "step": 6350
    },
    {
      "epoch": 0.6888386610698526,
      "grad_norm": 0.7501756548881531,
      "learning_rate": 0.0005592910246740653,
      "loss": 3.9828,
      "step": 6400
    },
    {
      "epoch": 0.6942202131094608,
      "grad_norm": 0.5967255234718323,
      "learning_rate": 0.0005589677836440038,
      "loss": 3.9566,
      "step": 6450
    },
    {
      "epoch": 0.699601765149069,
      "grad_norm": 0.6848652362823486,
      "learning_rate": 0.0005586445426139425,
      "loss": 3.977,
      "step": 6500
    },
    {
      "epoch": 0.7049833171886772,
      "grad_norm": 0.5733135938644409,
      "learning_rate": 0.000558321301583881,
      "loss": 3.9586,
      "step": 6550
    },
    {
      "epoch": 0.7103648692282855,
      "grad_norm": 0.6924303770065308,
      "learning_rate": 0.0005579980605538196,
      "loss": 3.9619,
      "step": 6600
    },
    {
      "epoch": 0.7157464212678937,
      "grad_norm": 0.600371778011322,
      "learning_rate": 0.0005576748195237582,
      "loss": 3.9636,
      "step": 6650
    },
    {
      "epoch": 0.7211279733075019,
      "grad_norm": 0.6366176605224609,
      "learning_rate": 0.0005573515784936967,
      "loss": 3.9436,
      "step": 6700
    },
    {
      "epoch": 0.7265095253471101,
      "grad_norm": 0.6640421152114868,
      "learning_rate": 0.0005570283374636353,
      "loss": 3.947,
      "step": 6750
    },
    {
      "epoch": 0.7318910773867183,
      "grad_norm": 0.608975887298584,
      "learning_rate": 0.0005567050964335739,
      "loss": 3.9586,
      "step": 6800
    },
    {
      "epoch": 0.7372726294263265,
      "grad_norm": 0.6333919167518616,
      "learning_rate": 0.0005563818554035126,
      "loss": 3.953,
      "step": 6850
    },
    {
      "epoch": 0.7426541814659348,
      "grad_norm": 0.6278727650642395,
      "learning_rate": 0.0005560586143734511,
      "loss": 3.9432,
      "step": 6900
    },
    {
      "epoch": 0.748035733505543,
      "grad_norm": 0.6481143832206726,
      "learning_rate": 0.0005557353733433896,
      "loss": 3.9275,
      "step": 6950
    },
    {
      "epoch": 0.7534172855451512,
      "grad_norm": 0.572243332862854,
      "learning_rate": 0.0005554121323133283,
      "loss": 3.9394,
      "step": 7000
    },
    {
      "epoch": 0.7534172855451512,
      "eval_accuracy": 0.33199063716335125,
      "eval_loss": 3.873633623123169,
      "eval_runtime": 185.2534,
      "eval_samples_per_second": 97.224,
      "eval_steps_per_second": 6.078,
      "step": 7000
    },
    {
      "epoch": 0.7587988375847594,
      "grad_norm": 0.6098677515983582,
      "learning_rate": 0.0005550888912832668,
      "loss": 3.933,
      "step": 7050
    },
    {
      "epoch": 0.7641803896243676,
      "grad_norm": 0.6433761715888977,
      "learning_rate": 0.0005547656502532055,
      "loss": 3.9594,
      "step": 7100
    },
    {
      "epoch": 0.7695619416639758,
      "grad_norm": 0.5848661065101624,
      "learning_rate": 0.000554442409223144,
      "loss": 3.9278,
      "step": 7150
    },
    {
      "epoch": 0.7749434937035841,
      "grad_norm": 0.5872640013694763,
      "learning_rate": 0.0005541191681930826,
      "loss": 3.9247,
      "step": 7200
    },
    {
      "epoch": 0.7803250457431924,
      "grad_norm": 0.6132564544677734,
      "learning_rate": 0.0005537959271630212,
      "loss": 3.9117,
      "step": 7250
    },
    {
      "epoch": 0.7857065977828006,
      "grad_norm": 0.6102502346038818,
      "learning_rate": 0.0005534726861329598,
      "loss": 3.9115,
      "step": 7300
    },
    {
      "epoch": 0.7910881498224088,
      "grad_norm": 0.6094052195549011,
      "learning_rate": 0.0005531494451028983,
      "loss": 3.9102,
      "step": 7350
    },
    {
      "epoch": 0.796469701862017,
      "grad_norm": 0.6002593636512756,
      "learning_rate": 0.0005528262040728369,
      "loss": 3.9191,
      "step": 7400
    },
    {
      "epoch": 0.8018512539016253,
      "grad_norm": 0.6475902795791626,
      "learning_rate": 0.0005525029630427755,
      "loss": 3.9288,
      "step": 7450
    },
    {
      "epoch": 0.8072328059412335,
      "grad_norm": 0.5730218887329102,
      "learning_rate": 0.0005521797220127141,
      "loss": 3.911,
      "step": 7500
    },
    {
      "epoch": 0.8126143579808417,
      "grad_norm": 0.5757044553756714,
      "learning_rate": 0.0005518564809826527,
      "loss": 3.9267,
      "step": 7550
    },
    {
      "epoch": 0.8179959100204499,
      "grad_norm": 0.5882857441902161,
      "learning_rate": 0.0005515332399525912,
      "loss": 3.9053,
      "step": 7600
    },
    {
      "epoch": 0.8233774620600581,
      "grad_norm": 0.6268393993377686,
      "learning_rate": 0.0005512099989225299,
      "loss": 3.922,
      "step": 7650
    },
    {
      "epoch": 0.8287590140996663,
      "grad_norm": 0.5791593194007874,
      "learning_rate": 0.0005508867578924685,
      "loss": 3.9071,
      "step": 7700
    },
    {
      "epoch": 0.8341405661392746,
      "grad_norm": 0.5377199053764343,
      "learning_rate": 0.000550563516862407,
      "loss": 3.9082,
      "step": 7750
    },
    {
      "epoch": 0.8395221181788828,
      "grad_norm": 0.6483974456787109,
      "learning_rate": 0.0005502402758323456,
      "loss": 3.9074,
      "step": 7800
    },
    {
      "epoch": 0.844903670218491,
      "grad_norm": 0.5979174375534058,
      "learning_rate": 0.0005499170348022841,
      "loss": 3.8974,
      "step": 7850
    },
    {
      "epoch": 0.8502852222580992,
      "grad_norm": 0.5983253717422485,
      "learning_rate": 0.0005495937937722228,
      "loss": 3.9058,
      "step": 7900
    },
    {
      "epoch": 0.8556667742977074,
      "grad_norm": 0.7089122533798218,
      "learning_rate": 0.0005492705527421614,
      "loss": 3.8775,
      "step": 7950
    },
    {
      "epoch": 0.8610483263373157,
      "grad_norm": 0.6263192892074585,
      "learning_rate": 0.0005489473117121,
      "loss": 3.8898,
      "step": 8000
    },
    {
      "epoch": 0.8610483263373157,
      "eval_accuracy": 0.33647593702522083,
      "eval_loss": 3.8263087272644043,
      "eval_runtime": 185.2286,
      "eval_samples_per_second": 97.237,
      "eval_steps_per_second": 6.079,
      "step": 8000
    },
    {
      "epoch": 0.8664298783769239,
      "grad_norm": 0.6070656180381775,
      "learning_rate": 0.0005486240706820385,
      "loss": 3.8872,
      "step": 8050
    },
    {
      "epoch": 0.8718114304165321,
      "grad_norm": 0.6354324817657471,
      "learning_rate": 0.0005483008296519772,
      "loss": 3.8835,
      "step": 8100
    },
    {
      "epoch": 0.8771929824561403,
      "grad_norm": 0.6234643459320068,
      "learning_rate": 0.0005479775886219157,
      "loss": 3.8919,
      "step": 8150
    },
    {
      "epoch": 0.8825745344957485,
      "grad_norm": 0.6626437306404114,
      "learning_rate": 0.0005476543475918542,
      "loss": 3.8667,
      "step": 8200
    },
    {
      "epoch": 0.8879560865353568,
      "grad_norm": 0.6201016902923584,
      "learning_rate": 0.0005473311065617929,
      "loss": 3.8832,
      "step": 8250
    },
    {
      "epoch": 0.8933376385749651,
      "grad_norm": 0.5627908706665039,
      "learning_rate": 0.0005470078655317314,
      "loss": 3.8667,
      "step": 8300
    },
    {
      "epoch": 0.8987191906145733,
      "grad_norm": 0.5496158003807068,
      "learning_rate": 0.0005466846245016701,
      "loss": 3.885,
      "step": 8350
    },
    {
      "epoch": 0.9041007426541815,
      "grad_norm": 0.5965729355812073,
      "learning_rate": 0.0005463613834716086,
      "loss": 3.8814,
      "step": 8400
    },
    {
      "epoch": 0.9094822946937897,
      "grad_norm": 0.5938984155654907,
      "learning_rate": 0.0005460381424415472,
      "loss": 3.8825,
      "step": 8450
    },
    {
      "epoch": 0.9148638467333979,
      "grad_norm": 0.7846994996070862,
      "learning_rate": 0.0005457149014114858,
      "loss": 3.8822,
      "step": 8500
    },
    {
      "epoch": 0.9202453987730062,
      "grad_norm": 0.6211069226264954,
      "learning_rate": 0.0005453916603814243,
      "loss": 3.8677,
      "step": 8550
    },
    {
      "epoch": 0.9256269508126144,
      "grad_norm": 0.6226423978805542,
      "learning_rate": 0.000545068419351363,
      "loss": 3.8837,
      "step": 8600
    },
    {
      "epoch": 0.9310085028522226,
      "grad_norm": 0.587722897529602,
      "learning_rate": 0.0005447451783213015,
      "loss": 3.858,
      "step": 8650
    },
    {
      "epoch": 0.9363900548918308,
      "grad_norm": 0.5344767570495605,
      "learning_rate": 0.0005444219372912401,
      "loss": 3.8538,
      "step": 8700
    },
    {
      "epoch": 0.941771606931439,
      "grad_norm": 0.5695492625236511,
      "learning_rate": 0.0005440986962611787,
      "loss": 3.8608,
      "step": 8750
    },
    {
      "epoch": 0.9471531589710472,
      "grad_norm": 0.6132233738899231,
      "learning_rate": 0.0005437754552311173,
      "loss": 3.8484,
      "step": 8800
    },
    {
      "epoch": 0.9525347110106555,
      "grad_norm": 0.592411458492279,
      "learning_rate": 0.0005434522142010559,
      "loss": 3.8428,
      "step": 8850
    },
    {
      "epoch": 0.9579162630502637,
      "grad_norm": 0.6013568043708801,
      "learning_rate": 0.0005431289731709945,
      "loss": 3.859,
      "step": 8900
    },
    {
      "epoch": 0.9632978150898719,
      "grad_norm": 0.5537044405937195,
      "learning_rate": 0.000542805732140933,
      "loss": 3.8675,
      "step": 8950
    },
    {
      "epoch": 0.9686793671294801,
      "grad_norm": 0.6780604720115662,
      "learning_rate": 0.0005424824911108716,
      "loss": 3.8549,
      "step": 9000
    },
    {
      "epoch": 0.9686793671294801,
      "eval_accuracy": 0.3402423893813098,
      "eval_loss": 3.786595582962036,
      "eval_runtime": 184.9927,
      "eval_samples_per_second": 97.361,
      "eval_steps_per_second": 6.087,
      "step": 9000
    },
    {
      "epoch": 0.9740609191690883,
      "grad_norm": 0.5153192281723022,
      "learning_rate": 0.0005421592500808102,
      "loss": 3.839,
      "step": 9050
    },
    {
      "epoch": 0.9794424712086965,
      "grad_norm": 0.6029755473136902,
      "learning_rate": 0.0005418360090507488,
      "loss": 3.8658,
      "step": 9100
    },
    {
      "epoch": 0.9848240232483048,
      "grad_norm": 0.5533767938613892,
      "learning_rate": 0.0005415127680206874,
      "loss": 3.8584,
      "step": 9150
    },
    {
      "epoch": 0.9902055752879131,
      "grad_norm": 0.5732285976409912,
      "learning_rate": 0.0005411895269906259,
      "loss": 3.8616,
      "step": 9200
    },
    {
      "epoch": 0.9955871273275213,
      "grad_norm": 0.4923030138015747,
      "learning_rate": 0.0005408662859605646,
      "loss": 3.8439,
      "step": 9250
    },
    {
      "epoch": 1.0009686793671295,
      "grad_norm": 0.6030671000480652,
      "learning_rate": 0.0005405430449305031,
      "loss": 3.8309,
      "step": 9300
    },
    {
      "epoch": 1.0063502314067376,
      "grad_norm": 0.632311224937439,
      "learning_rate": 0.0005402198039004416,
      "loss": 3.7747,
      "step": 9350
    },
    {
      "epoch": 1.011731783446346,
      "grad_norm": 0.6197382211685181,
      "learning_rate": 0.0005398965628703803,
      "loss": 3.7833,
      "step": 9400
    },
    {
      "epoch": 1.017113335485954,
      "grad_norm": 0.5493740439414978,
      "learning_rate": 0.0005395733218403189,
      "loss": 3.7919,
      "step": 9450
    },
    {
      "epoch": 1.0224948875255624,
      "grad_norm": 0.5452463030815125,
      "learning_rate": 0.0005392500808102575,
      "loss": 3.7921,
      "step": 9500
    },
    {
      "epoch": 1.0278764395651705,
      "grad_norm": 0.5887305736541748,
      "learning_rate": 0.000538926839780196,
      "loss": 3.7657,
      "step": 9550
    },
    {
      "epoch": 1.0332579916047788,
      "grad_norm": 0.5790954232215881,
      "learning_rate": 0.0005386035987501346,
      "loss": 3.7805,
      "step": 9600
    },
    {
      "epoch": 1.0386395436443872,
      "grad_norm": 0.6792527437210083,
      "learning_rate": 0.0005382803577200732,
      "loss": 3.7755,
      "step": 9650
    },
    {
      "epoch": 1.0440210956839953,
      "grad_norm": 0.5296483039855957,
      "learning_rate": 0.0005379571166900119,
      "loss": 3.7712,
      "step": 9700
    },
    {
      "epoch": 1.0494026477236036,
      "grad_norm": 0.562563955783844,
      "learning_rate": 0.0005376338756599504,
      "loss": 3.7794,
      "step": 9750
    },
    {
      "epoch": 1.0547841997632117,
      "grad_norm": 0.5288910865783691,
      "learning_rate": 0.0005373106346298889,
      "loss": 3.7798,
      "step": 9800
    },
    {
      "epoch": 1.06016575180282,
      "grad_norm": 0.5700885653495789,
      "learning_rate": 0.0005369873935998276,
      "loss": 3.7699,
      "step": 9850
    },
    {
      "epoch": 1.0655473038424281,
      "grad_norm": 0.61332106590271,
      "learning_rate": 0.0005366641525697661,
      "loss": 3.7863,
      "step": 9900
    },
    {
      "epoch": 1.0709288558820365,
      "grad_norm": 0.6568296551704407,
      "learning_rate": 0.0005363409115397048,
      "loss": 3.7771,
      "step": 9950
    },
    {
      "epoch": 1.0763104079216446,
      "grad_norm": 0.5686736106872559,
      "learning_rate": 0.0005360176705096433,
      "loss": 3.7728,
      "step": 10000
    },
    {
      "epoch": 1.0763104079216446,
      "eval_accuracy": 0.3436631082483731,
      "eval_loss": 3.7558090686798096,
      "eval_runtime": 185.26,
      "eval_samples_per_second": 97.22,
      "eval_steps_per_second": 6.078,
      "step": 10000
    },
    {
      "epoch": 1.081691959961253,
      "grad_norm": 0.5301960706710815,
      "learning_rate": 0.0005356944294795819,
      "loss": 3.7908,
      "step": 10050
    },
    {
      "epoch": 1.087073512000861,
      "grad_norm": 0.6727739572525024,
      "learning_rate": 0.0005353711884495205,
      "loss": 3.7768,
      "step": 10100
    },
    {
      "epoch": 1.0924550640404693,
      "grad_norm": 0.5785535573959351,
      "learning_rate": 0.000535047947419459,
      "loss": 3.7529,
      "step": 10150
    },
    {
      "epoch": 1.0978366160800774,
      "grad_norm": 0.6064712405204773,
      "learning_rate": 0.0005347247063893976,
      "loss": 3.7959,
      "step": 10200
    },
    {
      "epoch": 1.1032181681196858,
      "grad_norm": 0.5543254017829895,
      "learning_rate": 0.0005344014653593362,
      "loss": 3.7705,
      "step": 10250
    },
    {
      "epoch": 1.1085997201592939,
      "grad_norm": 0.5798726677894592,
      "learning_rate": 0.0005340782243292748,
      "loss": 3.7646,
      "step": 10300
    },
    {
      "epoch": 1.1139812721989022,
      "grad_norm": 0.5178470611572266,
      "learning_rate": 0.0005337549832992134,
      "loss": 3.7769,
      "step": 10350
    },
    {
      "epoch": 1.1193628242385103,
      "grad_norm": 0.5780012607574463,
      "learning_rate": 0.000533431742269152,
      "loss": 3.757,
      "step": 10400
    },
    {
      "epoch": 1.1247443762781186,
      "grad_norm": 0.5791512131690979,
      "learning_rate": 0.0005331085012390905,
      "loss": 3.7652,
      "step": 10450
    },
    {
      "epoch": 1.1301259283177267,
      "grad_norm": 0.5443729162216187,
      "learning_rate": 0.0005327852602090292,
      "loss": 3.7565,
      "step": 10500
    },
    {
      "epoch": 1.135507480357335,
      "grad_norm": 0.6184217929840088,
      "learning_rate": 0.0005324620191789678,
      "loss": 3.777,
      "step": 10550
    },
    {
      "epoch": 1.1408890323969434,
      "grad_norm": 0.5406556129455566,
      "learning_rate": 0.0005321387781489063,
      "loss": 3.7673,
      "step": 10600
    },
    {
      "epoch": 1.1462705844365515,
      "grad_norm": 0.6513271927833557,
      "learning_rate": 0.0005318155371188449,
      "loss": 3.7676,
      "step": 10650
    },
    {
      "epoch": 1.1516521364761596,
      "grad_norm": 0.5891707539558411,
      "learning_rate": 0.0005314922960887834,
      "loss": 3.7653,
      "step": 10700
    },
    {
      "epoch": 1.157033688515768,
      "grad_norm": 0.6060640811920166,
      "learning_rate": 0.0005311690550587221,
      "loss": 3.7767,
      "step": 10750
    },
    {
      "epoch": 1.1624152405553763,
      "grad_norm": 0.5950350761413574,
      "learning_rate": 0.0005308458140286607,
      "loss": 3.7667,
      "step": 10800
    },
    {
      "epoch": 1.1677967925949844,
      "grad_norm": 0.8926947116851807,
      "learning_rate": 0.0005305225729985993,
      "loss": 3.7527,
      "step": 10850
    },
    {
      "epoch": 1.1731783446345927,
      "grad_norm": 0.6724712252616882,
      "learning_rate": 0.0005301993319685378,
      "loss": 3.7644,
      "step": 10900
    },
    {
      "epoch": 1.1785598966742008,
      "grad_norm": 0.6039562821388245,
      "learning_rate": 0.0005298760909384765,
      "loss": 3.7528,
      "step": 10950
    },
    {
      "epoch": 1.1839414487138091,
      "grad_norm": 0.5568313598632812,
      "learning_rate": 0.000529552849908415,
      "loss": 3.756,
      "step": 11000
    },
    {
      "epoch": 1.1839414487138091,
      "eval_accuracy": 0.3459454708098041,
      "eval_loss": 3.732597827911377,
      "eval_runtime": 185.2698,
      "eval_samples_per_second": 97.215,
      "eval_steps_per_second": 6.078,
      "step": 11000
    },
    {
      "epoch": 1.1893230007534172,
      "grad_norm": 0.5328220129013062,
      "learning_rate": 0.0005292296088783535,
      "loss": 3.7575,
      "step": 11050
    },
    {
      "epoch": 1.1947045527930256,
      "grad_norm": 0.5747320055961609,
      "learning_rate": 0.0005289063678482922,
      "loss": 3.7423,
      "step": 11100
    },
    {
      "epoch": 1.2000861048326337,
      "grad_norm": 0.5885644555091858,
      "learning_rate": 0.0005285831268182307,
      "loss": 3.7712,
      "step": 11150
    },
    {
      "epoch": 1.205467656872242,
      "grad_norm": 0.6482129693031311,
      "learning_rate": 0.0005282598857881694,
      "loss": 3.7709,
      "step": 11200
    },
    {
      "epoch": 1.21084920891185,
      "grad_norm": 0.5408961772918701,
      "learning_rate": 0.0005279366447581079,
      "loss": 3.7681,
      "step": 11250
    },
    {
      "epoch": 1.2162307609514584,
      "grad_norm": 0.6509531736373901,
      "learning_rate": 0.0005276134037280465,
      "loss": 3.7586,
      "step": 11300
    },
    {
      "epoch": 1.2216123129910665,
      "grad_norm": 0.6265044808387756,
      "learning_rate": 0.0005272901626979851,
      "loss": 3.7503,
      "step": 11350
    },
    {
      "epoch": 1.2269938650306749,
      "grad_norm": 0.5076571702957153,
      "learning_rate": 0.0005269669216679236,
      "loss": 3.7552,
      "step": 11400
    },
    {
      "epoch": 1.232375417070283,
      "grad_norm": 0.6145132780075073,
      "learning_rate": 0.0005266436806378623,
      "loss": 3.7523,
      "step": 11450
    },
    {
      "epoch": 1.2377569691098913,
      "grad_norm": 0.5830681920051575,
      "learning_rate": 0.0005263204396078008,
      "loss": 3.7342,
      "step": 11500
    },
    {
      "epoch": 1.2431385211494996,
      "grad_norm": 0.5570287704467773,
      "learning_rate": 0.0005259971985777394,
      "loss": 3.7496,
      "step": 11550
    },
    {
      "epoch": 1.2485200731891077,
      "grad_norm": 0.5565011501312256,
      "learning_rate": 0.000525673957547678,
      "loss": 3.745,
      "step": 11600
    },
    {
      "epoch": 1.2539016252287158,
      "grad_norm": 0.5676853060722351,
      "learning_rate": 0.0005253507165176167,
      "loss": 3.7563,
      "step": 11650
    },
    {
      "epoch": 1.2592831772683242,
      "grad_norm": 0.5753624439239502,
      "learning_rate": 0.0005250274754875552,
      "loss": 3.7492,
      "step": 11700
    },
    {
      "epoch": 1.2646647293079325,
      "grad_norm": 0.6029525995254517,
      "learning_rate": 0.0005247042344574938,
      "loss": 3.745,
      "step": 11750
    },
    {
      "epoch": 1.2700462813475406,
      "grad_norm": 0.5230038166046143,
      "learning_rate": 0.0005243809934274323,
      "loss": 3.7632,
      "step": 11800
    },
    {
      "epoch": 1.275427833387149,
      "grad_norm": 0.5502301454544067,
      "learning_rate": 0.0005240577523973709,
      "loss": 3.7296,
      "step": 11850
    },
    {
      "epoch": 1.280809385426757,
      "grad_norm": 0.566849946975708,
      "learning_rate": 0.0005237345113673095,
      "loss": 3.7318,
      "step": 11900
    },
    {
      "epoch": 1.2861909374663654,
      "grad_norm": 0.5555686950683594,
      "learning_rate": 0.0005234112703372481,
      "loss": 3.7448,
      "step": 11950
    },
    {
      "epoch": 1.2915724895059735,
      "grad_norm": 0.520969033241272,
      "learning_rate": 0.0005230880293071867,
      "loss": 3.7433,
      "step": 12000
    },
    {
      "epoch": 1.2915724895059735,
      "eval_accuracy": 0.34817437615043034,
      "eval_loss": 3.7075250148773193,
      "eval_runtime": 185.027,
      "eval_samples_per_second": 97.343,
      "eval_steps_per_second": 6.086,
      "step": 12000
    },
    {
      "epoch": 1.2969540415455818,
      "grad_norm": 0.5476548075675964,
      "learning_rate": 0.0005227712530977265,
      "loss": 3.7293,
      "step": 12050
    },
    {
      "epoch": 1.30233559358519,
      "grad_norm": 0.6757012605667114,
      "learning_rate": 0.0005224480120676651,
      "loss": 3.7316,
      "step": 12100
    },
    {
      "epoch": 1.3077171456247982,
      "grad_norm": 0.7626763582229614,
      "learning_rate": 0.0005221247710376037,
      "loss": 3.7305,
      "step": 12150
    },
    {
      "epoch": 1.3130986976644063,
      "grad_norm": 0.5633838176727295,
      "learning_rate": 0.0005218015300075422,
      "loss": 3.7346,
      "step": 12200
    },
    {
      "epoch": 1.3184802497040147,
      "grad_norm": 0.573448121547699,
      "learning_rate": 0.0005214782889774809,
      "loss": 3.7292,
      "step": 12250
    },
    {
      "epoch": 1.3238618017436228,
      "grad_norm": 0.5759413242340088,
      "learning_rate": 0.0005211550479474194,
      "loss": 3.7435,
      "step": 12300
    },
    {
      "epoch": 1.329243353783231,
      "grad_norm": 0.531487762928009,
      "learning_rate": 0.000520831806917358,
      "loss": 3.7258,
      "step": 12350
    },
    {
      "epoch": 1.3346249058228392,
      "grad_norm": 0.6096369624137878,
      "learning_rate": 0.0005205085658872966,
      "loss": 3.7304,
      "step": 12400
    },
    {
      "epoch": 1.3400064578624475,
      "grad_norm": 0.5672060251235962,
      "learning_rate": 0.0005201853248572351,
      "loss": 3.7202,
      "step": 12450
    },
    {
      "epoch": 1.3453880099020559,
      "grad_norm": 0.5578292608261108,
      "learning_rate": 0.0005198620838271738,
      "loss": 3.7267,
      "step": 12500
    },
    {
      "epoch": 1.350769561941664,
      "grad_norm": 0.6003977060317993,
      "learning_rate": 0.0005195388427971123,
      "loss": 3.7263,
      "step": 12550
    },
    {
      "epoch": 1.356151113981272,
      "grad_norm": 0.5442166328430176,
      "learning_rate": 0.000519215601767051,
      "loss": 3.7171,
      "step": 12600
    },
    {
      "epoch": 1.3615326660208804,
      "grad_norm": 0.5496247410774231,
      "learning_rate": 0.0005188923607369895,
      "loss": 3.7428,
      "step": 12650
    },
    {
      "epoch": 1.3669142180604887,
      "grad_norm": 0.580812931060791,
      "learning_rate": 0.000518569119706928,
      "loss": 3.7242,
      "step": 12700
    },
    {
      "epoch": 1.3722957701000968,
      "grad_norm": 0.5713503956794739,
      "learning_rate": 0.0005182458786768667,
      "loss": 3.7265,
      "step": 12750
    },
    {
      "epoch": 1.3776773221397052,
      "grad_norm": 0.6114019155502319,
      "learning_rate": 0.0005179226376468052,
      "loss": 3.7249,
      "step": 12800
    },
    {
      "epoch": 1.3830588741793133,
      "grad_norm": 0.5706479549407959,
      "learning_rate": 0.0005175993966167438,
      "loss": 3.7194,
      "step": 12850
    },
    {
      "epoch": 1.3884404262189216,
      "grad_norm": 0.5517324209213257,
      "learning_rate": 0.0005172761555866824,
      "loss": 3.7084,
      "step": 12900
    },
    {
      "epoch": 1.3938219782585297,
      "grad_norm": 0.6003121137619019,
      "learning_rate": 0.0005169529145566211,
      "loss": 3.7325,
      "step": 12950
    },
    {
      "epoch": 1.399203530298138,
      "grad_norm": 0.6365465521812439,
      "learning_rate": 0.0005166296735265596,
      "loss": 3.6954,
      "step": 13000
    },
    {
      "epoch": 1.399203530298138,
      "eval_accuracy": 0.3507263065265291,
      "eval_loss": 3.6849923133850098,
      "eval_runtime": 185.465,
      "eval_samples_per_second": 97.113,
      "eval_steps_per_second": 6.071,
      "step": 13000
    },
    {
      "epoch": 1.4045850823377461,
      "grad_norm": 0.6039630770683289,
      "learning_rate": 0.0005163064324964982,
      "loss": 3.725,
      "step": 13050
    },
    {
      "epoch": 1.4099666343773545,
      "grad_norm": 0.5968127846717834,
      "learning_rate": 0.0005159831914664367,
      "loss": 3.7248,
      "step": 13100
    },
    {
      "epoch": 1.4153481864169626,
      "grad_norm": 0.5731820464134216,
      "learning_rate": 0.0005156599504363753,
      "loss": 3.7439,
      "step": 13150
    },
    {
      "epoch": 1.420729738456571,
      "grad_norm": 0.564530074596405,
      "learning_rate": 0.000515336709406314,
      "loss": 3.718,
      "step": 13200
    },
    {
      "epoch": 1.426111290496179,
      "grad_norm": 0.5262428522109985,
      "learning_rate": 0.0005150134683762525,
      "loss": 3.7107,
      "step": 13250
    },
    {
      "epoch": 1.4314928425357873,
      "grad_norm": 0.5388005971908569,
      "learning_rate": 0.0005146902273461911,
      "loss": 3.7054,
      "step": 13300
    },
    {
      "epoch": 1.4368743945753955,
      "grad_norm": 0.5280932784080505,
      "learning_rate": 0.0005143669863161297,
      "loss": 3.7194,
      "step": 13350
    },
    {
      "epoch": 1.4422559466150038,
      "grad_norm": 0.5483044385910034,
      "learning_rate": 0.0005140437452860683,
      "loss": 3.7228,
      "step": 13400
    },
    {
      "epoch": 1.447637498654612,
      "grad_norm": 0.5313496589660645,
      "learning_rate": 0.0005137205042560069,
      "loss": 3.7077,
      "step": 13450
    },
    {
      "epoch": 1.4530190506942202,
      "grad_norm": 0.5253715515136719,
      "learning_rate": 0.0005133972632259455,
      "loss": 3.7186,
      "step": 13500
    },
    {
      "epoch": 1.4584006027338283,
      "grad_norm": 0.5300552845001221,
      "learning_rate": 0.000513074022195884,
      "loss": 3.7023,
      "step": 13550
    },
    {
      "epoch": 1.4637821547734367,
      "grad_norm": 0.5582554340362549,
      "learning_rate": 0.0005127507811658226,
      "loss": 3.6958,
      "step": 13600
    },
    {
      "epoch": 1.469163706813045,
      "grad_norm": 0.6429547667503357,
      "learning_rate": 0.0005124275401357612,
      "loss": 3.7128,
      "step": 13650
    },
    {
      "epoch": 1.474545258852653,
      "grad_norm": 0.5178701281547546,
      "learning_rate": 0.0005121042991056997,
      "loss": 3.6942,
      "step": 13700
    },
    {
      "epoch": 1.4799268108922612,
      "grad_norm": 0.5510112047195435,
      "learning_rate": 0.0005117810580756384,
      "loss": 3.7081,
      "step": 13750
    },
    {
      "epoch": 1.4853083629318695,
      "grad_norm": 0.5764445662498474,
      "learning_rate": 0.0005114578170455769,
      "loss": 3.7178,
      "step": 13800
    },
    {
      "epoch": 1.4906899149714778,
      "grad_norm": 0.5534348487854004,
      "learning_rate": 0.0005111345760155156,
      "loss": 3.706,
      "step": 13850
    },
    {
      "epoch": 1.496071467011086,
      "grad_norm": 0.5160101056098938,
      "learning_rate": 0.0005108113349854541,
      "loss": 3.7143,
      "step": 13900
    },
    {
      "epoch": 1.501453019050694,
      "grad_norm": 0.5557759404182434,
      "learning_rate": 0.0005104880939553926,
      "loss": 3.7083,
      "step": 13950
    },
    {
      "epoch": 1.5068345710903024,
      "grad_norm": 0.5790023803710938,
      "learning_rate": 0.0005101648529253313,
      "loss": 3.7046,
      "step": 14000
    },
    {
      "epoch": 1.5068345710903024,
      "eval_accuracy": 0.3523945629660326,
      "eval_loss": 3.6652114391326904,
      "eval_runtime": 185.3912,
      "eval_samples_per_second": 97.151,
      "eval_steps_per_second": 6.074,
      "step": 14000
    },
    {
      "epoch": 1.5122161231299107,
      "grad_norm": 0.6131024956703186,
      "learning_rate": 0.0005098480767158711,
      "loss": 3.7117,
      "step": 14050
    },
    {
      "epoch": 1.5175976751695188,
      "grad_norm": 0.5883623361587524,
      "learning_rate": 0.0005095248356858097,
      "loss": 3.704,
      "step": 14100
    },
    {
      "epoch": 1.5229792272091272,
      "grad_norm": 0.6037416458129883,
      "learning_rate": 0.0005092015946557483,
      "loss": 3.6967,
      "step": 14150
    },
    {
      "epoch": 1.5283607792487355,
      "grad_norm": 0.5779008865356445,
      "learning_rate": 0.0005088783536256868,
      "loss": 3.7008,
      "step": 14200
    },
    {
      "epoch": 1.5337423312883436,
      "grad_norm": 0.5496140122413635,
      "learning_rate": 0.0005085551125956255,
      "loss": 3.6972,
      "step": 14250
    },
    {
      "epoch": 1.5391238833279517,
      "grad_norm": 0.5408839583396912,
      "learning_rate": 0.000508231871565564,
      "loss": 3.6907,
      "step": 14300
    },
    {
      "epoch": 1.54450543536756,
      "grad_norm": 0.49644845724105835,
      "learning_rate": 0.0005079086305355026,
      "loss": 3.7073,
      "step": 14350
    },
    {
      "epoch": 1.5498869874071683,
      "grad_norm": 0.5873992443084717,
      "learning_rate": 0.0005075853895054412,
      "loss": 3.7142,
      "step": 14400
    },
    {
      "epoch": 1.5552685394467765,
      "grad_norm": 0.6147649884223938,
      "learning_rate": 0.0005072621484753797,
      "loss": 3.7073,
      "step": 14450
    },
    {
      "epoch": 1.5606500914863846,
      "grad_norm": 0.5985692143440247,
      "learning_rate": 0.0005069389074453184,
      "loss": 3.6826,
      "step": 14500
    },
    {
      "epoch": 1.566031643525993,
      "grad_norm": 0.5603007674217224,
      "learning_rate": 0.0005066156664152569,
      "loss": 3.6996,
      "step": 14550
    },
    {
      "epoch": 1.5714131955656012,
      "grad_norm": 0.5217431783676147,
      "learning_rate": 0.0005062924253851955,
      "loss": 3.7008,
      "step": 14600
    },
    {
      "epoch": 1.5767947476052093,
      "grad_norm": 0.5642618536949158,
      "learning_rate": 0.0005059691843551341,
      "loss": 3.6809,
      "step": 14650
    },
    {
      "epoch": 1.5821762996448174,
      "grad_norm": 0.5684123039245605,
      "learning_rate": 0.0005056459433250727,
      "loss": 3.697,
      "step": 14700
    },
    {
      "epoch": 1.5875578516844258,
      "grad_norm": 0.5687706470489502,
      "learning_rate": 0.0005053227022950113,
      "loss": 3.6864,
      "step": 14750
    },
    {
      "epoch": 1.592939403724034,
      "grad_norm": 0.5460999608039856,
      "learning_rate": 0.0005049994612649499,
      "loss": 3.7047,
      "step": 14800
    },
    {
      "epoch": 1.5983209557636422,
      "grad_norm": 0.6303175687789917,
      "learning_rate": 0.0005046762202348884,
      "loss": 3.6926,
      "step": 14850
    },
    {
      "epoch": 1.6037025078032503,
      "grad_norm": 0.590775728225708,
      "learning_rate": 0.000504352979204827,
      "loss": 3.6951,
      "step": 14900
    },
    {
      "epoch": 1.6090840598428586,
      "grad_norm": 0.5631206631660461,
      "learning_rate": 0.0005040297381747656,
      "loss": 3.6979,
      "step": 14950
    },
    {
      "epoch": 1.614465611882467,
      "grad_norm": 0.6098004579544067,
      "learning_rate": 0.0005037064971447042,
      "loss": 3.6836,
      "step": 15000
    },
    {
      "epoch": 1.614465611882467,
      "eval_accuracy": 0.35486793730424143,
      "eval_loss": 3.644303798675537,
      "eval_runtime": 184.9336,
      "eval_samples_per_second": 97.392,
      "eval_steps_per_second": 6.089,
      "step": 15000
    },
    {
      "epoch": 1.619847163922075,
      "grad_norm": 0.6067646741867065,
      "learning_rate": 0.0005033832561146428,
      "loss": 3.6931,
      "step": 15050
    },
    {
      "epoch": 1.6252287159616834,
      "grad_norm": 0.5822307467460632,
      "learning_rate": 0.0005030600150845813,
      "loss": 3.68,
      "step": 15100
    },
    {
      "epoch": 1.6306102680012917,
      "grad_norm": 0.5807347297668457,
      "learning_rate": 0.00050273677405452,
      "loss": 3.6859,
      "step": 15150
    },
    {
      "epoch": 1.6359918200408998,
      "grad_norm": 0.5637763142585754,
      "learning_rate": 0.0005024135330244585,
      "loss": 3.6753,
      "step": 15200
    },
    {
      "epoch": 1.641373372080508,
      "grad_norm": 0.5217981338500977,
      "learning_rate": 0.0005020902919943972,
      "loss": 3.6977,
      "step": 15250
    },
    {
      "epoch": 1.6467549241201163,
      "grad_norm": 0.6031511425971985,
      "learning_rate": 0.0005017670509643357,
      "loss": 3.6661,
      "step": 15300
    },
    {
      "epoch": 1.6521364761597246,
      "grad_norm": 0.531609296798706,
      "learning_rate": 0.0005014438099342743,
      "loss": 3.6887,
      "step": 15350
    },
    {
      "epoch": 1.6575180281993327,
      "grad_norm": 0.5228476524353027,
      "learning_rate": 0.0005011205689042129,
      "loss": 3.6822,
      "step": 15400
    },
    {
      "epoch": 1.6628995802389408,
      "grad_norm": 0.5995671153068542,
      "learning_rate": 0.0005007973278741514,
      "loss": 3.7085,
      "step": 15450
    },
    {
      "epoch": 1.6682811322785491,
      "grad_norm": 0.6415295004844666,
      "learning_rate": 0.00050047408684409,
      "loss": 3.7013,
      "step": 15500
    },
    {
      "epoch": 1.6736626843181575,
      "grad_norm": 0.585547685623169,
      "learning_rate": 0.0005001508458140286,
      "loss": 3.6843,
      "step": 15550
    },
    {
      "epoch": 1.6790442363577656,
      "grad_norm": 0.5341925024986267,
      "learning_rate": 0.0004998276047839673,
      "loss": 3.7126,
      "step": 15600
    },
    {
      "epoch": 1.6844257883973737,
      "grad_norm": 0.6166404485702515,
      "learning_rate": 0.0004995043637539058,
      "loss": 3.6815,
      "step": 15650
    },
    {
      "epoch": 1.689807340436982,
      "grad_norm": 0.5578957200050354,
      "learning_rate": 0.0004991811227238443,
      "loss": 3.6761,
      "step": 15700
    },
    {
      "epoch": 1.6951888924765903,
      "grad_norm": 0.6500715613365173,
      "learning_rate": 0.0004988578816937829,
      "loss": 3.6974,
      "step": 15750
    },
    {
      "epoch": 1.7005704445161984,
      "grad_norm": 0.5625742077827454,
      "learning_rate": 0.0004985346406637215,
      "loss": 3.6699,
      "step": 15800
    },
    {
      "epoch": 1.7059519965558065,
      "grad_norm": 0.5425594449043274,
      "learning_rate": 0.0004982113996336602,
      "loss": 3.6727,
      "step": 15850
    },
    {
      "epoch": 1.7113335485954149,
      "grad_norm": 0.5652900338172913,
      "learning_rate": 0.0004978881586035987,
      "loss": 3.6937,
      "step": 15900
    },
    {
      "epoch": 1.7167151006350232,
      "grad_norm": 0.5479394793510437,
      "learning_rate": 0.0004975649175735373,
      "loss": 3.6701,
      "step": 15950
    },
    {
      "epoch": 1.7220966526746313,
      "grad_norm": 0.5041327476501465,
      "learning_rate": 0.0004972416765434759,
      "loss": 3.6705,
      "step": 16000
    },
    {
      "epoch": 1.7220966526746313,
      "eval_accuracy": 0.35625467411141765,
      "eval_loss": 3.6317310333251953,
      "eval_runtime": 185.1436,
      "eval_samples_per_second": 97.281,
      "eval_steps_per_second": 6.082,
      "step": 16000
    },
    {
      "epoch": 1.7274782047142396,
      "grad_norm": 0.529309093952179,
      "learning_rate": 0.0004969184355134145,
      "loss": 3.6741,
      "step": 16050
    },
    {
      "epoch": 1.732859756753848,
      "grad_norm": 0.6252491474151611,
      "learning_rate": 0.0004965951944833531,
      "loss": 3.6739,
      "step": 16100
    },
    {
      "epoch": 1.738241308793456,
      "grad_norm": 0.5377236604690552,
      "learning_rate": 0.0004962784182738928,
      "loss": 3.6919,
      "step": 16150
    },
    {
      "epoch": 1.7436228608330642,
      "grad_norm": 0.6052237153053284,
      "learning_rate": 0.0004959551772438314,
      "loss": 3.654,
      "step": 16200
    },
    {
      "epoch": 1.7490044128726725,
      "grad_norm": 0.5392761826515198,
      "learning_rate": 0.0004956319362137701,
      "loss": 3.6764,
      "step": 16250
    },
    {
      "epoch": 1.7543859649122808,
      "grad_norm": 0.5225568413734436,
      "learning_rate": 0.0004953086951837086,
      "loss": 3.6883,
      "step": 16300
    },
    {
      "epoch": 1.759767516951889,
      "grad_norm": 0.5589337348937988,
      "learning_rate": 0.0004949854541536472,
      "loss": 3.6697,
      "step": 16350
    },
    {
      "epoch": 1.765149068991497,
      "grad_norm": 0.5792645812034607,
      "learning_rate": 0.0004946622131235857,
      "loss": 3.686,
      "step": 16400
    },
    {
      "epoch": 1.7705306210311054,
      "grad_norm": 0.5120959281921387,
      "learning_rate": 0.0004943389720935244,
      "loss": 3.6547,
      "step": 16450
    },
    {
      "epoch": 1.7759121730707137,
      "grad_norm": 0.5924346446990967,
      "learning_rate": 0.000494015731063463,
      "loss": 3.6875,
      "step": 16500
    },
    {
      "epoch": 1.7812937251103218,
      "grad_norm": 0.5631136894226074,
      "learning_rate": 0.0004936924900334016,
      "loss": 3.6582,
      "step": 16550
    },
    {
      "epoch": 1.78667527714993,
      "grad_norm": 0.5707646012306213,
      "learning_rate": 0.0004933692490033401,
      "loss": 3.6777,
      "step": 16600
    },
    {
      "epoch": 1.7920568291895382,
      "grad_norm": 0.6018843650817871,
      "learning_rate": 0.0004930460079732786,
      "loss": 3.6792,
      "step": 16650
    },
    {
      "epoch": 1.7974383812291466,
      "grad_norm": 0.5362632870674133,
      "learning_rate": 0.0004927227669432173,
      "loss": 3.6595,
      "step": 16700
    },
    {
      "epoch": 1.8028199332687547,
      "grad_norm": 0.5789631009101868,
      "learning_rate": 0.0004923995259131558,
      "loss": 3.6609,
      "step": 16750
    },
    {
      "epoch": 1.8082014853083628,
      "grad_norm": 0.6050270199775696,
      "learning_rate": 0.0004920762848830945,
      "loss": 3.6883,
      "step": 16800
    },
    {
      "epoch": 1.813583037347971,
      "grad_norm": 0.5367583632469177,
      "learning_rate": 0.000491753043853033,
      "loss": 3.6694,
      "step": 16850
    },
    {
      "epoch": 1.8189645893875794,
      "grad_norm": 0.5649946331977844,
      "learning_rate": 0.0004914298028229717,
      "loss": 3.6738,
      "step": 16900
    },
    {
      "epoch": 1.8243461414271875,
      "grad_norm": 0.6238928437232971,
      "learning_rate": 0.0004911065617929102,
      "loss": 3.6669,
      "step": 16950
    },
    {
      "epoch": 1.8297276934667959,
      "grad_norm": 0.5264888405799866,
      "learning_rate": 0.0004907833207628487,
      "loss": 3.6766,
      "step": 17000
    },
    {
      "epoch": 1.8297276934667959,
      "eval_accuracy": 0.35823900180157353,
      "eval_loss": 3.6119563579559326,
      "eval_runtime": 185.1336,
      "eval_samples_per_second": 97.287,
      "eval_steps_per_second": 6.082,
      "step": 17000
    },
    {
      "epoch": 1.8351092455064042,
      "grad_norm": 0.6952211856842041,
      "learning_rate": 0.0004904600797327874,
      "loss": 3.6672,
      "step": 17050
    },
    {
      "epoch": 1.8404907975460123,
      "grad_norm": 0.5280624628067017,
      "learning_rate": 0.0004901368387027259,
      "loss": 3.6396,
      "step": 17100
    },
    {
      "epoch": 1.8458723495856204,
      "grad_norm": 0.6064589023590088,
      "learning_rate": 0.0004898135976726646,
      "loss": 3.6575,
      "step": 17150
    },
    {
      "epoch": 1.8512539016252287,
      "grad_norm": 0.5512820482254028,
      "learning_rate": 0.0004894903566426031,
      "loss": 3.6697,
      "step": 17200
    },
    {
      "epoch": 1.856635453664837,
      "grad_norm": 0.5815790295600891,
      "learning_rate": 0.0004891671156125417,
      "loss": 3.6543,
      "step": 17250
    },
    {
      "epoch": 1.8620170057044452,
      "grad_norm": 0.5231164693832397,
      "learning_rate": 0.0004888438745824803,
      "loss": 3.6726,
      "step": 17300
    },
    {
      "epoch": 1.8673985577440533,
      "grad_norm": 0.6134333610534668,
      "learning_rate": 0.0004885206335524189,
      "loss": 3.6545,
      "step": 17350
    },
    {
      "epoch": 1.8727801097836616,
      "grad_norm": 0.5191701650619507,
      "learning_rate": 0.0004881973925223575,
      "loss": 3.6635,
      "step": 17400
    },
    {
      "epoch": 1.87816166182327,
      "grad_norm": 0.5176146030426025,
      "learning_rate": 0.00048787415149229604,
      "loss": 3.6388,
      "step": 17450
    },
    {
      "epoch": 1.883543213862878,
      "grad_norm": 0.5695205330848694,
      "learning_rate": 0.00048755091046223464,
      "loss": 3.6604,
      "step": 17500
    },
    {
      "epoch": 1.8889247659024861,
      "grad_norm": 0.549403965473175,
      "learning_rate": 0.0004872276694321732,
      "loss": 3.6554,
      "step": 17550
    },
    {
      "epoch": 1.8943063179420945,
      "grad_norm": 0.5490947365760803,
      "learning_rate": 0.00048690442840211177,
      "loss": 3.6469,
      "step": 17600
    },
    {
      "epoch": 1.8996878699817028,
      "grad_norm": 0.6216672658920288,
      "learning_rate": 0.0004865811873720504,
      "loss": 3.6561,
      "step": 17650
    },
    {
      "epoch": 1.905069422021311,
      "grad_norm": 0.5394617915153503,
      "learning_rate": 0.00048625794634198896,
      "loss": 3.6456,
      "step": 17700
    },
    {
      "epoch": 1.910450974060919,
      "grad_norm": 0.5857135057449341,
      "learning_rate": 0.00048593470531192756,
      "loss": 3.6539,
      "step": 17750
    },
    {
      "epoch": 1.9158325261005273,
      "grad_norm": 0.5659494400024414,
      "learning_rate": 0.00048561146428186615,
      "loss": 3.6537,
      "step": 17800
    },
    {
      "epoch": 1.9212140781401357,
      "grad_norm": 0.5693187713623047,
      "learning_rate": 0.0004852882232518047,
      "loss": 3.6667,
      "step": 17850
    },
    {
      "epoch": 1.9265956301797438,
      "grad_norm": 0.4942563474178314,
      "learning_rate": 0.0004849714470423445,
      "loss": 3.6366,
      "step": 17900
    },
    {
      "epoch": 1.931977182219352,
      "grad_norm": 0.5239983201026917,
      "learning_rate": 0.0004846482060122831,
      "loss": 3.6539,
      "step": 17950
    },
    {
      "epoch": 1.9373587342589604,
      "grad_norm": 0.5823050141334534,
      "learning_rate": 0.00048432496498222174,
      "loss": 3.6642,
      "step": 18000
    },
    {
      "epoch": 1.9373587342589604,
      "eval_accuracy": 0.3593307460183334,
      "eval_loss": 3.5988430976867676,
      "eval_runtime": 185.4555,
      "eval_samples_per_second": 97.118,
      "eval_steps_per_second": 6.072,
      "step": 18000
    },
    {
      "epoch": 1.9427402862985685,
      "grad_norm": 0.546956479549408,
      "learning_rate": 0.0004840017239521603,
      "loss": 3.6703,
      "step": 18050
    },
    {
      "epoch": 1.9481218383381766,
      "grad_norm": 0.5563112497329712,
      "learning_rate": 0.0004836784829220989,
      "loss": 3.6315,
      "step": 18100
    },
    {
      "epoch": 1.953503390377785,
      "grad_norm": 0.5369589924812317,
      "learning_rate": 0.00048335524189203747,
      "loss": 3.6618,
      "step": 18150
    },
    {
      "epoch": 1.9588849424173933,
      "grad_norm": 0.5867165923118591,
      "learning_rate": 0.000483032000861976,
      "loss": 3.6482,
      "step": 18200
    },
    {
      "epoch": 1.9642664944570014,
      "grad_norm": 0.5733823776245117,
      "learning_rate": 0.0004827087598319146,
      "loss": 3.6353,
      "step": 18250
    },
    {
      "epoch": 1.9696480464966095,
      "grad_norm": 0.5795925855636597,
      "learning_rate": 0.00048238551880185325,
      "loss": 3.636,
      "step": 18300
    },
    {
      "epoch": 1.9750295985362178,
      "grad_norm": 0.5763929486274719,
      "learning_rate": 0.0004820622777717918,
      "loss": 3.641,
      "step": 18350
    },
    {
      "epoch": 1.9804111505758262,
      "grad_norm": 0.5630631446838379,
      "learning_rate": 0.0004817390367417304,
      "loss": 3.635,
      "step": 18400
    },
    {
      "epoch": 1.9857927026154343,
      "grad_norm": 0.5994884371757507,
      "learning_rate": 0.00048141579571166893,
      "loss": 3.6258,
      "step": 18450
    },
    {
      "epoch": 1.9911742546550424,
      "grad_norm": 0.5328142642974854,
      "learning_rate": 0.0004810925546816075,
      "loss": 3.6289,
      "step": 18500
    },
    {
      "epoch": 1.9965558066946507,
      "grad_norm": 0.6015245914459229,
      "learning_rate": 0.0004807693136515461,
      "loss": 3.6417,
      "step": 18550
    },
    {
      "epoch": 2.001937358734259,
      "grad_norm": 0.5309726595878601,
      "learning_rate": 0.0004804460726214847,
      "loss": 3.6061,
      "step": 18600
    },
    {
      "epoch": 2.007318910773867,
      "grad_norm": 0.53099125623703,
      "learning_rate": 0.0004801228315914233,
      "loss": 3.5522,
      "step": 18650
    },
    {
      "epoch": 2.0127004628134753,
      "grad_norm": 0.5386189818382263,
      "learning_rate": 0.0004797995905613619,
      "loss": 3.5501,
      "step": 18700
    },
    {
      "epoch": 2.018082014853084,
      "grad_norm": 0.5308017134666443,
      "learning_rate": 0.00047947634953130044,
      "loss": 3.5526,
      "step": 18750
    },
    {
      "epoch": 2.023463566892692,
      "grad_norm": 0.6009208559989929,
      "learning_rate": 0.00047915310850123904,
      "loss": 3.5499,
      "step": 18800
    },
    {
      "epoch": 2.0288451189323,
      "grad_norm": 0.6113578081130981,
      "learning_rate": 0.0004788298674711777,
      "loss": 3.5767,
      "step": 18850
    },
    {
      "epoch": 2.034226670971908,
      "grad_norm": 0.6289225220680237,
      "learning_rate": 0.00047850662644111623,
      "loss": 3.554,
      "step": 18900
    },
    {
      "epoch": 2.0396082230115167,
      "grad_norm": 0.6042494773864746,
      "learning_rate": 0.0004781833854110548,
      "loss": 3.5598,
      "step": 18950
    },
    {
      "epoch": 2.044989775051125,
      "grad_norm": 0.5708305835723877,
      "learning_rate": 0.00047786014438099336,
      "loss": 3.574,
      "step": 19000
    },
    {
      "epoch": 2.044989775051125,
      "eval_accuracy": 0.36072660966808606,
      "eval_loss": 3.586899757385254,
      "eval_runtime": 184.8382,
      "eval_samples_per_second": 97.442,
      "eval_steps_per_second": 6.092,
      "step": 19000
    },
    {
      "epoch": 2.050371327090733,
      "grad_norm": 0.5245704650878906,
      "learning_rate": 0.00047753690335093196,
      "loss": 3.5606,
      "step": 19050
    },
    {
      "epoch": 2.055752879130341,
      "grad_norm": 0.5576515197753906,
      "learning_rate": 0.00047721366232087055,
      "loss": 3.5491,
      "step": 19100
    },
    {
      "epoch": 2.0611344311699495,
      "grad_norm": 0.618057370185852,
      "learning_rate": 0.00047689042129080915,
      "loss": 3.5587,
      "step": 19150
    },
    {
      "epoch": 2.0665159832095576,
      "grad_norm": 0.5552718639373779,
      "learning_rate": 0.00047656718026074774,
      "loss": 3.5688,
      "step": 19200
    },
    {
      "epoch": 2.0718975352491658,
      "grad_norm": 0.5995255708694458,
      "learning_rate": 0.00047624393923068634,
      "loss": 3.558,
      "step": 19250
    },
    {
      "epoch": 2.0772790872887743,
      "grad_norm": 0.5941905379295349,
      "learning_rate": 0.0004759206982006249,
      "loss": 3.559,
      "step": 19300
    },
    {
      "epoch": 2.0826606393283824,
      "grad_norm": 0.6038264632225037,
      "learning_rate": 0.00047559745717056347,
      "loss": 3.5792,
      "step": 19350
    },
    {
      "epoch": 2.0880421913679905,
      "grad_norm": 0.5839534997940063,
      "learning_rate": 0.000475274216140502,
      "loss": 3.5724,
      "step": 19400
    },
    {
      "epoch": 2.0934237434075986,
      "grad_norm": 0.5897072553634644,
      "learning_rate": 0.00047495097511044066,
      "loss": 3.5659,
      "step": 19450
    },
    {
      "epoch": 2.098805295447207,
      "grad_norm": 0.5537237524986267,
      "learning_rate": 0.00047462773408037925,
      "loss": 3.5767,
      "step": 19500
    },
    {
      "epoch": 2.1041868474868153,
      "grad_norm": 0.5752836465835571,
      "learning_rate": 0.0004743044930503178,
      "loss": 3.5724,
      "step": 19550
    },
    {
      "epoch": 2.1095683995264234,
      "grad_norm": 0.5968397855758667,
      "learning_rate": 0.0004739812520202564,
      "loss": 3.5718,
      "step": 19600
    },
    {
      "epoch": 2.1149499515660315,
      "grad_norm": 0.5478861331939697,
      "learning_rate": 0.000473658010990195,
      "loss": 3.5814,
      "step": 19650
    },
    {
      "epoch": 2.12033150360564,
      "grad_norm": 0.5469426512718201,
      "learning_rate": 0.0004733347699601336,
      "loss": 3.5752,
      "step": 19700
    },
    {
      "epoch": 2.125713055645248,
      "grad_norm": 0.5446122884750366,
      "learning_rate": 0.0004730115289300722,
      "loss": 3.575,
      "step": 19750
    },
    {
      "epoch": 2.1310946076848563,
      "grad_norm": 0.6697937846183777,
      "learning_rate": 0.00047268828790001077,
      "loss": 3.5694,
      "step": 19800
    },
    {
      "epoch": 2.1364761597244644,
      "grad_norm": 0.5983029007911682,
      "learning_rate": 0.0004723650468699493,
      "loss": 3.5792,
      "step": 19850
    },
    {
      "epoch": 2.141857711764073,
      "grad_norm": 0.5419188737869263,
      "learning_rate": 0.0004720418058398879,
      "loss": 3.5583,
      "step": 19900
    },
    {
      "epoch": 2.147239263803681,
      "grad_norm": 0.5491231679916382,
      "learning_rate": 0.00047171856480982644,
      "loss": 3.5681,
      "step": 19950
    },
    {
      "epoch": 2.152620815843289,
      "grad_norm": 0.5777820944786072,
      "learning_rate": 0.0004713953237797651,
      "loss": 3.586,
      "step": 20000
    },
    {
      "epoch": 2.152620815843289,
      "eval_accuracy": 0.36172991043416497,
      "eval_loss": 3.579937696456909,
      "eval_runtime": 184.9995,
      "eval_samples_per_second": 97.357,
      "eval_steps_per_second": 6.087,
      "step": 20000
    },
    {
      "epoch": 2.1580023678828972,
      "grad_norm": 0.5954360365867615,
      "learning_rate": 0.0004710720827497037,
      "loss": 3.5572,
      "step": 20050
    },
    {
      "epoch": 2.163383919922506,
      "grad_norm": 0.5662351846694946,
      "learning_rate": 0.00047074884171964223,
      "loss": 3.589,
      "step": 20100
    },
    {
      "epoch": 2.168765471962114,
      "grad_norm": 0.5887290239334106,
      "learning_rate": 0.0004704256006895808,
      "loss": 3.5802,
      "step": 20150
    },
    {
      "epoch": 2.174147024001722,
      "grad_norm": 0.5584112405776978,
      "learning_rate": 0.00047010235965951936,
      "loss": 3.5675,
      "step": 20200
    },
    {
      "epoch": 2.1795285760413305,
      "grad_norm": 0.6759079694747925,
      "learning_rate": 0.00046977911862945796,
      "loss": 3.5579,
      "step": 20250
    },
    {
      "epoch": 2.1849101280809387,
      "grad_norm": 0.5682075619697571,
      "learning_rate": 0.0004694558775993966,
      "loss": 3.5598,
      "step": 20300
    },
    {
      "epoch": 2.1902916801205468,
      "grad_norm": 0.5859228372573853,
      "learning_rate": 0.00046913263656933515,
      "loss": 3.5695,
      "step": 20350
    },
    {
      "epoch": 2.195673232160155,
      "grad_norm": 0.5964463353157043,
      "learning_rate": 0.00046880939553927374,
      "loss": 3.5668,
      "step": 20400
    },
    {
      "epoch": 2.2010547841997634,
      "grad_norm": 0.6303418278694153,
      "learning_rate": 0.00046848615450921234,
      "loss": 3.5694,
      "step": 20450
    },
    {
      "epoch": 2.2064363362393715,
      "grad_norm": 0.584977924823761,
      "learning_rate": 0.0004681629134791509,
      "loss": 3.5611,
      "step": 20500
    },
    {
      "epoch": 2.2118178882789796,
      "grad_norm": 0.5613332986831665,
      "learning_rate": 0.00046783967244908947,
      "loss": 3.5823,
      "step": 20550
    },
    {
      "epoch": 2.2171994403185877,
      "grad_norm": 0.6148346066474915,
      "learning_rate": 0.0004675164314190281,
      "loss": 3.5691,
      "step": 20600
    },
    {
      "epoch": 2.2225809923581963,
      "grad_norm": 0.6257859468460083,
      "learning_rate": 0.00046719319038896666,
      "loss": 3.5705,
      "step": 20650
    },
    {
      "epoch": 2.2279625443978044,
      "grad_norm": 0.5708158016204834,
      "learning_rate": 0.00046686994935890526,
      "loss": 3.5666,
      "step": 20700
    },
    {
      "epoch": 2.2333440964374125,
      "grad_norm": 0.5658592581748962,
      "learning_rate": 0.0004665467083288438,
      "loss": 3.5436,
      "step": 20750
    },
    {
      "epoch": 2.2387256484770206,
      "grad_norm": 0.6272239685058594,
      "learning_rate": 0.0004662234672987824,
      "loss": 3.5503,
      "step": 20800
    },
    {
      "epoch": 2.244107200516629,
      "grad_norm": 0.5156669616699219,
      "learning_rate": 0.00046590022626872104,
      "loss": 3.5713,
      "step": 20850
    },
    {
      "epoch": 2.2494887525562373,
      "grad_norm": 0.5876966118812561,
      "learning_rate": 0.0004655769852386596,
      "loss": 3.5639,
      "step": 20900
    },
    {
      "epoch": 2.2548703045958454,
      "grad_norm": 0.6242596507072449,
      "learning_rate": 0.0004652537442085982,
      "loss": 3.5719,
      "step": 20950
    },
    {
      "epoch": 2.2602518566354535,
      "grad_norm": 0.6056633591651917,
      "learning_rate": 0.00046493050317853677,
      "loss": 3.579,
      "step": 21000
    },
    {
      "epoch": 2.2602518566354535,
      "eval_accuracy": 0.3628067692052943,
      "eval_loss": 3.5686612129211426,
      "eval_runtime": 185.1513,
      "eval_samples_per_second": 97.277,
      "eval_steps_per_second": 6.082,
      "step": 21000
    },
    {
      "epoch": 2.265633408675062,
      "grad_norm": 0.5574327707290649,
      "learning_rate": 0.0004646072621484753,
      "loss": 3.5771,
      "step": 21050
    },
    {
      "epoch": 2.27101496071467,
      "grad_norm": 0.5605047941207886,
      "learning_rate": 0.0004642840211184139,
      "loss": 3.5649,
      "step": 21100
    },
    {
      "epoch": 2.2763965127542782,
      "grad_norm": 0.5680267810821533,
      "learning_rate": 0.00046396078008835255,
      "loss": 3.5607,
      "step": 21150
    },
    {
      "epoch": 2.281778064793887,
      "grad_norm": 0.5700744390487671,
      "learning_rate": 0.0004636375390582911,
      "loss": 3.5688,
      "step": 21200
    },
    {
      "epoch": 2.287159616833495,
      "grad_norm": 0.5814450979232788,
      "learning_rate": 0.0004633142980282297,
      "loss": 3.5667,
      "step": 21250
    },
    {
      "epoch": 2.292541168873103,
      "grad_norm": 0.5372523069381714,
      "learning_rate": 0.00046299105699816823,
      "loss": 3.5721,
      "step": 21300
    },
    {
      "epoch": 2.297922720912711,
      "grad_norm": 0.5528806447982788,
      "learning_rate": 0.0004626678159681068,
      "loss": 3.5607,
      "step": 21350
    },
    {
      "epoch": 2.303304272952319,
      "grad_norm": 0.6022769808769226,
      "learning_rate": 0.0004623445749380454,
      "loss": 3.5699,
      "step": 21400
    },
    {
      "epoch": 2.3086858249919278,
      "grad_norm": 0.5523564219474792,
      "learning_rate": 0.000462021333907984,
      "loss": 3.5744,
      "step": 21450
    },
    {
      "epoch": 2.314067377031536,
      "grad_norm": 0.5891795754432678,
      "learning_rate": 0.0004616980928779226,
      "loss": 3.5686,
      "step": 21500
    },
    {
      "epoch": 2.319448929071144,
      "grad_norm": 0.5741997957229614,
      "learning_rate": 0.0004613748518478612,
      "loss": 3.5537,
      "step": 21550
    },
    {
      "epoch": 2.3248304811107525,
      "grad_norm": 0.5786455869674683,
      "learning_rate": 0.00046105161081779974,
      "loss": 3.5797,
      "step": 21600
    },
    {
      "epoch": 2.3302120331503606,
      "grad_norm": 0.5622592568397522,
      "learning_rate": 0.00046072836978773834,
      "loss": 3.5575,
      "step": 21650
    },
    {
      "epoch": 2.3355935851899687,
      "grad_norm": 0.6024913787841797,
      "learning_rate": 0.000460405128757677,
      "loss": 3.5572,
      "step": 21700
    },
    {
      "epoch": 2.340975137229577,
      "grad_norm": 0.584582507610321,
      "learning_rate": 0.0004600818877276155,
      "loss": 3.5753,
      "step": 21750
    },
    {
      "epoch": 2.3463566892691854,
      "grad_norm": 0.556788444519043,
      "learning_rate": 0.0004597586466975541,
      "loss": 3.5652,
      "step": 21800
    },
    {
      "epoch": 2.3517382413087935,
      "grad_norm": 0.5864278078079224,
      "learning_rate": 0.00045943540566749266,
      "loss": 3.5688,
      "step": 21850
    },
    {
      "epoch": 2.3571197933484016,
      "grad_norm": 0.587871253490448,
      "learning_rate": 0.00045911216463743126,
      "loss": 3.571,
      "step": 21900
    },
    {
      "epoch": 2.3625013453880097,
      "grad_norm": 0.5973302721977234,
      "learning_rate": 0.00045879538842797106,
      "loss": 3.5758,
      "step": 21950
    },
    {
      "epoch": 2.3678828974276183,
      "grad_norm": 0.5554845929145813,
      "learning_rate": 0.00045847214739790966,
      "loss": 3.5842,
      "step": 22000
    },
    {
      "epoch": 2.3678828974276183,
      "eval_accuracy": 0.3635833113945044,
      "eval_loss": 3.558429002761841,
      "eval_runtime": 185.3242,
      "eval_samples_per_second": 97.186,
      "eval_steps_per_second": 6.076,
      "step": 22000
    },
    {
      "epoch": 2.3732644494672264,
      "grad_norm": 0.5177135467529297,
      "learning_rate": 0.0004581489063678482,
      "loss": 3.5662,
      "step": 22050
    },
    {
      "epoch": 2.3786460015068345,
      "grad_norm": 0.667670726776123,
      "learning_rate": 0.00045782566533778685,
      "loss": 3.5808,
      "step": 22100
    },
    {
      "epoch": 2.384027553546443,
      "grad_norm": 0.5924534201622009,
      "learning_rate": 0.00045750242430772544,
      "loss": 3.568,
      "step": 22150
    },
    {
      "epoch": 2.389409105586051,
      "grad_norm": 0.5399534702301025,
      "learning_rate": 0.000457179183277664,
      "loss": 3.5782,
      "step": 22200
    },
    {
      "epoch": 2.3947906576256592,
      "grad_norm": 0.5430790781974792,
      "learning_rate": 0.0004568559422476026,
      "loss": 3.5533,
      "step": 22250
    },
    {
      "epoch": 2.4001722096652673,
      "grad_norm": 0.5829682946205139,
      "learning_rate": 0.00045653270121754117,
      "loss": 3.5716,
      "step": 22300
    },
    {
      "epoch": 2.4055537617048754,
      "grad_norm": 0.5541464686393738,
      "learning_rate": 0.0004562094601874797,
      "loss": 3.5646,
      "step": 22350
    },
    {
      "epoch": 2.410935313744484,
      "grad_norm": 0.5649983882904053,
      "learning_rate": 0.00045588621915741836,
      "loss": 3.5583,
      "step": 22400
    },
    {
      "epoch": 2.416316865784092,
      "grad_norm": 0.5958019495010376,
      "learning_rate": 0.00045556297812735696,
      "loss": 3.5506,
      "step": 22450
    },
    {
      "epoch": 2.4216984178237,
      "grad_norm": 0.5760444402694702,
      "learning_rate": 0.0004552397370972955,
      "loss": 3.5672,
      "step": 22500
    },
    {
      "epoch": 2.4270799698633088,
      "grad_norm": 0.5633614659309387,
      "learning_rate": 0.0004549164960672341,
      "loss": 3.5656,
      "step": 22550
    },
    {
      "epoch": 2.432461521902917,
      "grad_norm": 0.5957275629043579,
      "learning_rate": 0.00045459325503717263,
      "loss": 3.5573,
      "step": 22600
    },
    {
      "epoch": 2.437843073942525,
      "grad_norm": 0.6199848055839539,
      "learning_rate": 0.0004542700140071113,
      "loss": 3.5778,
      "step": 22650
    },
    {
      "epoch": 2.443224625982133,
      "grad_norm": 0.5850538611412048,
      "learning_rate": 0.0004539467729770499,
      "loss": 3.5652,
      "step": 22700
    },
    {
      "epoch": 2.4486061780217416,
      "grad_norm": 0.584876537322998,
      "learning_rate": 0.0004536235319469884,
      "loss": 3.5684,
      "step": 22750
    },
    {
      "epoch": 2.4539877300613497,
      "grad_norm": 0.5258706212043762,
      "learning_rate": 0.000453300290916927,
      "loss": 3.5463,
      "step": 22800
    },
    {
      "epoch": 2.459369282100958,
      "grad_norm": 0.5395996570587158,
      "learning_rate": 0.0004529770498868656,
      "loss": 3.5595,
      "step": 22850
    },
    {
      "epoch": 2.464750834140566,
      "grad_norm": 0.5797770619392395,
      "learning_rate": 0.00045265380885680414,
      "loss": 3.5744,
      "step": 22900
    },
    {
      "epoch": 2.4701323861801745,
      "grad_norm": 0.6019140481948853,
      "learning_rate": 0.0004523305678267428,
      "loss": 3.587,
      "step": 22950
    },
    {
      "epoch": 2.4755139382197826,
      "grad_norm": 0.5725117921829224,
      "learning_rate": 0.0004520073267966814,
      "loss": 3.5734,
      "step": 23000
    },
    {
      "epoch": 2.4755139382197826,
      "eval_accuracy": 0.3646976554119297,
      "eval_loss": 3.550741672515869,
      "eval_runtime": 185.1229,
      "eval_samples_per_second": 97.292,
      "eval_steps_per_second": 6.082,
      "step": 23000
    },
    {
      "epoch": 2.4808954902593907,
      "grad_norm": 0.5931921005249023,
      "learning_rate": 0.00045168408576661993,
      "loss": 3.5788,
      "step": 23050
    },
    {
      "epoch": 2.4862770422989993,
      "grad_norm": 0.5797882676124573,
      "learning_rate": 0.0004513608447365585,
      "loss": 3.535,
      "step": 23100
    },
    {
      "epoch": 2.4916585943386074,
      "grad_norm": 0.5638948082923889,
      "learning_rate": 0.00045103760370649706,
      "loss": 3.5692,
      "step": 23150
    },
    {
      "epoch": 2.4970401463782155,
      "grad_norm": 0.6029576063156128,
      "learning_rate": 0.00045071436267643566,
      "loss": 3.5509,
      "step": 23200
    },
    {
      "epoch": 2.5024216984178236,
      "grad_norm": 0.5442866086959839,
      "learning_rate": 0.0004503911216463743,
      "loss": 3.5577,
      "step": 23250
    },
    {
      "epoch": 2.5078032504574317,
      "grad_norm": 0.5774275660514832,
      "learning_rate": 0.00045006788061631285,
      "loss": 3.5548,
      "step": 23300
    },
    {
      "epoch": 2.5131848024970402,
      "grad_norm": 0.5645380616188049,
      "learning_rate": 0.00044974463958625144,
      "loss": 3.5557,
      "step": 23350
    },
    {
      "epoch": 2.5185663545366483,
      "grad_norm": 0.5402359962463379,
      "learning_rate": 0.00044942139855619004,
      "loss": 3.5655,
      "step": 23400
    },
    {
      "epoch": 2.5239479065762565,
      "grad_norm": 0.5606850385665894,
      "learning_rate": 0.0004490981575261286,
      "loss": 3.5516,
      "step": 23450
    },
    {
      "epoch": 2.529329458615865,
      "grad_norm": 0.5785205364227295,
      "learning_rate": 0.0004487749164960672,
      "loss": 3.5737,
      "step": 23500
    },
    {
      "epoch": 2.534711010655473,
      "grad_norm": 0.5383467674255371,
      "learning_rate": 0.0004484516754660058,
      "loss": 3.5615,
      "step": 23550
    },
    {
      "epoch": 2.540092562695081,
      "grad_norm": 0.5506654977798462,
      "learning_rate": 0.00044812843443594436,
      "loss": 3.5675,
      "step": 23600
    },
    {
      "epoch": 2.5454741147346893,
      "grad_norm": 0.5995807647705078,
      "learning_rate": 0.00044780519340588296,
      "loss": 3.5341,
      "step": 23650
    },
    {
      "epoch": 2.550855666774298,
      "grad_norm": 0.5914281606674194,
      "learning_rate": 0.0004474819523758215,
      "loss": 3.5595,
      "step": 23700
    },
    {
      "epoch": 2.556237218813906,
      "grad_norm": 0.5454679727554321,
      "learning_rate": 0.0004471587113457601,
      "loss": 3.56,
      "step": 23750
    },
    {
      "epoch": 2.561618770853514,
      "grad_norm": 0.5431504249572754,
      "learning_rate": 0.00044683547031569874,
      "loss": 3.5787,
      "step": 23800
    },
    {
      "epoch": 2.567000322893122,
      "grad_norm": 0.6172618865966797,
      "learning_rate": 0.0004465122292856373,
      "loss": 3.5569,
      "step": 23850
    },
    {
      "epoch": 2.5723818749327307,
      "grad_norm": 0.5649376511573792,
      "learning_rate": 0.0004461889882555759,
      "loss": 3.5538,
      "step": 23900
    },
    {
      "epoch": 2.577763426972339,
      "grad_norm": 0.5461429953575134,
      "learning_rate": 0.00044586574722551447,
      "loss": 3.5454,
      "step": 23950
    },
    {
      "epoch": 2.583144979011947,
      "grad_norm": 0.5643666982650757,
      "learning_rate": 0.000445542506195453,
      "loss": 3.5581,
      "step": 24000
    },
    {
      "epoch": 2.583144979011947,
      "eval_accuracy": 0.3659202177056183,
      "eval_loss": 3.5391037464141846,
      "eval_runtime": 185.4201,
      "eval_samples_per_second": 97.136,
      "eval_steps_per_second": 6.073,
      "step": 24000
    },
    {
      "epoch": 2.5885265310515555,
      "grad_norm": 0.5423243641853333,
      "learning_rate": 0.0004452257299859928,
      "loss": 3.5436,
      "step": 24050
    },
    {
      "epoch": 2.5939080830911636,
      "grad_norm": 0.6065905094146729,
      "learning_rate": 0.0004449024889559314,
      "loss": 3.5484,
      "step": 24100
    },
    {
      "epoch": 2.5992896351307717,
      "grad_norm": 0.5239328145980835,
      "learning_rate": 0.00044457924792587,
      "loss": 3.5602,
      "step": 24150
    },
    {
      "epoch": 2.60467118717038,
      "grad_norm": 0.5684859156608582,
      "learning_rate": 0.0004442560068958086,
      "loss": 3.5427,
      "step": 24200
    },
    {
      "epoch": 2.610052739209988,
      "grad_norm": 0.5955207347869873,
      "learning_rate": 0.0004439327658657472,
      "loss": 3.57,
      "step": 24250
    },
    {
      "epoch": 2.6154342912495965,
      "grad_norm": 0.5812932252883911,
      "learning_rate": 0.0004436095248356858,
      "loss": 3.5619,
      "step": 24300
    },
    {
      "epoch": 2.6208158432892046,
      "grad_norm": 0.6100989580154419,
      "learning_rate": 0.00044328628380562433,
      "loss": 3.5673,
      "step": 24350
    },
    {
      "epoch": 2.6261973953288127,
      "grad_norm": 0.5262783169746399,
      "learning_rate": 0.0004429630427755629,
      "loss": 3.5484,
      "step": 24400
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 0.5579462647438049,
      "learning_rate": 0.0004426398017455016,
      "loss": 3.5407,
      "step": 24450
    },
    {
      "epoch": 2.6369604994080293,
      "grad_norm": 0.5873637199401855,
      "learning_rate": 0.0004423165607154401,
      "loss": 3.5498,
      "step": 24500
    },
    {
      "epoch": 2.6423420514476375,
      "grad_norm": 0.5810472369194031,
      "learning_rate": 0.0004419933196853787,
      "loss": 3.5468,
      "step": 24550
    },
    {
      "epoch": 2.6477236034872456,
      "grad_norm": 0.6093787550926208,
      "learning_rate": 0.00044167007865531725,
      "loss": 3.5498,
      "step": 24600
    },
    {
      "epoch": 2.653105155526854,
      "grad_norm": 0.5575881004333496,
      "learning_rate": 0.00044134683762525584,
      "loss": 3.5471,
      "step": 24650
    },
    {
      "epoch": 2.658486707566462,
      "grad_norm": 0.5593315362930298,
      "learning_rate": 0.00044102359659519444,
      "loss": 3.5487,
      "step": 24700
    },
    {
      "epoch": 2.6638682596060703,
      "grad_norm": 0.5515527129173279,
      "learning_rate": 0.00044070035556513303,
      "loss": 3.5674,
      "step": 24750
    },
    {
      "epoch": 2.6692498116456784,
      "grad_norm": 0.5453667044639587,
      "learning_rate": 0.00044037711453507163,
      "loss": 3.5493,
      "step": 24800
    },
    {
      "epoch": 2.674631363685287,
      "grad_norm": 0.6053645610809326,
      "learning_rate": 0.0004400538735050102,
      "loss": 3.561,
      "step": 24850
    },
    {
      "epoch": 2.680012915724895,
      "grad_norm": 0.5807383060455322,
      "learning_rate": 0.00043973063247494876,
      "loss": 3.5511,
      "step": 24900
    },
    {
      "epoch": 2.685394467764503,
      "grad_norm": 0.5875633955001831,
      "learning_rate": 0.00043940739144488736,
      "loss": 3.555,
      "step": 24950
    },
    {
      "epoch": 2.6907760198041117,
      "grad_norm": 0.6244288682937622,
      "learning_rate": 0.0004390841504148259,
      "loss": 3.5515,
      "step": 25000
    },
    {
      "epoch": 2.6907760198041117,
      "eval_accuracy": 0.36675293343782844,
      "eval_loss": 3.531473398208618,
      "eval_runtime": 184.949,
      "eval_samples_per_second": 97.384,
      "eval_steps_per_second": 6.088,
      "step": 25000
    },
    {
      "epoch": 2.69615757184372,
      "grad_norm": 0.5658705830574036,
      "learning_rate": 0.00043876090938476455,
      "loss": 3.578,
      "step": 25050
    },
    {
      "epoch": 2.701539123883328,
      "grad_norm": 0.5584858059883118,
      "learning_rate": 0.00043843766835470314,
      "loss": 3.5602,
      "step": 25100
    },
    {
      "epoch": 2.706920675922936,
      "grad_norm": 0.7065500617027283,
      "learning_rate": 0.0004381144273246417,
      "loss": 3.5402,
      "step": 25150
    },
    {
      "epoch": 2.712302227962544,
      "grad_norm": 0.5647648572921753,
      "learning_rate": 0.0004377911862945803,
      "loss": 3.522,
      "step": 25200
    },
    {
      "epoch": 2.7176837800021527,
      "grad_norm": 0.6321160793304443,
      "learning_rate": 0.00043746794526451887,
      "loss": 3.5556,
      "step": 25250
    },
    {
      "epoch": 2.723065332041761,
      "grad_norm": 0.6080904006958008,
      "learning_rate": 0.00043714470423445747,
      "loss": 3.5507,
      "step": 25300
    },
    {
      "epoch": 2.728446884081369,
      "grad_norm": 0.6239748597145081,
      "learning_rate": 0.00043682146320439606,
      "loss": 3.5336,
      "step": 25350
    },
    {
      "epoch": 2.7338284361209775,
      "grad_norm": 0.5854498744010925,
      "learning_rate": 0.00043649822217433466,
      "loss": 3.5417,
      "step": 25400
    },
    {
      "epoch": 2.7392099881605856,
      "grad_norm": 0.597295343875885,
      "learning_rate": 0.0004361749811442732,
      "loss": 3.5644,
      "step": 25450
    },
    {
      "epoch": 2.7445915402001937,
      "grad_norm": 0.6007148623466492,
      "learning_rate": 0.0004358517401142118,
      "loss": 3.5541,
      "step": 25500
    },
    {
      "epoch": 2.749973092239802,
      "grad_norm": 0.5275408625602722,
      "learning_rate": 0.00043552849908415033,
      "loss": 3.5399,
      "step": 25550
    },
    {
      "epoch": 2.7553546442794103,
      "grad_norm": 0.5821923017501831,
      "learning_rate": 0.000435205258054089,
      "loss": 3.5615,
      "step": 25600
    },
    {
      "epoch": 2.7607361963190185,
      "grad_norm": 0.6047101616859436,
      "learning_rate": 0.0004348820170240276,
      "loss": 3.5498,
      "step": 25650
    },
    {
      "epoch": 2.7661177483586266,
      "grad_norm": 0.5546850562095642,
      "learning_rate": 0.0004345587759939661,
      "loss": 3.535,
      "step": 25700
    },
    {
      "epoch": 2.7714993003982347,
      "grad_norm": 0.5674479603767395,
      "learning_rate": 0.0004342355349639047,
      "loss": 3.5392,
      "step": 25750
    },
    {
      "epoch": 2.776880852437843,
      "grad_norm": 0.564844012260437,
      "learning_rate": 0.0004339122939338433,
      "loss": 3.5534,
      "step": 25800
    },
    {
      "epoch": 2.7822624044774513,
      "grad_norm": 0.5702014565467834,
      "learning_rate": 0.00043358905290378184,
      "loss": 3.5486,
      "step": 25850
    },
    {
      "epoch": 2.7876439565170594,
      "grad_norm": 0.5721421837806702,
      "learning_rate": 0.0004332658118737205,
      "loss": 3.5526,
      "step": 25900
    },
    {
      "epoch": 2.793025508556668,
      "grad_norm": 0.5494389533996582,
      "learning_rate": 0.0004329425708436591,
      "loss": 3.5538,
      "step": 25950
    },
    {
      "epoch": 2.798407060596276,
      "grad_norm": 0.5940845012664795,
      "learning_rate": 0.00043261932981359763,
      "loss": 3.5423,
      "step": 26000
    },
    {
      "epoch": 2.798407060596276,
      "eval_accuracy": 0.36804133938153255,
      "eval_loss": 3.5232934951782227,
      "eval_runtime": 185.1304,
      "eval_samples_per_second": 97.288,
      "eval_steps_per_second": 6.082,
      "step": 26000
    },
    {
      "epoch": 2.803788612635884,
      "grad_norm": 0.5880907773971558,
      "learning_rate": 0.00043230255360413744,
      "loss": 3.5406,
      "step": 26050
    },
    {
      "epoch": 2.8091701646754923,
      "grad_norm": 0.5378446578979492,
      "learning_rate": 0.00043197931257407603,
      "loss": 3.5487,
      "step": 26100
    },
    {
      "epoch": 2.8145517167151004,
      "grad_norm": 0.5413463711738586,
      "learning_rate": 0.0004316560715440146,
      "loss": 3.542,
      "step": 26150
    },
    {
      "epoch": 2.819933268754709,
      "grad_norm": 0.5668351650238037,
      "learning_rate": 0.00043133283051395317,
      "loss": 3.55,
      "step": 26200
    },
    {
      "epoch": 2.825314820794317,
      "grad_norm": 0.5521743893623352,
      "learning_rate": 0.0004310095894838918,
      "loss": 3.5502,
      "step": 26250
    },
    {
      "epoch": 2.830696372833925,
      "grad_norm": 0.5967365503311157,
      "learning_rate": 0.0004306863484538304,
      "loss": 3.5285,
      "step": 26300
    },
    {
      "epoch": 2.8360779248735337,
      "grad_norm": 0.5773559212684631,
      "learning_rate": 0.00043036310742376895,
      "loss": 3.5357,
      "step": 26350
    },
    {
      "epoch": 2.841459476913142,
      "grad_norm": 0.616309642791748,
      "learning_rate": 0.00043003986639370754,
      "loss": 3.5282,
      "step": 26400
    },
    {
      "epoch": 2.84684102895275,
      "grad_norm": 0.5834537744522095,
      "learning_rate": 0.0004297166253636461,
      "loss": 3.5359,
      "step": 26450
    },
    {
      "epoch": 2.852222580992358,
      "grad_norm": 0.5581462979316711,
      "learning_rate": 0.0004293933843335847,
      "loss": 3.5465,
      "step": 26500
    },
    {
      "epoch": 2.857604133031966,
      "grad_norm": 0.5349182486534119,
      "learning_rate": 0.00042907014330352333,
      "loss": 3.5251,
      "step": 26550
    },
    {
      "epoch": 2.8629856850715747,
      "grad_norm": 0.5827826857566833,
      "learning_rate": 0.00042874690227346187,
      "loss": 3.538,
      "step": 26600
    },
    {
      "epoch": 2.868367237111183,
      "grad_norm": 0.5464391112327576,
      "learning_rate": 0.00042842366124340046,
      "loss": 3.5237,
      "step": 26650
    },
    {
      "epoch": 2.873748789150791,
      "grad_norm": 0.5807905197143555,
      "learning_rate": 0.00042810042021333906,
      "loss": 3.5479,
      "step": 26700
    },
    {
      "epoch": 2.8791303411903995,
      "grad_norm": 0.584532618522644,
      "learning_rate": 0.0004277771791832776,
      "loss": 3.546,
      "step": 26750
    },
    {
      "epoch": 2.8845118932300076,
      "grad_norm": 0.5569878816604614,
      "learning_rate": 0.0004274539381532162,
      "loss": 3.5326,
      "step": 26800
    },
    {
      "epoch": 2.8898934452696157,
      "grad_norm": 0.5832167267799377,
      "learning_rate": 0.00042713069712315484,
      "loss": 3.5423,
      "step": 26850
    },
    {
      "epoch": 2.895274997309224,
      "grad_norm": 0.531145453453064,
      "learning_rate": 0.0004268074560930934,
      "loss": 3.5367,
      "step": 26900
    },
    {
      "epoch": 2.9006565493488323,
      "grad_norm": 0.5769475102424622,
      "learning_rate": 0.000426484215063032,
      "loss": 3.5419,
      "step": 26950
    },
    {
      "epoch": 2.9060381013884404,
      "grad_norm": 0.5462967157363892,
      "learning_rate": 0.0004261609740329705,
      "loss": 3.5243,
      "step": 27000
    },
    {
      "epoch": 2.9060381013884404,
      "eval_accuracy": 0.36837675084621585,
      "eval_loss": 3.511310338973999,
      "eval_runtime": 184.9613,
      "eval_samples_per_second": 97.377,
      "eval_steps_per_second": 6.088,
      "step": 27000
    },
    {
      "epoch": 2.9114196534280485,
      "grad_norm": 0.5693956017494202,
      "learning_rate": 0.0004258377330029091,
      "loss": 3.534,
      "step": 27050
    },
    {
      "epoch": 2.9168012054676566,
      "grad_norm": 0.6412118673324585,
      "learning_rate": 0.00042551449197284776,
      "loss": 3.5489,
      "step": 27100
    },
    {
      "epoch": 2.922182757507265,
      "grad_norm": 0.591422438621521,
      "learning_rate": 0.0004251912509427863,
      "loss": 3.525,
      "step": 27150
    },
    {
      "epoch": 2.9275643095468733,
      "grad_norm": 0.5912882089614868,
      "learning_rate": 0.0004248680099127249,
      "loss": 3.5463,
      "step": 27200
    },
    {
      "epoch": 2.9329458615864814,
      "grad_norm": 0.5609670281410217,
      "learning_rate": 0.0004245447688826635,
      "loss": 3.5404,
      "step": 27250
    },
    {
      "epoch": 2.93832741362609,
      "grad_norm": 0.6347305774688721,
      "learning_rate": 0.00042422152785260203,
      "loss": 3.5479,
      "step": 27300
    },
    {
      "epoch": 2.943708965665698,
      "grad_norm": 0.5777581334114075,
      "learning_rate": 0.0004238982868225406,
      "loss": 3.5253,
      "step": 27350
    },
    {
      "epoch": 2.949090517705306,
      "grad_norm": 0.5795454382896423,
      "learning_rate": 0.0004235750457924793,
      "loss": 3.5315,
      "step": 27400
    },
    {
      "epoch": 2.9544720697449143,
      "grad_norm": 0.5652514100074768,
      "learning_rate": 0.0004232518047624178,
      "loss": 3.5254,
      "step": 27450
    },
    {
      "epoch": 2.9598536217845224,
      "grad_norm": 0.5999734997749329,
      "learning_rate": 0.0004229285637323564,
      "loss": 3.5425,
      "step": 27500
    },
    {
      "epoch": 2.965235173824131,
      "grad_norm": 0.6210274696350098,
      "learning_rate": 0.00042260532270229495,
      "loss": 3.5249,
      "step": 27550
    },
    {
      "epoch": 2.970616725863739,
      "grad_norm": 0.5799051523208618,
      "learning_rate": 0.00042228208167223354,
      "loss": 3.5389,
      "step": 27600
    },
    {
      "epoch": 2.975998277903347,
      "grad_norm": 0.5363778471946716,
      "learning_rate": 0.00042195884064217214,
      "loss": 3.5328,
      "step": 27650
    },
    {
      "epoch": 2.9813798299429557,
      "grad_norm": 0.5385113954544067,
      "learning_rate": 0.00042163559961211073,
      "loss": 3.5298,
      "step": 27700
    },
    {
      "epoch": 2.986761381982564,
      "grad_norm": 0.5480571985244751,
      "learning_rate": 0.00042131235858204933,
      "loss": 3.5293,
      "step": 27750
    },
    {
      "epoch": 2.992142934022172,
      "grad_norm": 0.736081600189209,
      "learning_rate": 0.0004209891175519879,
      "loss": 3.5378,
      "step": 27800
    },
    {
      "epoch": 2.9975244860617805,
      "grad_norm": 0.6315052509307861,
      "learning_rate": 0.00042066587652192646,
      "loss": 3.5216,
      "step": 27850
    },
    {
      "epoch": 3.0029060381013886,
      "grad_norm": 0.5893605947494507,
      "learning_rate": 0.00042034263549186506,
      "loss": 3.4902,
      "step": 27900
    },
    {
      "epoch": 3.0082875901409967,
      "grad_norm": 0.5847485661506653,
      "learning_rate": 0.0004200193944618036,
      "loss": 3.4621,
      "step": 27950
    },
    {
      "epoch": 3.0136691421806048,
      "grad_norm": 0.5579429268836975,
      "learning_rate": 0.00041969615343174225,
      "loss": 3.4538,
      "step": 28000
    },
    {
      "epoch": 3.0136691421806048,
      "eval_accuracy": 0.36958986033866453,
      "eval_loss": 3.50691819190979,
      "eval_runtime": 185.2417,
      "eval_samples_per_second": 97.23,
      "eval_steps_per_second": 6.079,
      "step": 28000
    },
    {
      "epoch": 3.0190506942202133,
      "grad_norm": 0.5732946395874023,
      "learning_rate": 0.00041937291240168084,
      "loss": 3.4402,
      "step": 28050
    },
    {
      "epoch": 3.0244322462598214,
      "grad_norm": 0.541876494884491,
      "learning_rate": 0.00041905613619222065,
      "loss": 3.4322,
      "step": 28100
    },
    {
      "epoch": 3.0298137982994295,
      "grad_norm": 0.5866724848747253,
      "learning_rate": 0.00041873289516215924,
      "loss": 3.4276,
      "step": 28150
    },
    {
      "epoch": 3.0351953503390376,
      "grad_norm": 0.6123706102371216,
      "learning_rate": 0.0004184096541320978,
      "loss": 3.463,
      "step": 28200
    },
    {
      "epoch": 3.040576902378646,
      "grad_norm": 0.5802378058433533,
      "learning_rate": 0.0004180864131020364,
      "loss": 3.4622,
      "step": 28250
    },
    {
      "epoch": 3.0459584544182543,
      "grad_norm": 0.5802274942398071,
      "learning_rate": 0.0004177631720719749,
      "loss": 3.4607,
      "step": 28300
    },
    {
      "epoch": 3.0513400064578624,
      "grad_norm": 0.5760840177536011,
      "learning_rate": 0.00041743993104191357,
      "loss": 3.435,
      "step": 28350
    },
    {
      "epoch": 3.0567215584974705,
      "grad_norm": 0.6052194833755493,
      "learning_rate": 0.00041711669001185216,
      "loss": 3.4559,
      "step": 28400
    },
    {
      "epoch": 3.062103110537079,
      "grad_norm": 0.5779029726982117,
      "learning_rate": 0.0004167934489817907,
      "loss": 3.4453,
      "step": 28450
    },
    {
      "epoch": 3.067484662576687,
      "grad_norm": 0.5903552174568176,
      "learning_rate": 0.0004164702079517293,
      "loss": 3.4478,
      "step": 28500
    },
    {
      "epoch": 3.0728662146162953,
      "grad_norm": 0.5654973983764648,
      "learning_rate": 0.0004161469669216679,
      "loss": 3.4583,
      "step": 28550
    },
    {
      "epoch": 3.0782477666559034,
      "grad_norm": 0.5794385671615601,
      "learning_rate": 0.00041582372589160643,
      "loss": 3.4467,
      "step": 28600
    },
    {
      "epoch": 3.083629318695512,
      "grad_norm": 0.5537961721420288,
      "learning_rate": 0.0004155004848615451,
      "loss": 3.4501,
      "step": 28650
    },
    {
      "epoch": 3.08901087073512,
      "grad_norm": 0.5887314081192017,
      "learning_rate": 0.0004151772438314837,
      "loss": 3.4452,
      "step": 28700
    },
    {
      "epoch": 3.094392422774728,
      "grad_norm": 0.5306625366210938,
      "learning_rate": 0.0004148540028014222,
      "loss": 3.4747,
      "step": 28750
    },
    {
      "epoch": 3.0997739748143363,
      "grad_norm": 0.6064127683639526,
      "learning_rate": 0.0004145307617713608,
      "loss": 3.4821,
      "step": 28800
    },
    {
      "epoch": 3.105155526853945,
      "grad_norm": 0.5997468829154968,
      "learning_rate": 0.00041420752074129935,
      "loss": 3.4619,
      "step": 28850
    },
    {
      "epoch": 3.110537078893553,
      "grad_norm": 0.5563057065010071,
      "learning_rate": 0.000413884279711238,
      "loss": 3.4953,
      "step": 28900
    },
    {
      "epoch": 3.115918630933161,
      "grad_norm": 0.5288698673248291,
      "learning_rate": 0.0004135610386811766,
      "loss": 3.4621,
      "step": 28950
    },
    {
      "epoch": 3.121300182972769,
      "grad_norm": 0.6248419880867004,
      "learning_rate": 0.00041323779765111514,
      "loss": 3.4722,
      "step": 29000
    },
    {
      "epoch": 3.121300182972769,
      "eval_accuracy": 0.3696654827485834,
      "eval_loss": 3.504861354827881,
      "eval_runtime": 184.903,
      "eval_samples_per_second": 97.408,
      "eval_steps_per_second": 6.09,
      "step": 29000
    },
    {
      "epoch": 3.1266817350123777,
      "grad_norm": 0.5577928423881531,
      "learning_rate": 0.00041291455662105373,
      "loss": 3.4692,
      "step": 29050
    },
    {
      "epoch": 3.132063287051986,
      "grad_norm": 0.5774070024490356,
      "learning_rate": 0.0004125913155909923,
      "loss": 3.4736,
      "step": 29100
    },
    {
      "epoch": 3.137444839091594,
      "grad_norm": 0.5555630922317505,
      "learning_rate": 0.00041226807456093087,
      "loss": 3.4673,
      "step": 29150
    },
    {
      "epoch": 3.1428263911312024,
      "grad_norm": 0.5902732610702515,
      "learning_rate": 0.0004119448335308695,
      "loss": 3.4549,
      "step": 29200
    },
    {
      "epoch": 3.1482079431708105,
      "grad_norm": 0.5566281080245972,
      "learning_rate": 0.0004116215925008081,
      "loss": 3.4549,
      "step": 29250
    },
    {
      "epoch": 3.1535894952104186,
      "grad_norm": 0.591916024684906,
      "learning_rate": 0.00041129835147074665,
      "loss": 3.4726,
      "step": 29300
    },
    {
      "epoch": 3.1589710472500268,
      "grad_norm": 0.5458685159683228,
      "learning_rate": 0.00041097511044068524,
      "loss": 3.4696,
      "step": 29350
    },
    {
      "epoch": 3.1643525992896353,
      "grad_norm": 0.6111952066421509,
      "learning_rate": 0.0004106518694106238,
      "loss": 3.4716,
      "step": 29400
    },
    {
      "epoch": 3.1697341513292434,
      "grad_norm": 0.6616466045379639,
      "learning_rate": 0.0004103286283805624,
      "loss": 3.4673,
      "step": 29450
    },
    {
      "epoch": 3.1751157033688515,
      "grad_norm": 0.5770179629325867,
      "learning_rate": 0.00041000538735050103,
      "loss": 3.4589,
      "step": 29500
    },
    {
      "epoch": 3.1804972554084596,
      "grad_norm": 0.5820581912994385,
      "learning_rate": 0.00040968214632043957,
      "loss": 3.4696,
      "step": 29550
    },
    {
      "epoch": 3.185878807448068,
      "grad_norm": 0.6496612429618835,
      "learning_rate": 0.00040935890529037816,
      "loss": 3.4746,
      "step": 29600
    },
    {
      "epoch": 3.1912603594876763,
      "grad_norm": 0.6107315421104431,
      "learning_rate": 0.00040903566426031676,
      "loss": 3.4712,
      "step": 29650
    },
    {
      "epoch": 3.1966419115272844,
      "grad_norm": 0.5959407091140747,
      "learning_rate": 0.0004087124232302553,
      "loss": 3.4557,
      "step": 29700
    },
    {
      "epoch": 3.2020234635668925,
      "grad_norm": 0.5461238622665405,
      "learning_rate": 0.0004083891822001939,
      "loss": 3.4646,
      "step": 29750
    },
    {
      "epoch": 3.207405015606501,
      "grad_norm": 0.6032586097717285,
      "learning_rate": 0.00040806594117013254,
      "loss": 3.4598,
      "step": 29800
    },
    {
      "epoch": 3.212786567646109,
      "grad_norm": 0.537319540977478,
      "learning_rate": 0.0004077427001400711,
      "loss": 3.4649,
      "step": 29850
    },
    {
      "epoch": 3.2181681196857173,
      "grad_norm": 0.5607523918151855,
      "learning_rate": 0.0004074194591100097,
      "loss": 3.4566,
      "step": 29900
    },
    {
      "epoch": 3.2235496717253254,
      "grad_norm": 0.5795313119888306,
      "learning_rate": 0.0004070962180799482,
      "loss": 3.4576,
      "step": 29950
    },
    {
      "epoch": 3.228931223764934,
      "grad_norm": 0.5723983645439148,
      "learning_rate": 0.0004067729770498868,
      "loss": 3.4524,
      "step": 30000
    },
    {
      "epoch": 3.228931223764934,
      "eval_accuracy": 0.3707188724959448,
      "eval_loss": 3.49776029586792,
      "eval_runtime": 185.1177,
      "eval_samples_per_second": 97.295,
      "eval_steps_per_second": 6.083,
      "step": 30000
    },
    {
      "epoch": 3.234312775804542,
      "grad_norm": 0.6038468480110168,
      "learning_rate": 0.00040644973601982546,
      "loss": 3.4494,
      "step": 30050
    },
    {
      "epoch": 3.23969432784415,
      "grad_norm": 0.5517382025718689,
      "learning_rate": 0.000406126494989764,
      "loss": 3.4749,
      "step": 30100
    },
    {
      "epoch": 3.2450758798837587,
      "grad_norm": 0.6411867737770081,
      "learning_rate": 0.0004058032539597026,
      "loss": 3.4579,
      "step": 30150
    },
    {
      "epoch": 3.250457431923367,
      "grad_norm": 0.607178270816803,
      "learning_rate": 0.00040548001292964114,
      "loss": 3.4644,
      "step": 30200
    },
    {
      "epoch": 3.255838983962975,
      "grad_norm": Infinity,
      "learning_rate": 0.000405163236720181,
      "loss": 3.4585,
      "step": 30250
    },
    {
      "epoch": 3.261220536002583,
      "grad_norm": 0.5709850788116455,
      "learning_rate": 0.00040483999569011954,
      "loss": 3.4721,
      "step": 30300
    },
    {
      "epoch": 3.2666020880421915,
      "grad_norm": 0.5674587488174438,
      "learning_rate": 0.00040451675466005813,
      "loss": 3.4626,
      "step": 30350
    },
    {
      "epoch": 3.2719836400817996,
      "grad_norm": 0.5631972551345825,
      "learning_rate": 0.0004041935136299967,
      "loss": 3.4719,
      "step": 30400
    },
    {
      "epoch": 3.2773651921214078,
      "grad_norm": 0.5942214727401733,
      "learning_rate": 0.0004038702725999353,
      "loss": 3.4597,
      "step": 30450
    },
    {
      "epoch": 3.282746744161016,
      "grad_norm": 0.5321446061134338,
      "learning_rate": 0.0004035470315698739,
      "loss": 3.4562,
      "step": 30500
    },
    {
      "epoch": 3.2881282962006244,
      "grad_norm": 0.5662004947662354,
      "learning_rate": 0.0004032237905398125,
      "loss": 3.4801,
      "step": 30550
    },
    {
      "epoch": 3.2935098482402325,
      "grad_norm": 0.5357405543327332,
      "learning_rate": 0.00040290054950975105,
      "loss": 3.4737,
      "step": 30600
    },
    {
      "epoch": 3.2988914002798406,
      "grad_norm": 0.6253825426101685,
      "learning_rate": 0.00040257730847968965,
      "loss": 3.4632,
      "step": 30650
    },
    {
      "epoch": 3.304272952319449,
      "grad_norm": 0.5765745639801025,
      "learning_rate": 0.0004022540674496283,
      "loss": 3.4847,
      "step": 30700
    },
    {
      "epoch": 3.3096545043590573,
      "grad_norm": 0.60694819688797,
      "learning_rate": 0.00040193082641956684,
      "loss": 3.4657,
      "step": 30750
    },
    {
      "epoch": 3.3150360563986654,
      "grad_norm": 0.6164451837539673,
      "learning_rate": 0.00040160758538950543,
      "loss": 3.4748,
      "step": 30800
    },
    {
      "epoch": 3.3204176084382735,
      "grad_norm": 0.5834468007087708,
      "learning_rate": 0.00040128434435944397,
      "loss": 3.4726,
      "step": 30850
    },
    {
      "epoch": 3.3257991604778816,
      "grad_norm": 0.5414763689041138,
      "learning_rate": 0.00040096110332938257,
      "loss": 3.4621,
      "step": 30900
    },
    {
      "epoch": 3.33118071251749,
      "grad_norm": 0.5613992810249329,
      "learning_rate": 0.00040063786229932116,
      "loss": 3.4607,
      "step": 30950
    },
    {
      "epoch": 3.3365622645570983,
      "grad_norm": 0.5616213083267212,
      "learning_rate": 0.00040031462126925975,
      "loss": 3.4688,
      "step": 31000
    },
    {
      "epoch": 3.3365622645570983,
      "eval_accuracy": 0.37120759318533436,
      "eval_loss": 3.4941439628601074,
      "eval_runtime": 185.2821,
      "eval_samples_per_second": 97.209,
      "eval_steps_per_second": 6.077,
      "step": 31000
    },
    {
      "epoch": 3.3419438165967064,
      "grad_norm": 0.6164249777793884,
      "learning_rate": 0.00039999138023919835,
      "loss": 3.4664,
      "step": 31050
    },
    {
      "epoch": 3.347325368636315,
      "grad_norm": 0.5682007670402527,
      "learning_rate": 0.00039966813920913694,
      "loss": 3.4711,
      "step": 31100
    },
    {
      "epoch": 3.352706920675923,
      "grad_norm": 0.5876696109771729,
      "learning_rate": 0.0003993448981790755,
      "loss": 3.4691,
      "step": 31150
    },
    {
      "epoch": 3.358088472715531,
      "grad_norm": 0.6580924391746521,
      "learning_rate": 0.0003990216571490141,
      "loss": 3.4781,
      "step": 31200
    },
    {
      "epoch": 3.3634700247551392,
      "grad_norm": 0.5489287972450256,
      "learning_rate": 0.0003986984161189526,
      "loss": 3.4588,
      "step": 31250
    },
    {
      "epoch": 3.368851576794748,
      "grad_norm": 0.5927851796150208,
      "learning_rate": 0.00039837517508889127,
      "loss": 3.465,
      "step": 31300
    },
    {
      "epoch": 3.374233128834356,
      "grad_norm": 0.5911576151847839,
      "learning_rate": 0.00039805193405882986,
      "loss": 3.4652,
      "step": 31350
    },
    {
      "epoch": 3.379614680873964,
      "grad_norm": 0.534320592880249,
      "learning_rate": 0.0003977286930287684,
      "loss": 3.4811,
      "step": 31400
    },
    {
      "epoch": 3.384996232913572,
      "grad_norm": 0.5626870393753052,
      "learning_rate": 0.000397405451998707,
      "loss": 3.47,
      "step": 31450
    },
    {
      "epoch": 3.3903777849531807,
      "grad_norm": 0.5819312930107117,
      "learning_rate": 0.00039708221096864554,
      "loss": 3.4702,
      "step": 31500
    },
    {
      "epoch": 3.3957593369927888,
      "grad_norm": 0.5503734350204468,
      "learning_rate": 0.00039675896993858413,
      "loss": 3.4701,
      "step": 31550
    },
    {
      "epoch": 3.401140889032397,
      "grad_norm": 0.5898651480674744,
      "learning_rate": 0.0003964357289085228,
      "loss": 3.4736,
      "step": 31600
    },
    {
      "epoch": 3.4065224410720054,
      "grad_norm": 0.6126618981361389,
      "learning_rate": 0.0003961124878784613,
      "loss": 3.4664,
      "step": 31650
    },
    {
      "epoch": 3.4119039931116135,
      "grad_norm": 0.5383553504943848,
      "learning_rate": 0.0003957892468483999,
      "loss": 3.4685,
      "step": 31700
    },
    {
      "epoch": 3.4172855451512216,
      "grad_norm": 0.5667200088500977,
      "learning_rate": 0.0003954660058183385,
      "loss": 3.4513,
      "step": 31750
    },
    {
      "epoch": 3.4226670971908297,
      "grad_norm": 0.6069493293762207,
      "learning_rate": 0.00039514276478827705,
      "loss": 3.4609,
      "step": 31800
    },
    {
      "epoch": 3.428048649230438,
      "grad_norm": 0.5866658687591553,
      "learning_rate": 0.0003948195237582157,
      "loss": 3.469,
      "step": 31850
    },
    {
      "epoch": 3.4334302012700464,
      "grad_norm": 0.6032720804214478,
      "learning_rate": 0.0003944962827281543,
      "loss": 3.4837,
      "step": 31900
    },
    {
      "epoch": 3.4388117533096545,
      "grad_norm": 0.5675144791603088,
      "learning_rate": 0.00039417304169809284,
      "loss": 3.4723,
      "step": 31950
    },
    {
      "epoch": 3.4441933053492626,
      "grad_norm": 0.5984659194946289,
      "learning_rate": 0.00039384980066803143,
      "loss": 3.4785,
      "step": 32000
    },
    {
      "epoch": 3.4441933053492626,
      "eval_accuracy": 0.3718997120807126,
      "eval_loss": 3.4841418266296387,
      "eval_runtime": 185.0435,
      "eval_samples_per_second": 97.334,
      "eval_steps_per_second": 6.085,
      "step": 32000
    },
    {
      "epoch": 3.449574857388871,
      "grad_norm": 0.5830808877944946,
      "learning_rate": 0.00039352655963796997,
      "loss": 3.4587,
      "step": 32050
    },
    {
      "epoch": 3.4549564094284793,
      "grad_norm": 0.5691652297973633,
      "learning_rate": 0.00039320331860790857,
      "loss": 3.4469,
      "step": 32100
    },
    {
      "epoch": 3.4603379614680874,
      "grad_norm": 0.6014079451560974,
      "learning_rate": 0.0003928800775778472,
      "loss": 3.4651,
      "step": 32150
    },
    {
      "epoch": 3.4657195135076955,
      "grad_norm": 0.6116219162940979,
      "learning_rate": 0.00039255683654778576,
      "loss": 3.4843,
      "step": 32200
    },
    {
      "epoch": 3.471101065547304,
      "grad_norm": 0.5671770572662354,
      "learning_rate": 0.00039223359551772435,
      "loss": 3.4678,
      "step": 32250
    },
    {
      "epoch": 3.476482617586912,
      "grad_norm": 0.5948795080184937,
      "learning_rate": 0.00039191681930826416,
      "loss": 3.4701,
      "step": 32300
    },
    {
      "epoch": 3.4818641696265202,
      "grad_norm": 0.6259046792984009,
      "learning_rate": 0.00039159357827820275,
      "loss": 3.4466,
      "step": 32350
    },
    {
      "epoch": 3.4872457216661283,
      "grad_norm": 0.6268399953842163,
      "learning_rate": 0.00039127033724814135,
      "loss": 3.4671,
      "step": 32400
    },
    {
      "epoch": 3.492627273705737,
      "grad_norm": 0.5916853547096252,
      "learning_rate": 0.0003909470962180799,
      "loss": 3.4701,
      "step": 32450
    },
    {
      "epoch": 3.498008825745345,
      "grad_norm": 0.6028156280517578,
      "learning_rate": 0.00039062385518801854,
      "loss": 3.4675,
      "step": 32500
    },
    {
      "epoch": 3.503390377784953,
      "grad_norm": 0.6164430379867554,
      "learning_rate": 0.00039030061415795713,
      "loss": 3.4734,
      "step": 32550
    },
    {
      "epoch": 3.5087719298245617,
      "grad_norm": 0.5968393087387085,
      "learning_rate": 0.00038997737312789567,
      "loss": 3.46,
      "step": 32600
    },
    {
      "epoch": 3.5141534818641698,
      "grad_norm": 0.5949715375900269,
      "learning_rate": 0.00038965413209783426,
      "loss": 3.4725,
      "step": 32650
    },
    {
      "epoch": 3.519535033903778,
      "grad_norm": 0.6135655641555786,
      "learning_rate": 0.0003893308910677728,
      "loss": 3.4849,
      "step": 32700
    },
    {
      "epoch": 3.524916585943386,
      "grad_norm": 0.5821818709373474,
      "learning_rate": 0.0003890076500377114,
      "loss": 3.4577,
      "step": 32750
    },
    {
      "epoch": 3.530298137982994,
      "grad_norm": 0.5963186621665955,
      "learning_rate": 0.00038868440900765005,
      "loss": 3.4839,
      "step": 32800
    },
    {
      "epoch": 3.5356796900226026,
      "grad_norm": 0.5958220362663269,
      "learning_rate": 0.0003883611679775886,
      "loss": 3.4784,
      "step": 32850
    },
    {
      "epoch": 3.5410612420622107,
      "grad_norm": 0.5943959951400757,
      "learning_rate": 0.0003880379269475272,
      "loss": 3.4655,
      "step": 32900
    },
    {
      "epoch": 3.546442794101819,
      "grad_norm": 0.6073140501976013,
      "learning_rate": 0.0003877146859174657,
      "loss": 3.4658,
      "step": 32950
    },
    {
      "epoch": 3.5518243461414274,
      "grad_norm": 0.5739242434501648,
      "learning_rate": 0.0003873914448874043,
      "loss": 3.4753,
      "step": 33000
    },
    {
      "epoch": 3.5518243461414274,
      "eval_accuracy": 0.3729022522765768,
      "eval_loss": 3.4780356884002686,
      "eval_runtime": 185.1443,
      "eval_samples_per_second": 97.281,
      "eval_steps_per_second": 6.082,
      "step": 33000
    },
    {
      "epoch": 3.5572058981810355,
      "grad_norm": 0.5834700465202332,
      "learning_rate": 0.0003870682038573429,
      "loss": 3.4607,
      "step": 33050
    },
    {
      "epoch": 3.5625874502206436,
      "grad_norm": 0.5756320357322693,
      "learning_rate": 0.0003867449628272815,
      "loss": 3.4727,
      "step": 33100
    },
    {
      "epoch": 3.5679690022602517,
      "grad_norm": 0.5543476343154907,
      "learning_rate": 0.0003864217217972201,
      "loss": 3.4763,
      "step": 33150
    },
    {
      "epoch": 3.57335055429986,
      "grad_norm": 0.5621790289878845,
      "learning_rate": 0.0003860984807671587,
      "loss": 3.455,
      "step": 33200
    },
    {
      "epoch": 3.5787321063394684,
      "grad_norm": 0.6197202801704407,
      "learning_rate": 0.00038577523973709724,
      "loss": 3.4623,
      "step": 33250
    },
    {
      "epoch": 3.5841136583790765,
      "grad_norm": 0.565671443939209,
      "learning_rate": 0.00038545199870703583,
      "loss": 3.4724,
      "step": 33300
    },
    {
      "epoch": 3.5894952104186846,
      "grad_norm": 0.5738323330879211,
      "learning_rate": 0.0003851287576769744,
      "loss": 3.4814,
      "step": 33350
    },
    {
      "epoch": 3.594876762458293,
      "grad_norm": 0.611393392086029,
      "learning_rate": 0.000384805516646913,
      "loss": 3.4659,
      "step": 33400
    },
    {
      "epoch": 3.6002583144979012,
      "grad_norm": 0.5913741588592529,
      "learning_rate": 0.0003844822756168516,
      "loss": 3.4689,
      "step": 33450
    },
    {
      "epoch": 3.6056398665375093,
      "grad_norm": 0.6085612177848816,
      "learning_rate": 0.00038415903458679016,
      "loss": 3.4796,
      "step": 33500
    },
    {
      "epoch": 3.611021418577118,
      "grad_norm": 0.5542953610420227,
      "learning_rate": 0.00038383579355672875,
      "loss": 3.4777,
      "step": 33550
    },
    {
      "epoch": 3.616402970616726,
      "grad_norm": 0.5730182528495789,
      "learning_rate": 0.00038351255252666735,
      "loss": 3.4677,
      "step": 33600
    },
    {
      "epoch": 3.621784522656334,
      "grad_norm": 0.5870245695114136,
      "learning_rate": 0.00038318931149660594,
      "loss": 3.4829,
      "step": 33650
    },
    {
      "epoch": 3.627166074695942,
      "grad_norm": 0.5914954543113708,
      "learning_rate": 0.00038286607046654454,
      "loss": 3.4645,
      "step": 33700
    },
    {
      "epoch": 3.6325476267355503,
      "grad_norm": 0.6237531304359436,
      "learning_rate": 0.00038254282943648313,
      "loss": 3.4608,
      "step": 33750
    },
    {
      "epoch": 3.637929178775159,
      "grad_norm": 0.5946942567825317,
      "learning_rate": 0.00038221958840642167,
      "loss": 3.4513,
      "step": 33800
    },
    {
      "epoch": 3.643310730814767,
      "grad_norm": 0.5923416614532471,
      "learning_rate": 0.00038189634737636027,
      "loss": 3.4839,
      "step": 33850
    },
    {
      "epoch": 3.648692282854375,
      "grad_norm": 0.5869930982589722,
      "learning_rate": 0.0003815731063462988,
      "loss": 3.4562,
      "step": 33900
    },
    {
      "epoch": 3.6540738348939836,
      "grad_norm": 0.5318478345870972,
      "learning_rate": 0.00038124986531623745,
      "loss": 3.4751,
      "step": 33950
    },
    {
      "epoch": 3.6594553869335917,
      "grad_norm": 0.7541871666908264,
      "learning_rate": 0.00038092662428617605,
      "loss": 3.4545,
      "step": 34000
    },
    {
      "epoch": 3.6594553869335917,
      "eval_accuracy": 0.37322821094002023,
      "eval_loss": 3.4730207920074463,
      "eval_runtime": 185.2424,
      "eval_samples_per_second": 97.229,
      "eval_steps_per_second": 6.079,
      "step": 34000
    },
    {
      "epoch": 3.6648369389732,
      "grad_norm": 0.564893901348114,
      "learning_rate": 0.0003806033832561146,
      "loss": 3.467,
      "step": 34050
    },
    {
      "epoch": 3.670218491012808,
      "grad_norm": 0.5987644791603088,
      "learning_rate": 0.0003802801422260532,
      "loss": 3.4678,
      "step": 34100
    },
    {
      "epoch": 3.675600043052416,
      "grad_norm": 0.5589121580123901,
      "learning_rate": 0.0003799569011959918,
      "loss": 3.4554,
      "step": 34150
    },
    {
      "epoch": 3.6809815950920246,
      "grad_norm": 0.574065089225769,
      "learning_rate": 0.0003796336601659303,
      "loss": 3.4735,
      "step": 34200
    },
    {
      "epoch": 3.6863631471316327,
      "grad_norm": 0.6161201000213623,
      "learning_rate": 0.00037931041913586897,
      "loss": 3.4657,
      "step": 34250
    },
    {
      "epoch": 3.691744699171241,
      "grad_norm": 0.5860987305641174,
      "learning_rate": 0.00037898717810580756,
      "loss": 3.464,
      "step": 34300
    },
    {
      "epoch": 3.6971262512108494,
      "grad_norm": 0.5846574306488037,
      "learning_rate": 0.0003786639370757461,
      "loss": 3.4506,
      "step": 34350
    },
    {
      "epoch": 3.7025078032504575,
      "grad_norm": 0.6496445536613464,
      "learning_rate": 0.0003783406960456847,
      "loss": 3.4898,
      "step": 34400
    },
    {
      "epoch": 3.7078893552900656,
      "grad_norm": 0.6443366408348083,
      "learning_rate": 0.0003780239198362245,
      "loss": 3.4684,
      "step": 34450
    },
    {
      "epoch": 3.713270907329674,
      "grad_norm": 0.57025545835495,
      "learning_rate": 0.0003777006788061631,
      "loss": 3.4405,
      "step": 34500
    },
    {
      "epoch": 3.7186524593692822,
      "grad_norm": 0.5932455658912659,
      "learning_rate": 0.00037737743777610164,
      "loss": 3.4706,
      "step": 34550
    },
    {
      "epoch": 3.7240340114088903,
      "grad_norm": 0.6124030351638794,
      "learning_rate": 0.0003770541967460403,
      "loss": 3.4607,
      "step": 34600
    },
    {
      "epoch": 3.7294155634484984,
      "grad_norm": 0.5861355066299438,
      "learning_rate": 0.0003767309557159789,
      "loss": 3.4802,
      "step": 34650
    },
    {
      "epoch": 3.7347971154881066,
      "grad_norm": 0.5718616843223572,
      "learning_rate": 0.0003764077146859174,
      "loss": 3.4549,
      "step": 34700
    },
    {
      "epoch": 3.740178667527715,
      "grad_norm": 0.5628113746643066,
      "learning_rate": 0.000376084473655856,
      "loss": 3.4763,
      "step": 34750
    },
    {
      "epoch": 3.745560219567323,
      "grad_norm": 0.6013800501823425,
      "learning_rate": 0.00037576123262579456,
      "loss": 3.4656,
      "step": 34800
    },
    {
      "epoch": 3.7509417716069313,
      "grad_norm": 0.6024396419525146,
      "learning_rate": 0.00037543799159573315,
      "loss": 3.4701,
      "step": 34850
    },
    {
      "epoch": 3.75632332364654,
      "grad_norm": 0.6883943676948547,
      "learning_rate": 0.0003751147505656718,
      "loss": 3.4542,
      "step": 34900
    },
    {
      "epoch": 3.761704875686148,
      "grad_norm": 0.558179497718811,
      "learning_rate": 0.00037479150953561034,
      "loss": 3.4838,
      "step": 34950
    },
    {
      "epoch": 3.767086427725756,
      "grad_norm": 0.6130647659301758,
      "learning_rate": 0.00037446826850554894,
      "loss": 3.4517,
      "step": 35000
    },
    {
      "epoch": 3.767086427725756,
      "eval_accuracy": 0.3737176921996245,
      "eval_loss": 3.4650635719299316,
      "eval_runtime": 185.3118,
      "eval_samples_per_second": 97.193,
      "eval_steps_per_second": 6.076,
      "step": 35000
    },
    {
      "epoch": 3.772467979765364,
      "grad_norm": 0.6228893399238586,
      "learning_rate": 0.00037414502747548753,
      "loss": 3.4709,
      "step": 35050
    },
    {
      "epoch": 3.7778495318049723,
      "grad_norm": 0.6668744683265686,
      "learning_rate": 0.00037382178644542607,
      "loss": 3.4727,
      "step": 35100
    },
    {
      "epoch": 3.783231083844581,
      "grad_norm": 0.5576463937759399,
      "learning_rate": 0.00037349854541536467,
      "loss": 3.4832,
      "step": 35150
    },
    {
      "epoch": 3.788612635884189,
      "grad_norm": 0.6484488844871521,
      "learning_rate": 0.0003731753043853033,
      "loss": 3.4536,
      "step": 35200
    },
    {
      "epoch": 3.793994187923797,
      "grad_norm": 0.6269448399543762,
      "learning_rate": 0.00037285206335524186,
      "loss": 3.4712,
      "step": 35250
    },
    {
      "epoch": 3.7993757399634056,
      "grad_norm": 0.5891007781028748,
      "learning_rate": 0.00037252882232518045,
      "loss": 3.4616,
      "step": 35300
    },
    {
      "epoch": 3.8047572920030137,
      "grad_norm": 0.5909605622291565,
      "learning_rate": 0.000372205581295119,
      "loss": 3.4482,
      "step": 35350
    },
    {
      "epoch": 3.810138844042622,
      "grad_norm": 0.5795333981513977,
      "learning_rate": 0.0003718823402650576,
      "loss": 3.4754,
      "step": 35400
    },
    {
      "epoch": 3.8155203960822304,
      "grad_norm": 0.5827459096908569,
      "learning_rate": 0.00037155909923499624,
      "loss": 3.4647,
      "step": 35450
    },
    {
      "epoch": 3.8209019481218385,
      "grad_norm": 0.5595166683197021,
      "learning_rate": 0.0003712358582049348,
      "loss": 3.4658,
      "step": 35500
    },
    {
      "epoch": 3.8262835001614466,
      "grad_norm": 0.6286375522613525,
      "learning_rate": 0.00037091261717487337,
      "loss": 3.4525,
      "step": 35550
    },
    {
      "epoch": 3.8316650522010547,
      "grad_norm": 0.5683946013450623,
      "learning_rate": 0.00037058937614481197,
      "loss": 3.4528,
      "step": 35600
    },
    {
      "epoch": 3.837046604240663,
      "grad_norm": 0.589246928691864,
      "learning_rate": 0.0003702661351147505,
      "loss": 3.4576,
      "step": 35650
    },
    {
      "epoch": 3.8424281562802713,
      "grad_norm": 0.5900267362594604,
      "learning_rate": 0.0003699428940846891,
      "loss": 3.4754,
      "step": 35700
    },
    {
      "epoch": 3.8478097083198795,
      "grad_norm": 0.5886877775192261,
      "learning_rate": 0.00036961965305462775,
      "loss": 3.4576,
      "step": 35750
    },
    {
      "epoch": 3.8531912603594876,
      "grad_norm": 0.5695244669914246,
      "learning_rate": 0.0003692964120245663,
      "loss": 3.4605,
      "step": 35800
    },
    {
      "epoch": 3.858572812399096,
      "grad_norm": 0.5602601766586304,
      "learning_rate": 0.0003689731709945049,
      "loss": 3.4631,
      "step": 35850
    },
    {
      "epoch": 3.863954364438704,
      "grad_norm": 0.6068081259727478,
      "learning_rate": 0.0003686499299644434,
      "loss": 3.4506,
      "step": 35900
    },
    {
      "epoch": 3.8693359164783123,
      "grad_norm": 0.5892680883407593,
      "learning_rate": 0.000368326688934382,
      "loss": 3.4749,
      "step": 35950
    },
    {
      "epoch": 3.8747174685179204,
      "grad_norm": 0.5890790820121765,
      "learning_rate": 0.0003680034479043206,
      "loss": 3.4909,
      "step": 36000
    },
    {
      "epoch": 3.8747174685179204,
      "eval_accuracy": 0.3744507732337088,
      "eval_loss": 3.459874391555786,
      "eval_runtime": 184.9384,
      "eval_samples_per_second": 97.389,
      "eval_steps_per_second": 6.089,
      "step": 36000
    },
    {
      "epoch": 3.8800990205575285,
      "grad_norm": 0.5913886427879333,
      "learning_rate": 0.0003676802068742592,
      "loss": 3.4775,
      "step": 36050
    },
    {
      "epoch": 3.885480572597137,
      "grad_norm": 0.6496626138687134,
      "learning_rate": 0.0003673569658441978,
      "loss": 3.4521,
      "step": 36100
    },
    {
      "epoch": 3.890862124636745,
      "grad_norm": 0.5794878005981445,
      "learning_rate": 0.0003670337248141364,
      "loss": 3.4606,
      "step": 36150
    },
    {
      "epoch": 3.8962436766763533,
      "grad_norm": 0.6062841415405273,
      "learning_rate": 0.00036671048378407494,
      "loss": 3.4604,
      "step": 36200
    },
    {
      "epoch": 3.901625228715962,
      "grad_norm": 0.597095251083374,
      "learning_rate": 0.00036638724275401353,
      "loss": 3.4558,
      "step": 36250
    },
    {
      "epoch": 3.90700678075557,
      "grad_norm": 0.5854865312576294,
      "learning_rate": 0.0003660640017239522,
      "loss": 3.4666,
      "step": 36300
    },
    {
      "epoch": 3.912388332795178,
      "grad_norm": 0.570051372051239,
      "learning_rate": 0.0003657407606938907,
      "loss": 3.4525,
      "step": 36350
    },
    {
      "epoch": 3.9177698848347866,
      "grad_norm": 0.6018213033676147,
      "learning_rate": 0.0003654175196638293,
      "loss": 3.4561,
      "step": 36400
    },
    {
      "epoch": 3.9231514368743947,
      "grad_norm": 0.5964256525039673,
      "learning_rate": 0.00036509427863376786,
      "loss": 3.4557,
      "step": 36450
    },
    {
      "epoch": 3.928532988914003,
      "grad_norm": 0.5669384002685547,
      "learning_rate": 0.00036477103760370645,
      "loss": 3.46,
      "step": 36500
    },
    {
      "epoch": 3.933914540953611,
      "grad_norm": 0.5647692084312439,
      "learning_rate": 0.00036445426139424626,
      "loss": 3.4593,
      "step": 36550
    },
    {
      "epoch": 3.939296092993219,
      "grad_norm": 0.599595844745636,
      "learning_rate": 0.00036413102036418485,
      "loss": 3.4597,
      "step": 36600
    },
    {
      "epoch": 3.9446776450328276,
      "grad_norm": 0.601189911365509,
      "learning_rate": 0.0003638077793341234,
      "loss": 3.453,
      "step": 36650
    },
    {
      "epoch": 3.9500591970724357,
      "grad_norm": 0.5915253758430481,
      "learning_rate": 0.00036348453830406204,
      "loss": 3.4619,
      "step": 36700
    },
    {
      "epoch": 3.955440749112044,
      "grad_norm": 0.6263431906700134,
      "learning_rate": 0.00036316129727400064,
      "loss": 3.4506,
      "step": 36750
    },
    {
      "epoch": 3.9608223011516523,
      "grad_norm": 0.5876361131668091,
      "learning_rate": 0.0003628380562439392,
      "loss": 3.4579,
      "step": 36800
    },
    {
      "epoch": 3.9662038531912605,
      "grad_norm": 0.6339970827102661,
      "learning_rate": 0.00036251481521387777,
      "loss": 3.4642,
      "step": 36850
    },
    {
      "epoch": 3.9715854052308686,
      "grad_norm": 0.6295678019523621,
      "learning_rate": 0.00036219157418381637,
      "loss": 3.4417,
      "step": 36900
    },
    {
      "epoch": 3.9769669572704767,
      "grad_norm": 0.5976482629776001,
      "learning_rate": 0.0003618683331537549,
      "loss": 3.4559,
      "step": 36950
    },
    {
      "epoch": 3.9823485093100848,
      "grad_norm": 0.5901889204978943,
      "learning_rate": 0.00036154509212369356,
      "loss": 3.4628,
      "step": 37000
    },
    {
      "epoch": 3.9823485093100848,
      "eval_accuracy": 0.37490146541236324,
      "eval_loss": 3.4546635150909424,
      "eval_runtime": 185.0719,
      "eval_samples_per_second": 97.319,
      "eval_steps_per_second": 6.084,
      "step": 37000
    },
    {
      "epoch": 3.9877300613496933,
      "grad_norm": 0.5821202397346497,
      "learning_rate": 0.00036122185109363215,
      "loss": 3.4679,
      "step": 37050
    },
    {
      "epoch": 3.9931116133893014,
      "grad_norm": 0.6086378693580627,
      "learning_rate": 0.0003608986100635707,
      "loss": 3.452,
      "step": 37100
    },
    {
      "epoch": 3.9984931654289095,
      "grad_norm": 0.6031500697135925,
      "learning_rate": 0.0003605753690335093,
      "loss": 3.4467,
      "step": 37150
    },
    {
      "epoch": 4.003874717468518,
      "grad_norm": 0.6034141182899475,
      "learning_rate": 0.0003602521280034478,
      "loss": 3.409,
      "step": 37200
    },
    {
      "epoch": 4.009256269508126,
      "grad_norm": 0.6197435855865479,
      "learning_rate": 0.0003599288869733865,
      "loss": 3.3649,
      "step": 37250
    },
    {
      "epoch": 4.014637821547734,
      "grad_norm": 0.6221106052398682,
      "learning_rate": 0.00035960564594332507,
      "loss": 3.3763,
      "step": 37300
    },
    {
      "epoch": 4.020019373587343,
      "grad_norm": 0.5424720048904419,
      "learning_rate": 0.0003592824049132636,
      "loss": 3.3671,
      "step": 37350
    },
    {
      "epoch": 4.0254009256269505,
      "grad_norm": 0.6059404015541077,
      "learning_rate": 0.0003589591638832022,
      "loss": 3.3528,
      "step": 37400
    },
    {
      "epoch": 4.030782477666559,
      "grad_norm": 0.576892614364624,
      "learning_rate": 0.0003586359228531408,
      "loss": 3.3652,
      "step": 37450
    },
    {
      "epoch": 4.036164029706168,
      "grad_norm": 0.6605005860328674,
      "learning_rate": 0.00035831268182307934,
      "loss": 3.3497,
      "step": 37500
    },
    {
      "epoch": 4.041545581745775,
      "grad_norm": 0.6000856161117554,
      "learning_rate": 0.000357989440793018,
      "loss": 3.385,
      "step": 37550
    },
    {
      "epoch": 4.046927133785384,
      "grad_norm": 0.5874991416931152,
      "learning_rate": 0.0003576661997629566,
      "loss": 3.3832,
      "step": 37600
    },
    {
      "epoch": 4.0523086858249915,
      "grad_norm": 0.5919212102890015,
      "learning_rate": 0.0003573429587328951,
      "loss": 3.3731,
      "step": 37650
    },
    {
      "epoch": 4.0576902378646,
      "grad_norm": 0.6303485631942749,
      "learning_rate": 0.0003570197177028337,
      "loss": 3.3712,
      "step": 37700
    },
    {
      "epoch": 4.063071789904209,
      "grad_norm": 0.6002392172813416,
      "learning_rate": 0.00035669647667277226,
      "loss": 3.3736,
      "step": 37750
    },
    {
      "epoch": 4.068453341943816,
      "grad_norm": 0.6628726720809937,
      "learning_rate": 0.00035637323564271085,
      "loss": 3.3761,
      "step": 37800
    },
    {
      "epoch": 4.073834893983425,
      "grad_norm": 0.6533042192459106,
      "learning_rate": 0.0003560499946126495,
      "loss": 3.3621,
      "step": 37850
    },
    {
      "epoch": 4.079216446023033,
      "grad_norm": 0.6461078524589539,
      "learning_rate": 0.00035572675358258804,
      "loss": 3.3777,
      "step": 37900
    },
    {
      "epoch": 4.084597998062641,
      "grad_norm": 0.6249018907546997,
      "learning_rate": 0.00035540351255252664,
      "loss": 3.3767,
      "step": 37950
    },
    {
      "epoch": 4.08997955010225,
      "grad_norm": 0.60147625207901,
      "learning_rate": 0.00035508027152246523,
      "loss": 3.3911,
      "step": 38000
    },
    {
      "epoch": 4.08997955010225,
      "eval_accuracy": 0.37528794373431934,
      "eval_loss": 3.4564130306243896,
      "eval_runtime": 185.4356,
      "eval_samples_per_second": 97.128,
      "eval_steps_per_second": 6.072,
      "step": 38000
    },
    {
      "epoch": 4.095361102141858,
      "grad_norm": 0.6334101557731628,
      "learning_rate": 0.0003547570304924038,
      "loss": 3.3511,
      "step": 38050
    },
    {
      "epoch": 4.100742654181466,
      "grad_norm": 0.648288369178772,
      "learning_rate": 0.0003544337894623424,
      "loss": 3.3713,
      "step": 38100
    },
    {
      "epoch": 4.106124206221074,
      "grad_norm": 0.5762490630149841,
      "learning_rate": 0.000354110548432281,
      "loss": 3.375,
      "step": 38150
    },
    {
      "epoch": 4.111505758260682,
      "grad_norm": 0.6332446336746216,
      "learning_rate": 0.00035378730740221956,
      "loss": 3.3811,
      "step": 38200
    },
    {
      "epoch": 4.1168873103002905,
      "grad_norm": 0.5964016914367676,
      "learning_rate": 0.00035346406637215815,
      "loss": 3.3832,
      "step": 38250
    },
    {
      "epoch": 4.122268862339899,
      "grad_norm": 0.6264393329620361,
      "learning_rate": 0.0003531408253420967,
      "loss": 3.4097,
      "step": 38300
    },
    {
      "epoch": 4.127650414379507,
      "grad_norm": 0.6108452677726746,
      "learning_rate": 0.0003528175843120353,
      "loss": 3.3868,
      "step": 38350
    },
    {
      "epoch": 4.133031966419115,
      "grad_norm": 0.602462112903595,
      "learning_rate": 0.00035249434328197394,
      "loss": 3.3876,
      "step": 38400
    },
    {
      "epoch": 4.138413518458724,
      "grad_norm": 0.6299335360527039,
      "learning_rate": 0.0003521711022519125,
      "loss": 3.3969,
      "step": 38450
    },
    {
      "epoch": 4.1437950704983315,
      "grad_norm": 0.6169623136520386,
      "learning_rate": 0.00035184786122185107,
      "loss": 3.3865,
      "step": 38500
    },
    {
      "epoch": 4.14917662253794,
      "grad_norm": 0.5948913097381592,
      "learning_rate": 0.0003515310850123909,
      "loss": 3.3873,
      "step": 38550
    },
    {
      "epoch": 4.154558174577549,
      "grad_norm": 0.6054384708404541,
      "learning_rate": 0.00035120784398232947,
      "loss": 3.3802,
      "step": 38600
    },
    {
      "epoch": 4.159939726617156,
      "grad_norm": 0.6338880658149719,
      "learning_rate": 0.000350884602952268,
      "loss": 3.365,
      "step": 38650
    },
    {
      "epoch": 4.165321278656765,
      "grad_norm": 0.6192794442176819,
      "learning_rate": 0.0003505613619222066,
      "loss": 3.3837,
      "step": 38700
    },
    {
      "epoch": 4.1707028306963725,
      "grad_norm": 0.5782403945922852,
      "learning_rate": 0.0003502381208921452,
      "loss": 3.3893,
      "step": 38750
    },
    {
      "epoch": 4.176084382735981,
      "grad_norm": 0.5997782945632935,
      "learning_rate": 0.0003499148798620838,
      "loss": 3.3841,
      "step": 38800
    },
    {
      "epoch": 4.18146593477559,
      "grad_norm": 0.6096587181091309,
      "learning_rate": 0.0003495916388320224,
      "loss": 3.3911,
      "step": 38850
    },
    {
      "epoch": 4.186847486815197,
      "grad_norm": 0.5727220177650452,
      "learning_rate": 0.000349268397801961,
      "loss": 3.4081,
      "step": 38900
    },
    {
      "epoch": 4.192229038854806,
      "grad_norm": 0.6632119417190552,
      "learning_rate": 0.0003489451567718995,
      "loss": 3.4049,
      "step": 38950
    },
    {
      "epoch": 4.197610590894414,
      "grad_norm": 0.6273840069770813,
      "learning_rate": 0.0003486219157418381,
      "loss": 3.3895,
      "step": 39000
    },
    {
      "epoch": 4.197610590894414,
      "eval_accuracy": 0.3758614136762042,
      "eval_loss": 3.451568126678467,
      "eval_runtime": 185.0398,
      "eval_samples_per_second": 97.336,
      "eval_steps_per_second": 6.085,
      "step": 39000
    },
    {
      "epoch": 4.202992142934022,
      "grad_norm": 0.5886096358299255,
      "learning_rate": 0.00034829867471177677,
      "loss": 3.4036,
      "step": 39050
    },
    {
      "epoch": 4.208373694973631,
      "grad_norm": 0.6243075728416443,
      "learning_rate": 0.0003479754336817153,
      "loss": 3.3923,
      "step": 39100
    },
    {
      "epoch": 4.213755247013238,
      "grad_norm": 0.6715163588523865,
      "learning_rate": 0.0003476521926516539,
      "loss": 3.3949,
      "step": 39150
    },
    {
      "epoch": 4.219136799052847,
      "grad_norm": 0.5588281750679016,
      "learning_rate": 0.00034732895162159245,
      "loss": 3.3995,
      "step": 39200
    },
    {
      "epoch": 4.224518351092455,
      "grad_norm": 0.5913317203521729,
      "learning_rate": 0.00034700571059153104,
      "loss": 3.4051,
      "step": 39250
    },
    {
      "epoch": 4.229899903132063,
      "grad_norm": 0.610282301902771,
      "learning_rate": 0.00034668246956146963,
      "loss": 3.403,
      "step": 39300
    },
    {
      "epoch": 4.2352814551716715,
      "grad_norm": 0.6324890851974487,
      "learning_rate": 0.00034635922853140823,
      "loss": 3.3998,
      "step": 39350
    },
    {
      "epoch": 4.24066300721128,
      "grad_norm": 0.65585857629776,
      "learning_rate": 0.0003460359875013468,
      "loss": 3.3895,
      "step": 39400
    },
    {
      "epoch": 4.246044559250888,
      "grad_norm": 0.611068844795227,
      "learning_rate": 0.0003457127464712854,
      "loss": 3.3939,
      "step": 39450
    },
    {
      "epoch": 4.251426111290496,
      "grad_norm": 0.5995203852653503,
      "learning_rate": 0.00034538950544122396,
      "loss": 3.4055,
      "step": 39500
    },
    {
      "epoch": 4.256807663330104,
      "grad_norm": 0.5972570180892944,
      "learning_rate": 0.00034506626441116255,
      "loss": 3.4015,
      "step": 39550
    },
    {
      "epoch": 4.2621892153697125,
      "grad_norm": 0.6344018578529358,
      "learning_rate": 0.0003447430233811011,
      "loss": 3.3952,
      "step": 39600
    },
    {
      "epoch": 4.267570767409321,
      "grad_norm": 0.6562846302986145,
      "learning_rate": 0.00034441978235103974,
      "loss": 3.3818,
      "step": 39650
    },
    {
      "epoch": 4.272952319448929,
      "grad_norm": 0.6400189399719238,
      "learning_rate": 0.00034409654132097834,
      "loss": 3.3984,
      "step": 39700
    },
    {
      "epoch": 4.278333871488537,
      "grad_norm": 0.5819385647773743,
      "learning_rate": 0.0003437733002909169,
      "loss": 3.3956,
      "step": 39750
    },
    {
      "epoch": 4.283715423528146,
      "grad_norm": 0.623574435710907,
      "learning_rate": 0.00034345005926085547,
      "loss": 3.3922,
      "step": 39800
    },
    {
      "epoch": 4.2890969755677535,
      "grad_norm": 0.6206246018409729,
      "learning_rate": 0.00034312681823079407,
      "loss": 3.3921,
      "step": 39850
    },
    {
      "epoch": 4.294478527607362,
      "grad_norm": 0.699115514755249,
      "learning_rate": 0.00034280357720073266,
      "loss": 3.3933,
      "step": 39900
    },
    {
      "epoch": 4.299860079646971,
      "grad_norm": 0.6255781650543213,
      "learning_rate": 0.00034248033617067126,
      "loss": 3.3828,
      "step": 39950
    },
    {
      "epoch": 4.305241631686578,
      "grad_norm": 0.5946710109710693,
      "learning_rate": 0.00034215709514060985,
      "loss": 3.4174,
      "step": 40000
    },
    {
      "epoch": 4.305241631686578,
      "eval_accuracy": 0.37623094214766123,
      "eval_loss": 3.447413444519043,
      "eval_runtime": 185.2428,
      "eval_samples_per_second": 97.229,
      "eval_steps_per_second": 6.079,
      "step": 40000
    },
    {
      "epoch": 4.310623183726187,
      "grad_norm": 0.6533955931663513,
      "learning_rate": 0.0003418338541105484,
      "loss": 3.4007,
      "step": 40050
    },
    {
      "epoch": 4.3160047357657945,
      "grad_norm": 0.6439366340637207,
      "learning_rate": 0.000341510613080487,
      "loss": 3.4044,
      "step": 40100
    },
    {
      "epoch": 4.321386287805403,
      "grad_norm": 0.6164541244506836,
      "learning_rate": 0.0003411873720504255,
      "loss": 3.4045,
      "step": 40150
    },
    {
      "epoch": 4.326767839845012,
      "grad_norm": 0.5683732628822327,
      "learning_rate": 0.0003408641310203642,
      "loss": 3.3895,
      "step": 40200
    },
    {
      "epoch": 4.332149391884619,
      "grad_norm": 0.6738073825836182,
      "learning_rate": 0.00034054088999030277,
      "loss": 3.3922,
      "step": 40250
    },
    {
      "epoch": 4.337530943924228,
      "grad_norm": 0.6270463466644287,
      "learning_rate": 0.0003402176489602413,
      "loss": 3.3995,
      "step": 40300
    },
    {
      "epoch": 4.342912495963836,
      "grad_norm": 0.6303966641426086,
      "learning_rate": 0.0003398944079301799,
      "loss": 3.4159,
      "step": 40350
    },
    {
      "epoch": 4.348294048003444,
      "grad_norm": 0.6783425807952881,
      "learning_rate": 0.0003395711669001185,
      "loss": 3.4157,
      "step": 40400
    },
    {
      "epoch": 4.3536756000430525,
      "grad_norm": 0.6219685673713684,
      "learning_rate": 0.00033924792587005704,
      "loss": 3.3939,
      "step": 40450
    },
    {
      "epoch": 4.359057152082661,
      "grad_norm": 0.663224458694458,
      "learning_rate": 0.0003389246848399957,
      "loss": 3.3942,
      "step": 40500
    },
    {
      "epoch": 4.364438704122269,
      "grad_norm": 0.6304126381874084,
      "learning_rate": 0.0003386014438099343,
      "loss": 3.4078,
      "step": 40550
    },
    {
      "epoch": 4.369820256161877,
      "grad_norm": 0.6078771948814392,
      "learning_rate": 0.0003382846676004741,
      "loss": 3.4051,
      "step": 40600
    },
    {
      "epoch": 4.375201808201485,
      "grad_norm": 0.6536765694618225,
      "learning_rate": 0.00033796142657041263,
      "loss": 3.4022,
      "step": 40650
    },
    {
      "epoch": 4.3805833602410935,
      "grad_norm": 0.6436047554016113,
      "learning_rate": 0.0003376381855403512,
      "loss": 3.4129,
      "step": 40700
    },
    {
      "epoch": 4.385964912280702,
      "grad_norm": 0.6399409174919128,
      "learning_rate": 0.0003373149445102898,
      "loss": 3.3883,
      "step": 40750
    },
    {
      "epoch": 4.39134646432031,
      "grad_norm": 0.573988139629364,
      "learning_rate": 0.00033699170348022836,
      "loss": 3.4083,
      "step": 40800
    },
    {
      "epoch": 4.396728016359918,
      "grad_norm": 0.5928240418434143,
      "learning_rate": 0.000336668462450167,
      "loss": 3.4044,
      "step": 40850
    },
    {
      "epoch": 4.402109568399527,
      "grad_norm": 0.5938860774040222,
      "learning_rate": 0.0003363452214201056,
      "loss": 3.4019,
      "step": 40900
    },
    {
      "epoch": 4.4074911204391345,
      "grad_norm": 0.6259658336639404,
      "learning_rate": 0.00033602198039004415,
      "loss": 3.3903,
      "step": 40950
    },
    {
      "epoch": 4.412872672478743,
      "grad_norm": 0.6614382266998291,
      "learning_rate": 0.00033569873935998274,
      "loss": 3.4022,
      "step": 41000
    },
    {
      "epoch": 4.412872672478743,
      "eval_accuracy": 0.3767523673562829,
      "eval_loss": 3.4427523612976074,
      "eval_runtime": 185.3986,
      "eval_samples_per_second": 97.147,
      "eval_steps_per_second": 6.073,
      "step": 41000
    },
    {
      "epoch": 4.418254224518351,
      "grad_norm": 0.6325867772102356,
      "learning_rate": 0.0003353754983299213,
      "loss": 3.4178,
      "step": 41050
    },
    {
      "epoch": 4.423635776557959,
      "grad_norm": 0.6308853626251221,
      "learning_rate": 0.0003350522572998599,
      "loss": 3.3847,
      "step": 41100
    },
    {
      "epoch": 4.429017328597568,
      "grad_norm": 0.6155990958213806,
      "learning_rate": 0.0003347290162697985,
      "loss": 3.3941,
      "step": 41150
    },
    {
      "epoch": 4.4343988806371755,
      "grad_norm": 0.6045761704444885,
      "learning_rate": 0.00033440577523973706,
      "loss": 3.404,
      "step": 41200
    },
    {
      "epoch": 4.439780432676784,
      "grad_norm": 0.6093679666519165,
      "learning_rate": 0.00033408253420967566,
      "loss": 3.381,
      "step": 41250
    },
    {
      "epoch": 4.445161984716393,
      "grad_norm": 0.5896681547164917,
      "learning_rate": 0.00033375929317961425,
      "loss": 3.403,
      "step": 41300
    },
    {
      "epoch": 4.450543536756,
      "grad_norm": 0.6145768165588379,
      "learning_rate": 0.0003334360521495528,
      "loss": 3.3984,
      "step": 41350
    },
    {
      "epoch": 4.455925088795609,
      "grad_norm": 0.601361870765686,
      "learning_rate": 0.0003331128111194914,
      "loss": 3.3936,
      "step": 41400
    },
    {
      "epoch": 4.461306640835216,
      "grad_norm": 0.6163118481636047,
      "learning_rate": 0.00033278957008943004,
      "loss": 3.415,
      "step": 41450
    },
    {
      "epoch": 4.466688192874825,
      "grad_norm": 0.6618328094482422,
      "learning_rate": 0.0003324663290593686,
      "loss": 3.3996,
      "step": 41500
    },
    {
      "epoch": 4.4720697449144335,
      "grad_norm": 0.6380786895751953,
      "learning_rate": 0.00033214308802930717,
      "loss": 3.4126,
      "step": 41550
    },
    {
      "epoch": 4.477451296954041,
      "grad_norm": 0.622982382774353,
      "learning_rate": 0.0003318198469992457,
      "loss": 3.4209,
      "step": 41600
    },
    {
      "epoch": 4.48283284899365,
      "grad_norm": 0.6435645222663879,
      "learning_rate": 0.0003314966059691843,
      "loss": 3.3899,
      "step": 41650
    },
    {
      "epoch": 4.488214401033258,
      "grad_norm": 0.6214344501495361,
      "learning_rate": 0.00033117336493912296,
      "loss": 3.4082,
      "step": 41700
    },
    {
      "epoch": 4.493595953072866,
      "grad_norm": 0.6280515789985657,
      "learning_rate": 0.0003308501239090615,
      "loss": 3.4029,
      "step": 41750
    },
    {
      "epoch": 4.4989775051124745,
      "grad_norm": 0.6123694777488708,
      "learning_rate": 0.0003305268828790001,
      "loss": 3.4089,
      "step": 41800
    },
    {
      "epoch": 4.504359057152083,
      "grad_norm": 0.612644612789154,
      "learning_rate": 0.0003302036418489387,
      "loss": 3.4163,
      "step": 41850
    },
    {
      "epoch": 4.509740609191691,
      "grad_norm": 0.6429576873779297,
      "learning_rate": 0.0003298804008188772,
      "loss": 3.4081,
      "step": 41900
    },
    {
      "epoch": 4.515122161231299,
      "grad_norm": 0.6211088299751282,
      "learning_rate": 0.0003295571597888158,
      "loss": 3.4016,
      "step": 41950
    },
    {
      "epoch": 4.520503713270907,
      "grad_norm": 0.6325125098228455,
      "learning_rate": 0.00032923391875875447,
      "loss": 3.4035,
      "step": 42000
    },
    {
      "epoch": 4.520503713270907,
      "eval_accuracy": 0.3773413746611252,
      "eval_loss": 3.438572645187378,
      "eval_runtime": 185.3283,
      "eval_samples_per_second": 97.184,
      "eval_steps_per_second": 6.076,
      "step": 42000
    },
    {
      "epoch": 4.5258852653105155,
      "grad_norm": 0.6462650895118713,
      "learning_rate": 0.000328910677728693,
      "loss": 3.4052,
      "step": 42050
    },
    {
      "epoch": 4.531266817350124,
      "grad_norm": 0.6761447787284851,
      "learning_rate": 0.0003285874366986316,
      "loss": 3.3923,
      "step": 42100
    },
    {
      "epoch": 4.536648369389732,
      "grad_norm": 0.5771544575691223,
      "learning_rate": 0.00032826419566857015,
      "loss": 3.4027,
      "step": 42150
    },
    {
      "epoch": 4.54202992142934,
      "grad_norm": 0.6672393679618835,
      "learning_rate": 0.00032794095463850874,
      "loss": 3.409,
      "step": 42200
    },
    {
      "epoch": 4.547411473468949,
      "grad_norm": 0.66680908203125,
      "learning_rate": 0.0003276177136084473,
      "loss": 3.4134,
      "step": 42250
    },
    {
      "epoch": 4.5527930255085565,
      "grad_norm": 0.6311260461807251,
      "learning_rate": 0.00032729447257838593,
      "loss": 3.3963,
      "step": 42300
    },
    {
      "epoch": 4.558174577548165,
      "grad_norm": 0.6836269497871399,
      "learning_rate": 0.0003269712315483245,
      "loss": 3.3978,
      "step": 42350
    },
    {
      "epoch": 4.563556129587774,
      "grad_norm": 0.6160884499549866,
      "learning_rate": 0.00032664799051826306,
      "loss": 3.4117,
      "step": 42400
    },
    {
      "epoch": 4.568937681627381,
      "grad_norm": 0.6402059197425842,
      "learning_rate": 0.00032632474948820166,
      "loss": 3.3927,
      "step": 42450
    },
    {
      "epoch": 4.57431923366699,
      "grad_norm": 0.6368441581726074,
      "learning_rate": 0.00032600150845814025,
      "loss": 3.3969,
      "step": 42500
    },
    {
      "epoch": 4.579700785706597,
      "grad_norm": 0.6122279167175293,
      "learning_rate": 0.0003256782674280788,
      "loss": 3.399,
      "step": 42550
    },
    {
      "epoch": 4.585082337746206,
      "grad_norm": 0.6490753293037415,
      "learning_rate": 0.00032536149121861866,
      "loss": 3.4171,
      "step": 42600
    },
    {
      "epoch": 4.5904638897858145,
      "grad_norm": 0.6444741487503052,
      "learning_rate": 0.00032503825018855725,
      "loss": 3.4091,
      "step": 42650
    },
    {
      "epoch": 4.595845441825422,
      "grad_norm": 0.5838130116462708,
      "learning_rate": 0.00032471500915849584,
      "loss": 3.4126,
      "step": 42700
    },
    {
      "epoch": 4.601226993865031,
      "grad_norm": 0.6206106543540955,
      "learning_rate": 0.00032439176812843444,
      "loss": 3.3903,
      "step": 42750
    },
    {
      "epoch": 4.606608545904638,
      "grad_norm": 0.6526898145675659,
      "learning_rate": 0.000324068527098373,
      "loss": 3.4028,
      "step": 42800
    },
    {
      "epoch": 4.611990097944247,
      "grad_norm": 0.602979838848114,
      "learning_rate": 0.0003237452860683116,
      "loss": 3.4218,
      "step": 42850
    },
    {
      "epoch": 4.6173716499838555,
      "grad_norm": 0.625683605670929,
      "learning_rate": 0.0003234220450382501,
      "loss": 3.4032,
      "step": 42900
    },
    {
      "epoch": 4.622753202023463,
      "grad_norm": 0.6133497953414917,
      "learning_rate": 0.00032309880400818876,
      "loss": 3.41,
      "step": 42950
    },
    {
      "epoch": 4.628134754063072,
      "grad_norm": 0.6578198075294495,
      "learning_rate": 0.00032277556297812736,
      "loss": 3.4213,
      "step": 43000
    },
    {
      "epoch": 4.628134754063072,
      "eval_accuracy": 0.3779550461715014,
      "eval_loss": 3.4305732250213623,
      "eval_runtime": 185.5469,
      "eval_samples_per_second": 97.07,
      "eval_steps_per_second": 6.069,
      "step": 43000
    },
    {
      "epoch": 4.63351630610268,
      "grad_norm": 0.585598349571228,
      "learning_rate": 0.0003224523219480659,
      "loss": 3.3965,
      "step": 43050
    },
    {
      "epoch": 4.638897858142288,
      "grad_norm": 0.6515240669250488,
      "learning_rate": 0.0003221290809180045,
      "loss": 3.3968,
      "step": 43100
    },
    {
      "epoch": 4.6442794101818965,
      "grad_norm": 0.6437974572181702,
      "learning_rate": 0.0003218058398879431,
      "loss": 3.381,
      "step": 43150
    },
    {
      "epoch": 4.649660962221505,
      "grad_norm": 0.6343697905540466,
      "learning_rate": 0.00032148259885788163,
      "loss": 3.3995,
      "step": 43200
    },
    {
      "epoch": 4.655042514261113,
      "grad_norm": 0.6162006258964539,
      "learning_rate": 0.0003211593578278203,
      "loss": 3.4146,
      "step": 43250
    },
    {
      "epoch": 4.660424066300721,
      "grad_norm": 0.6167454123497009,
      "learning_rate": 0.00032083611679775887,
      "loss": 3.4109,
      "step": 43300
    },
    {
      "epoch": 4.665805618340329,
      "grad_norm": 0.6448920369148254,
      "learning_rate": 0.0003205128757676974,
      "loss": 3.407,
      "step": 43350
    },
    {
      "epoch": 4.6711871703799375,
      "grad_norm": 0.6741233468055725,
      "learning_rate": 0.000320189634737636,
      "loss": 3.3963,
      "step": 43400
    },
    {
      "epoch": 4.676568722419546,
      "grad_norm": 0.6615515947341919,
      "learning_rate": 0.00031986639370757455,
      "loss": 3.386,
      "step": 43450
    },
    {
      "epoch": 4.681950274459154,
      "grad_norm": 0.6240947246551514,
      "learning_rate": 0.0003195431526775132,
      "loss": 3.3822,
      "step": 43500
    },
    {
      "epoch": 4.687331826498762,
      "grad_norm": 0.6300475001335144,
      "learning_rate": 0.0003192199116474518,
      "loss": 3.3948,
      "step": 43550
    },
    {
      "epoch": 4.692713378538371,
      "grad_norm": 0.6011607050895691,
      "learning_rate": 0.00031889667061739033,
      "loss": 3.4019,
      "step": 43600
    },
    {
      "epoch": 4.6980949305779784,
      "grad_norm": 0.6514333486557007,
      "learning_rate": 0.0003185734295873289,
      "loss": 3.4142,
      "step": 43650
    },
    {
      "epoch": 4.703476482617587,
      "grad_norm": 0.6134124398231506,
      "learning_rate": 0.00031825018855726747,
      "loss": 3.4136,
      "step": 43700
    },
    {
      "epoch": 4.7088580346571955,
      "grad_norm": 0.63430255651474,
      "learning_rate": 0.00031792694752720606,
      "loss": 3.397,
      "step": 43750
    },
    {
      "epoch": 4.714239586696803,
      "grad_norm": 0.6157966256141663,
      "learning_rate": 0.0003176037064971447,
      "loss": 3.4264,
      "step": 43800
    },
    {
      "epoch": 4.719621138736412,
      "grad_norm": 0.6450691819190979,
      "learning_rate": 0.0003172804654670833,
      "loss": 3.4054,
      "step": 43850
    },
    {
      "epoch": 4.725002690776019,
      "grad_norm": 0.6598862409591675,
      "learning_rate": 0.00031695722443702185,
      "loss": 3.4,
      "step": 43900
    },
    {
      "epoch": 4.730384242815628,
      "grad_norm": 0.6398562788963318,
      "learning_rate": 0.00031663398340696044,
      "loss": 3.4092,
      "step": 43950
    },
    {
      "epoch": 4.7357657948552365,
      "grad_norm": 0.6426649689674377,
      "learning_rate": 0.000316310742376899,
      "loss": 3.3859,
      "step": 44000
    },
    {
      "epoch": 4.7357657948552365,
      "eval_accuracy": 0.37860435582908075,
      "eval_loss": 3.426556348800659,
      "eval_runtime": 185.3381,
      "eval_samples_per_second": 97.179,
      "eval_steps_per_second": 6.075,
      "step": 44000
    },
    {
      "epoch": 4.741147346894844,
      "grad_norm": 0.6626866459846497,
      "learning_rate": 0.0003159875013468376,
      "loss": 3.3932,
      "step": 44050
    },
    {
      "epoch": 4.746528898934453,
      "grad_norm": 0.6173332333564758,
      "learning_rate": 0.0003156642603167762,
      "loss": 3.4053,
      "step": 44100
    },
    {
      "epoch": 4.751910450974061,
      "grad_norm": 0.6142896413803101,
      "learning_rate": 0.00031534101928671476,
      "loss": 3.4,
      "step": 44150
    },
    {
      "epoch": 4.757292003013669,
      "grad_norm": 0.6264691352844238,
      "learning_rate": 0.00031501777825665336,
      "loss": 3.3941,
      "step": 44200
    },
    {
      "epoch": 4.7626735550532775,
      "grad_norm": 0.6633380055427551,
      "learning_rate": 0.0003146945372265919,
      "loss": 3.4032,
      "step": 44250
    },
    {
      "epoch": 4.768055107092886,
      "grad_norm": 0.6285715699195862,
      "learning_rate": 0.0003143712961965305,
      "loss": 3.3913,
      "step": 44300
    },
    {
      "epoch": 4.773436659132494,
      "grad_norm": 0.6682639122009277,
      "learning_rate": 0.0003140480551664691,
      "loss": 3.4215,
      "step": 44350
    },
    {
      "epoch": 4.778818211172102,
      "grad_norm": 0.6822831034660339,
      "learning_rate": 0.0003137248141364077,
      "loss": 3.3814,
      "step": 44400
    },
    {
      "epoch": 4.78419976321171,
      "grad_norm": 0.5919625163078308,
      "learning_rate": 0.0003134015731063463,
      "loss": 3.3865,
      "step": 44450
    },
    {
      "epoch": 4.7895813152513185,
      "grad_norm": 0.6576287150382996,
      "learning_rate": 0.0003130783320762849,
      "loss": 3.4026,
      "step": 44500
    },
    {
      "epoch": 4.794962867290927,
      "grad_norm": 0.6284410357475281,
      "learning_rate": 0.0003127550910462234,
      "loss": 3.4125,
      "step": 44550
    },
    {
      "epoch": 4.800344419330535,
      "grad_norm": 0.6227229833602905,
      "learning_rate": 0.000312431850016162,
      "loss": 3.3905,
      "step": 44600
    },
    {
      "epoch": 4.805725971370143,
      "grad_norm": 0.6507493257522583,
      "learning_rate": 0.00031210860898610066,
      "loss": 3.39,
      "step": 44650
    },
    {
      "epoch": 4.811107523409751,
      "grad_norm": 0.660424530506134,
      "learning_rate": 0.0003117918327766404,
      "loss": 3.391,
      "step": 44700
    },
    {
      "epoch": 4.8164890754493594,
      "grad_norm": 0.6176576614379883,
      "learning_rate": 0.00031146859174657906,
      "loss": 3.4107,
      "step": 44750
    },
    {
      "epoch": 4.821870627488968,
      "grad_norm": 0.612305223941803,
      "learning_rate": 0.0003111453507165176,
      "loss": 3.4162,
      "step": 44800
    },
    {
      "epoch": 4.827252179528576,
      "grad_norm": 0.5901872515678406,
      "learning_rate": 0.0003108221096864562,
      "loss": 3.3884,
      "step": 44850
    },
    {
      "epoch": 4.832633731568184,
      "grad_norm": 0.6249916553497314,
      "learning_rate": 0.00031049886865639473,
      "loss": 3.381,
      "step": 44900
    },
    {
      "epoch": 4.838015283607793,
      "grad_norm": 0.6434414386749268,
      "learning_rate": 0.00031017562762633333,
      "loss": 3.3978,
      "step": 44950
    },
    {
      "epoch": 4.8433968356474,
      "grad_norm": 0.634209394454956,
      "learning_rate": 0.00030985238659627187,
      "loss": 3.3887,
      "step": 45000
    },
    {
      "epoch": 4.8433968356474,
      "eval_accuracy": 0.3787655967145974,
      "eval_loss": 3.421621084213257,
      "eval_runtime": 185.0892,
      "eval_samples_per_second": 97.31,
      "eval_steps_per_second": 6.084,
      "step": 45000
    },
    {
      "epoch": 4.848778387687009,
      "grad_norm": 0.6518102884292603,
      "learning_rate": 0.0003095291455662105,
      "loss": 3.3883,
      "step": 45050
    },
    {
      "epoch": 4.8541599397266175,
      "grad_norm": 0.6235784292221069,
      "learning_rate": 0.0003092059045361491,
      "loss": 3.4005,
      "step": 45100
    },
    {
      "epoch": 4.859541491766225,
      "grad_norm": 0.6313170790672302,
      "learning_rate": 0.00030888266350608765,
      "loss": 3.4204,
      "step": 45150
    },
    {
      "epoch": 4.864923043805834,
      "grad_norm": 0.5994978547096252,
      "learning_rate": 0.00030855942247602625,
      "loss": 3.4059,
      "step": 45200
    },
    {
      "epoch": 4.870304595845441,
      "grad_norm": 0.6309564113616943,
      "learning_rate": 0.00030823618144596484,
      "loss": 3.4227,
      "step": 45250
    },
    {
      "epoch": 4.87568614788505,
      "grad_norm": 0.6527310013771057,
      "learning_rate": 0.0003079129404159035,
      "loss": 3.3912,
      "step": 45300
    },
    {
      "epoch": 4.8810676999246585,
      "grad_norm": 0.6751705408096313,
      "learning_rate": 0.00030758969938584203,
      "loss": 3.3909,
      "step": 45350
    },
    {
      "epoch": 4.886449251964266,
      "grad_norm": 0.6809637546539307,
      "learning_rate": 0.0003072664583557806,
      "loss": 3.3995,
      "step": 45400
    },
    {
      "epoch": 4.891830804003875,
      "grad_norm": 0.677666187286377,
      "learning_rate": 0.00030694321732571917,
      "loss": 3.3895,
      "step": 45450
    },
    {
      "epoch": 4.897212356043483,
      "grad_norm": 0.6385020017623901,
      "learning_rate": 0.00030661997629565776,
      "loss": 3.3938,
      "step": 45500
    },
    {
      "epoch": 4.902593908083091,
      "grad_norm": 0.6369286179542542,
      "learning_rate": 0.0003062967352655963,
      "loss": 3.4226,
      "step": 45550
    },
    {
      "epoch": 4.9079754601226995,
      "grad_norm": 0.6255587935447693,
      "learning_rate": 0.00030597349423553495,
      "loss": 3.4069,
      "step": 45600
    },
    {
      "epoch": 4.913357012162308,
      "grad_norm": 0.6073203682899475,
      "learning_rate": 0.00030565025320547355,
      "loss": 3.3941,
      "step": 45650
    },
    {
      "epoch": 4.918738564201916,
      "grad_norm": 0.6055741310119629,
      "learning_rate": 0.0003053270121754121,
      "loss": 3.3977,
      "step": 45700
    },
    {
      "epoch": 4.924120116241524,
      "grad_norm": 0.6943128108978271,
      "learning_rate": 0.0003050037711453507,
      "loss": 3.3949,
      "step": 45750
    },
    {
      "epoch": 4.929501668281132,
      "grad_norm": 0.6536029577255249,
      "learning_rate": 0.0003046805301152893,
      "loss": 3.3976,
      "step": 45800
    },
    {
      "epoch": 4.9348832203207404,
      "grad_norm": 0.6314778327941895,
      "learning_rate": 0.0003043572890852278,
      "loss": 3.4104,
      "step": 45850
    },
    {
      "epoch": 4.940264772360349,
      "grad_norm": 0.6491816639900208,
      "learning_rate": 0.00030403404805516646,
      "loss": 3.4021,
      "step": 45900
    },
    {
      "epoch": 4.945646324399957,
      "grad_norm": 0.639607310295105,
      "learning_rate": 0.00030371080702510506,
      "loss": 3.4021,
      "step": 45950
    },
    {
      "epoch": 4.951027876439565,
      "grad_norm": 0.6291335225105286,
      "learning_rate": 0.0003033875659950436,
      "loss": 3.4066,
      "step": 46000
    },
    {
      "epoch": 4.951027876439565,
      "eval_accuracy": 0.37932983116101804,
      "eval_loss": 3.416398763656616,
      "eval_runtime": 185.3345,
      "eval_samples_per_second": 97.181,
      "eval_steps_per_second": 6.076,
      "step": 46000
    },
    {
      "epoch": 4.956409428479174,
      "grad_norm": 0.6686611771583557,
      "learning_rate": 0.0003030643249649822,
      "loss": 3.4019,
      "step": 46050
    },
    {
      "epoch": 4.961790980518781,
      "grad_norm": 0.6632118821144104,
      "learning_rate": 0.00030274108393492073,
      "loss": 3.405,
      "step": 46100
    },
    {
      "epoch": 4.96717253255839,
      "grad_norm": 0.6167105436325073,
      "learning_rate": 0.00030241784290485933,
      "loss": 3.3986,
      "step": 46150
    },
    {
      "epoch": 4.9725540845979985,
      "grad_norm": 0.6737709045410156,
      "learning_rate": 0.000302094601874798,
      "loss": 3.4027,
      "step": 46200
    },
    {
      "epoch": 4.977935636637606,
      "grad_norm": 0.650943398475647,
      "learning_rate": 0.0003017713608447365,
      "loss": 3.3746,
      "step": 46250
    },
    {
      "epoch": 4.983317188677215,
      "grad_norm": 0.6425312161445618,
      "learning_rate": 0.0003014481198146751,
      "loss": 3.4089,
      "step": 46300
    },
    {
      "epoch": 4.988698740716822,
      "grad_norm": 0.6737054586410522,
      "learning_rate": 0.0003011248787846137,
      "loss": 3.4108,
      "step": 46350
    },
    {
      "epoch": 4.994080292756431,
      "grad_norm": 0.5824667811393738,
      "learning_rate": 0.00030080163775455225,
      "loss": 3.4089,
      "step": 46400
    },
    {
      "epoch": 4.9994618447960395,
      "grad_norm": 0.6069629788398743,
      "learning_rate": 0.0003004783967244909,
      "loss": 3.3961,
      "step": 46450
    },
    {
      "epoch": 5.004843396835647,
      "grad_norm": 0.6461926102638245,
      "learning_rate": 0.0003001551556944295,
      "loss": 3.3149,
      "step": 46500
    },
    {
      "epoch": 5.010224948875256,
      "grad_norm": 0.6721562147140503,
      "learning_rate": 0.00029983191466436803,
      "loss": 3.324,
      "step": 46550
    },
    {
      "epoch": 5.015606500914864,
      "grad_norm": 0.6682996153831482,
      "learning_rate": 0.0002995086736343066,
      "loss": 3.3002,
      "step": 46600
    },
    {
      "epoch": 5.020988052954472,
      "grad_norm": 0.6803022027015686,
      "learning_rate": 0.00029918543260424517,
      "loss": 3.3082,
      "step": 46650
    },
    {
      "epoch": 5.0263696049940805,
      "grad_norm": 0.6562891006469727,
      "learning_rate": 0.0002988621915741838,
      "loss": 3.3043,
      "step": 46700
    },
    {
      "epoch": 5.031751157033688,
      "grad_norm": 0.6049568057060242,
      "learning_rate": 0.00029853895054412236,
      "loss": 3.3242,
      "step": 46750
    },
    {
      "epoch": 5.037132709073297,
      "grad_norm": 0.7194983959197998,
      "learning_rate": 0.00029822217433466216,
      "loss": 3.3293,
      "step": 46800
    },
    {
      "epoch": 5.042514261112905,
      "grad_norm": 0.6507198810577393,
      "learning_rate": 0.00029789893330460076,
      "loss": 3.3187,
      "step": 46850
    },
    {
      "epoch": 5.047895813152513,
      "grad_norm": 0.6023919582366943,
      "learning_rate": 0.00029757569227453935,
      "loss": 3.2911,
      "step": 46900
    },
    {
      "epoch": 5.0532773651921215,
      "grad_norm": 0.6877101063728333,
      "learning_rate": 0.00029725245124447795,
      "loss": 3.3207,
      "step": 46950
    },
    {
      "epoch": 5.05865891723173,
      "grad_norm": 0.614548921585083,
      "learning_rate": 0.0002969292102144165,
      "loss": 3.3124,
      "step": 47000
    },
    {
      "epoch": 5.05865891723173,
      "eval_accuracy": 0.3794301177764708,
      "eval_loss": 3.4212470054626465,
      "eval_runtime": 185.1572,
      "eval_samples_per_second": 97.274,
      "eval_steps_per_second": 6.081,
      "step": 47000
    },
    {
      "epoch": 5.064040469271338,
      "grad_norm": 0.632571816444397,
      "learning_rate": 0.00029660596918435514,
      "loss": 3.3087,
      "step": 47050
    },
    {
      "epoch": 5.069422021310946,
      "grad_norm": 0.6783804297447205,
      "learning_rate": 0.0002962827281542937,
      "loss": 3.3279,
      "step": 47100
    },
    {
      "epoch": 5.074803573350554,
      "grad_norm": 0.613727331161499,
      "learning_rate": 0.00029595948712423227,
      "loss": 3.3147,
      "step": 47150
    },
    {
      "epoch": 5.080185125390162,
      "grad_norm": 0.6511032581329346,
      "learning_rate": 0.00029563624609417087,
      "loss": 3.3327,
      "step": 47200
    },
    {
      "epoch": 5.085566677429771,
      "grad_norm": 0.6946690678596497,
      "learning_rate": 0.00029531300506410946,
      "loss": 3.3201,
      "step": 47250
    },
    {
      "epoch": 5.090948229469379,
      "grad_norm": 0.6138713359832764,
      "learning_rate": 0.000294989764034048,
      "loss": 3.3152,
      "step": 47300
    },
    {
      "epoch": 5.096329781508987,
      "grad_norm": 0.6477729082107544,
      "learning_rate": 0.0002946665230039866,
      "loss": 3.3439,
      "step": 47350
    },
    {
      "epoch": 5.101711333548596,
      "grad_norm": 0.6678525805473328,
      "learning_rate": 0.0002943432819739252,
      "loss": 3.3308,
      "step": 47400
    },
    {
      "epoch": 5.107092885588203,
      "grad_norm": 0.648914098739624,
      "learning_rate": 0.0002940200409438638,
      "loss": 3.33,
      "step": 47450
    },
    {
      "epoch": 5.112474437627812,
      "grad_norm": 0.6616610884666443,
      "learning_rate": 0.0002936967999138024,
      "loss": 3.3208,
      "step": 47500
    },
    {
      "epoch": 5.1178559896674205,
      "grad_norm": 0.688664972782135,
      "learning_rate": 0.0002933735588837409,
      "loss": 3.3241,
      "step": 47550
    },
    {
      "epoch": 5.123237541707028,
      "grad_norm": 0.645365297794342,
      "learning_rate": 0.00029305031785367957,
      "loss": 3.3187,
      "step": 47600
    },
    {
      "epoch": 5.128619093746637,
      "grad_norm": 0.6506872773170471,
      "learning_rate": 0.0002927270768236181,
      "loss": 3.3471,
      "step": 47650
    },
    {
      "epoch": 5.134000645786244,
      "grad_norm": 0.6527488231658936,
      "learning_rate": 0.0002924038357935567,
      "loss": 3.3469,
      "step": 47700
    },
    {
      "epoch": 5.139382197825853,
      "grad_norm": 0.6231043934822083,
      "learning_rate": 0.0002920805947634953,
      "loss": 3.3163,
      "step": 47750
    },
    {
      "epoch": 5.1447637498654615,
      "grad_norm": 0.6428726315498352,
      "learning_rate": 0.0002917573537334339,
      "loss": 3.3221,
      "step": 47800
    },
    {
      "epoch": 5.150145301905069,
      "grad_norm": 0.6974531412124634,
      "learning_rate": 0.00029143411270337243,
      "loss": 3.3213,
      "step": 47850
    },
    {
      "epoch": 5.155526853944678,
      "grad_norm": 0.6757194399833679,
      "learning_rate": 0.00029111087167331103,
      "loss": 3.335,
      "step": 47900
    },
    {
      "epoch": 5.160908405984286,
      "grad_norm": 0.6837787628173828,
      "learning_rate": 0.0002907876306432496,
      "loss": 3.3368,
      "step": 47950
    },
    {
      "epoch": 5.166289958023894,
      "grad_norm": 0.6229298710823059,
      "learning_rate": 0.0002904643896131882,
      "loss": 3.3621,
      "step": 48000
    },
    {
      "epoch": 5.166289958023894,
      "eval_accuracy": 0.3799755552732995,
      "eval_loss": 3.418832540512085,
      "eval_runtime": 185.4586,
      "eval_samples_per_second": 97.116,
      "eval_steps_per_second": 6.071,
      "step": 48000
    },
    {
      "epoch": 5.1716715100635025,
      "grad_norm": 0.64594966173172,
      "learning_rate": 0.0002901411485831268,
      "loss": 3.3507,
      "step": 48050
    },
    {
      "epoch": 5.17705306210311,
      "grad_norm": 0.7221371531486511,
      "learning_rate": 0.00028981790755306535,
      "loss": 3.3299,
      "step": 48100
    },
    {
      "epoch": 5.182434614142719,
      "grad_norm": 0.6576762199401855,
      "learning_rate": 0.00028949466652300395,
      "loss": 3.3191,
      "step": 48150
    },
    {
      "epoch": 5.187816166182327,
      "grad_norm": 0.6678467988967896,
      "learning_rate": 0.00028917142549294254,
      "loss": 3.3366,
      "step": 48200
    },
    {
      "epoch": 5.193197718221935,
      "grad_norm": 0.6873053908348083,
      "learning_rate": 0.00028884818446288114,
      "loss": 3.3336,
      "step": 48250
    },
    {
      "epoch": 5.198579270261543,
      "grad_norm": 0.6549661159515381,
      "learning_rate": 0.0002885249434328197,
      "loss": 3.3324,
      "step": 48300
    },
    {
      "epoch": 5.203960822301152,
      "grad_norm": 0.6529666185379028,
      "learning_rate": 0.0002882017024027583,
      "loss": 3.3318,
      "step": 48350
    },
    {
      "epoch": 5.20934237434076,
      "grad_norm": 0.6662092804908752,
      "learning_rate": 0.00028787846137269687,
      "loss": 3.3325,
      "step": 48400
    },
    {
      "epoch": 5.214723926380368,
      "grad_norm": 0.6745813488960266,
      "learning_rate": 0.00028755522034263546,
      "loss": 3.3409,
      "step": 48450
    },
    {
      "epoch": 5.220105478419977,
      "grad_norm": 0.6083055734634399,
      "learning_rate": 0.00028723197931257406,
      "loss": 3.3308,
      "step": 48500
    },
    {
      "epoch": 5.225487030459584,
      "grad_norm": 0.6546552181243896,
      "learning_rate": 0.00028690873828251265,
      "loss": 3.327,
      "step": 48550
    },
    {
      "epoch": 5.230868582499193,
      "grad_norm": 0.684239387512207,
      "learning_rate": 0.00028658549725245125,
      "loss": 3.3514,
      "step": 48600
    },
    {
      "epoch": 5.236250134538801,
      "grad_norm": 0.648367166519165,
      "learning_rate": 0.0002862622562223898,
      "loss": 3.3401,
      "step": 48650
    },
    {
      "epoch": 5.241631686578409,
      "grad_norm": 0.688685417175293,
      "learning_rate": 0.0002859390151923284,
      "loss": 3.3243,
      "step": 48700
    },
    {
      "epoch": 5.247013238618018,
      "grad_norm": 0.6698136925697327,
      "learning_rate": 0.000285615774162267,
      "loss": 3.3273,
      "step": 48750
    },
    {
      "epoch": 5.252394790657625,
      "grad_norm": 0.6517601609230042,
      "learning_rate": 0.00028529253313220557,
      "loss": 3.3498,
      "step": 48800
    },
    {
      "epoch": 5.257776342697234,
      "grad_norm": 0.6846751570701599,
      "learning_rate": 0.0002849692921021441,
      "loss": 3.3413,
      "step": 48850
    },
    {
      "epoch": 5.2631578947368425,
      "grad_norm": 0.6320788264274597,
      "learning_rate": 0.00028464605107208276,
      "loss": 3.3401,
      "step": 48900
    },
    {
      "epoch": 5.26853944677645,
      "grad_norm": 0.6687496900558472,
      "learning_rate": 0.0002843228100420213,
      "loss": 3.3471,
      "step": 48950
    },
    {
      "epoch": 5.273920998816059,
      "grad_norm": 0.6405502557754517,
      "learning_rate": 0.0002839995690119599,
      "loss": 3.3335,
      "step": 49000
    },
    {
      "epoch": 5.273920998816059,
      "eval_accuracy": 0.3801706958598143,
      "eval_loss": 3.4140377044677734,
      "eval_runtime": 185.2146,
      "eval_samples_per_second": 97.244,
      "eval_steps_per_second": 6.079,
      "step": 49000
    },
    {
      "epoch": 5.279302550855666,
      "grad_norm": 0.653680145740509,
      "learning_rate": 0.0002836763279818985,
      "loss": 3.3264,
      "step": 49050
    },
    {
      "epoch": 5.284684102895275,
      "grad_norm": 0.724675178527832,
      "learning_rate": 0.0002833595517724383,
      "loss": 3.3418,
      "step": 49100
    },
    {
      "epoch": 5.2900656549348835,
      "grad_norm": 0.6484974026679993,
      "learning_rate": 0.0002830363107423769,
      "loss": 3.3476,
      "step": 49150
    },
    {
      "epoch": 5.295447206974491,
      "grad_norm": 0.688917338848114,
      "learning_rate": 0.00028271306971231543,
      "loss": 3.3359,
      "step": 49200
    },
    {
      "epoch": 5.3008287590141,
      "grad_norm": 0.6998127102851868,
      "learning_rate": 0.000282389828682254,
      "loss": 3.3372,
      "step": 49250
    },
    {
      "epoch": 5.306210311053708,
      "grad_norm": 0.6400441527366638,
      "learning_rate": 0.0002820665876521926,
      "loss": 3.3493,
      "step": 49300
    },
    {
      "epoch": 5.311591863093316,
      "grad_norm": 0.6734283566474915,
      "learning_rate": 0.0002817433466221312,
      "loss": 3.3457,
      "step": 49350
    },
    {
      "epoch": 5.316973415132924,
      "grad_norm": 0.6428185701370239,
      "learning_rate": 0.0002814201055920698,
      "loss": 3.3362,
      "step": 49400
    },
    {
      "epoch": 5.322354967172533,
      "grad_norm": 0.6745774149894714,
      "learning_rate": 0.0002810968645620084,
      "loss": 3.3338,
      "step": 49450
    },
    {
      "epoch": 5.327736519212141,
      "grad_norm": 0.6905531883239746,
      "learning_rate": 0.00028077362353194694,
      "loss": 3.3263,
      "step": 49500
    },
    {
      "epoch": 5.333118071251749,
      "grad_norm": 0.6836384534835815,
      "learning_rate": 0.00028045038250188554,
      "loss": 3.345,
      "step": 49550
    },
    {
      "epoch": 5.338499623291357,
      "grad_norm": 0.6578437089920044,
      "learning_rate": 0.00028012714147182413,
      "loss": 3.3499,
      "step": 49600
    },
    {
      "epoch": 5.343881175330965,
      "grad_norm": 0.6770058274269104,
      "learning_rate": 0.00027980390044176273,
      "loss": 3.3272,
      "step": 49650
    },
    {
      "epoch": 5.349262727370574,
      "grad_norm": 0.631992757320404,
      "learning_rate": 0.0002794806594117013,
      "loss": 3.3182,
      "step": 49700
    },
    {
      "epoch": 5.354644279410182,
      "grad_norm": 0.6202480792999268,
      "learning_rate": 0.00027915741838163986,
      "loss": 3.3483,
      "step": 49750
    },
    {
      "epoch": 5.36002583144979,
      "grad_norm": 0.6883432865142822,
      "learning_rate": 0.00027883417735157846,
      "loss": 3.3323,
      "step": 49800
    },
    {
      "epoch": 5.365407383489399,
      "grad_norm": 0.6618931889533997,
      "learning_rate": 0.00027851093632151705,
      "loss": 3.3424,
      "step": 49850
    },
    {
      "epoch": 5.370788935529006,
      "grad_norm": 0.6649948954582214,
      "learning_rate": 0.00027818769529145565,
      "loss": 3.3525,
      "step": 49900
    },
    {
      "epoch": 5.376170487568615,
      "grad_norm": 0.7178868055343628,
      "learning_rate": 0.0002778644542613942,
      "loss": 3.3406,
      "step": 49950
    },
    {
      "epoch": 5.3815520396082235,
      "grad_norm": 0.6427643895149231,
      "learning_rate": 0.00027754121323133284,
      "loss": 3.3529,
      "step": 50000
    },
    {
      "epoch": 5.3815520396082235,
      "eval_accuracy": 0.3805308801829193,
      "eval_loss": 3.4105281829833984,
      "eval_runtime": 185.2582,
      "eval_samples_per_second": 97.221,
      "eval_steps_per_second": 6.078,
      "step": 50000
    },
    {
      "epoch": 5.386933591647831,
      "grad_norm": 0.7206417322158813,
      "learning_rate": 0.0002772179722012714,
      "loss": 3.3407,
      "step": 50050
    },
    {
      "epoch": 5.39231514368744,
      "grad_norm": 0.6234815716743469,
      "learning_rate": 0.00027689473117120997,
      "loss": 3.3508,
      "step": 50100
    },
    {
      "epoch": 5.397696695727047,
      "grad_norm": 0.6713009476661682,
      "learning_rate": 0.00027657149014114857,
      "loss": 3.3503,
      "step": 50150
    },
    {
      "epoch": 5.403078247766656,
      "grad_norm": 0.6315914988517761,
      "learning_rate": 0.00027624824911108716,
      "loss": 3.3342,
      "step": 50200
    },
    {
      "epoch": 5.4084597998062645,
      "grad_norm": 0.6887965798377991,
      "learning_rate": 0.0002759250080810257,
      "loss": 3.3488,
      "step": 50250
    },
    {
      "epoch": 5.413841351845872,
      "grad_norm": 0.6944553256034851,
      "learning_rate": 0.0002756017670509643,
      "loss": 3.3473,
      "step": 50300
    },
    {
      "epoch": 5.419222903885481,
      "grad_norm": 0.6763197183609009,
      "learning_rate": 0.0002752785260209029,
      "loss": 3.3387,
      "step": 50350
    },
    {
      "epoch": 5.424604455925088,
      "grad_norm": 0.631121814250946,
      "learning_rate": 0.0002749552849908415,
      "loss": 3.3409,
      "step": 50400
    },
    {
      "epoch": 5.429986007964697,
      "grad_norm": 0.6401315331459045,
      "learning_rate": 0.0002746320439607801,
      "loss": 3.3238,
      "step": 50450
    },
    {
      "epoch": 5.435367560004305,
      "grad_norm": 0.6535040736198425,
      "learning_rate": 0.0002743088029307186,
      "loss": 3.337,
      "step": 50500
    },
    {
      "epoch": 5.440749112043913,
      "grad_norm": 0.648989200592041,
      "learning_rate": 0.00027398556190065727,
      "loss": 3.3271,
      "step": 50550
    },
    {
      "epoch": 5.446130664083522,
      "grad_norm": 0.6286004781723022,
      "learning_rate": 0.0002736623208705958,
      "loss": 3.3497,
      "step": 50600
    },
    {
      "epoch": 5.45151221612313,
      "grad_norm": 0.6893504858016968,
      "learning_rate": 0.0002733390798405344,
      "loss": 3.3576,
      "step": 50650
    },
    {
      "epoch": 5.456893768162738,
      "grad_norm": 0.6353843808174133,
      "learning_rate": 0.000273015838810473,
      "loss": 3.3488,
      "step": 50700
    },
    {
      "epoch": 5.462275320202346,
      "grad_norm": 0.7283032536506653,
      "learning_rate": 0.0002726925977804116,
      "loss": 3.3427,
      "step": 50750
    },
    {
      "epoch": 5.467656872241955,
      "grad_norm": 0.7136728167533875,
      "learning_rate": 0.00027236935675035013,
      "loss": 3.3393,
      "step": 50800
    },
    {
      "epoch": 5.473038424281563,
      "grad_norm": 0.6591393947601318,
      "learning_rate": 0.00027204611572028873,
      "loss": 3.3538,
      "step": 50850
    },
    {
      "epoch": 5.478419976321171,
      "grad_norm": 0.6900110840797424,
      "learning_rate": 0.0002717228746902273,
      "loss": 3.3338,
      "step": 50900
    },
    {
      "epoch": 5.483801528360779,
      "grad_norm": 0.6398887038230896,
      "learning_rate": 0.0002713996336601659,
      "loss": 3.35,
      "step": 50950
    },
    {
      "epoch": 5.489183080400387,
      "grad_norm": 0.6916953325271606,
      "learning_rate": 0.0002710763926301045,
      "loss": 3.3445,
      "step": 51000
    },
    {
      "epoch": 5.489183080400387,
      "eval_accuracy": 0.38107001581225475,
      "eval_loss": 3.4076125621795654,
      "eval_runtime": 185.5513,
      "eval_samples_per_second": 97.068,
      "eval_steps_per_second": 6.068,
      "step": 51000
    },
    {
      "epoch": 5.494564632439996,
      "grad_norm": 0.6774631142616272,
      "learning_rate": 0.00027075315160004305,
      "loss": 3.3555,
      "step": 51050
    },
    {
      "epoch": 5.499946184479604,
      "grad_norm": 0.651495635509491,
      "learning_rate": 0.0002704363753905829,
      "loss": 3.3554,
      "step": 51100
    },
    {
      "epoch": 5.505327736519212,
      "grad_norm": 0.6358013153076172,
      "learning_rate": 0.00027011313436052145,
      "loss": 3.3412,
      "step": 51150
    },
    {
      "epoch": 5.510709288558821,
      "grad_norm": 0.6620592474937439,
      "learning_rate": 0.00026978989333046005,
      "loss": 3.3546,
      "step": 51200
    },
    {
      "epoch": 5.516090840598428,
      "grad_norm": 0.6456208229064941,
      "learning_rate": 0.00026946665230039864,
      "loss": 3.3238,
      "step": 51250
    },
    {
      "epoch": 5.521472392638037,
      "grad_norm": 0.7263305187225342,
      "learning_rate": 0.00026914341127033724,
      "loss": 3.352,
      "step": 51300
    },
    {
      "epoch": 5.5268539446776455,
      "grad_norm": 0.6708289980888367,
      "learning_rate": 0.00026882017024027583,
      "loss": 3.3555,
      "step": 51350
    },
    {
      "epoch": 5.532235496717253,
      "grad_norm": 0.6622492074966431,
      "learning_rate": 0.0002684969292102144,
      "loss": 3.3594,
      "step": 51400
    },
    {
      "epoch": 5.537617048756862,
      "grad_norm": 0.7074928283691406,
      "learning_rate": 0.00026817368818015297,
      "loss": 3.352,
      "step": 51450
    },
    {
      "epoch": 5.542998600796469,
      "grad_norm": 0.6487237215042114,
      "learning_rate": 0.00026785044715009156,
      "loss": 3.3599,
      "step": 51500
    },
    {
      "epoch": 5.548380152836078,
      "grad_norm": 0.6759780645370483,
      "learning_rate": 0.00026752720612003016,
      "loss": 3.3461,
      "step": 51550
    },
    {
      "epoch": 5.553761704875686,
      "grad_norm": 0.6818773150444031,
      "learning_rate": 0.0002672039650899687,
      "loss": 3.3629,
      "step": 51600
    },
    {
      "epoch": 5.559143256915294,
      "grad_norm": 0.6837232112884521,
      "learning_rate": 0.00026688072405990735,
      "loss": 3.3529,
      "step": 51650
    },
    {
      "epoch": 5.564524808954903,
      "grad_norm": 0.687944769859314,
      "learning_rate": 0.0002665574830298459,
      "loss": 3.3491,
      "step": 51700
    },
    {
      "epoch": 5.569906360994511,
      "grad_norm": 0.7152137160301208,
      "learning_rate": 0.0002662342419997845,
      "loss": 3.3496,
      "step": 51750
    },
    {
      "epoch": 5.575287913034119,
      "grad_norm": 0.640335738658905,
      "learning_rate": 0.0002659110009697231,
      "loss": 3.3444,
      "step": 51800
    },
    {
      "epoch": 5.580669465073727,
      "grad_norm": 0.71235191822052,
      "learning_rate": 0.00026558775993966167,
      "loss": 3.3316,
      "step": 51850
    },
    {
      "epoch": 5.586051017113336,
      "grad_norm": 0.6604979634284973,
      "learning_rate": 0.0002652645189096002,
      "loss": 3.3539,
      "step": 51900
    },
    {
      "epoch": 5.591432569152944,
      "grad_norm": 0.7094579935073853,
      "learning_rate": 0.0002649412778795388,
      "loss": 3.3608,
      "step": 51950
    },
    {
      "epoch": 5.596814121192552,
      "grad_norm": 0.7073338031768799,
      "learning_rate": 0.0002646180368494774,
      "loss": 3.3472,
      "step": 52000
    },
    {
      "epoch": 5.596814121192552,
      "eval_accuracy": 0.381449974960942,
      "eval_loss": 3.4006683826446533,
      "eval_runtime": 185.0029,
      "eval_samples_per_second": 97.355,
      "eval_steps_per_second": 6.086,
      "step": 52000
    },
    {
      "epoch": 5.60219567323216,
      "grad_norm": 1.0897701978683472,
      "learning_rate": 0.000264294795819416,
      "loss": 3.3612,
      "step": 52050
    },
    {
      "epoch": 5.607577225271768,
      "grad_norm": 0.6901730298995972,
      "learning_rate": 0.0002639715547893546,
      "loss": 3.3614,
      "step": 52100
    },
    {
      "epoch": 5.612958777311377,
      "grad_norm": 0.6666170358657837,
      "learning_rate": 0.00026364831375929313,
      "loss": 3.3625,
      "step": 52150
    },
    {
      "epoch": 5.618340329350985,
      "grad_norm": 0.6621705293655396,
      "learning_rate": 0.0002633250727292318,
      "loss": 3.3553,
      "step": 52200
    },
    {
      "epoch": 5.623721881390593,
      "grad_norm": 0.6942315697669983,
      "learning_rate": 0.0002630018316991703,
      "loss": 3.3546,
      "step": 52250
    },
    {
      "epoch": 5.629103433430201,
      "grad_norm": 0.6472527384757996,
      "learning_rate": 0.0002626785906691089,
      "loss": 3.3379,
      "step": 52300
    },
    {
      "epoch": 5.634484985469809,
      "grad_norm": 0.6996361017227173,
      "learning_rate": 0.0002623553496390475,
      "loss": 3.3448,
      "step": 52350
    },
    {
      "epoch": 5.639866537509418,
      "grad_norm": 0.6833305954933167,
      "learning_rate": 0.0002620321086089861,
      "loss": 3.3521,
      "step": 52400
    },
    {
      "epoch": 5.645248089549026,
      "grad_norm": 0.6437026262283325,
      "learning_rate": 0.00026170886757892464,
      "loss": 3.3554,
      "step": 52450
    },
    {
      "epoch": 5.650629641588634,
      "grad_norm": 0.7210367918014526,
      "learning_rate": 0.00026138562654886324,
      "loss": 3.3436,
      "step": 52500
    },
    {
      "epoch": 5.656011193628243,
      "grad_norm": 0.647635281085968,
      "learning_rate": 0.00026106238551880183,
      "loss": 3.3564,
      "step": 52550
    },
    {
      "epoch": 5.66139274566785,
      "grad_norm": 0.6487835049629211,
      "learning_rate": 0.00026073914448874043,
      "loss": 3.3402,
      "step": 52600
    },
    {
      "epoch": 5.666774297707459,
      "grad_norm": 2.1879024505615234,
      "learning_rate": 0.000260415903458679,
      "loss": 3.3497,
      "step": 52650
    },
    {
      "epoch": 5.672155849747067,
      "grad_norm": 0.6743273735046387,
      "learning_rate": 0.00026009266242861756,
      "loss": 3.3691,
      "step": 52700
    },
    {
      "epoch": 5.677537401786675,
      "grad_norm": 0.6630558371543884,
      "learning_rate": 0.00025976942139855616,
      "loss": 3.3534,
      "step": 52750
    },
    {
      "epoch": 5.682918953826284,
      "grad_norm": 0.641095757484436,
      "learning_rate": 0.00025944618036849475,
      "loss": 3.3688,
      "step": 52800
    },
    {
      "epoch": 5.688300505865891,
      "grad_norm": 0.6753203868865967,
      "learning_rate": 0.00025912293933843335,
      "loss": 3.3498,
      "step": 52850
    },
    {
      "epoch": 5.6936820579055,
      "grad_norm": 0.6644752621650696,
      "learning_rate": 0.0002587996983083719,
      "loss": 3.3341,
      "step": 52900
    },
    {
      "epoch": 5.699063609945108,
      "grad_norm": 0.6631327867507935,
      "learning_rate": 0.00025847645727831054,
      "loss": 3.3551,
      "step": 52950
    },
    {
      "epoch": 5.704445161984716,
      "grad_norm": 0.6692296862602234,
      "learning_rate": 0.0002581532162482491,
      "loss": 3.3447,
      "step": 53000
    },
    {
      "epoch": 5.704445161984716,
      "eval_accuracy": 0.38218175216037253,
      "eval_loss": 3.396669864654541,
      "eval_runtime": 185.3184,
      "eval_samples_per_second": 97.189,
      "eval_steps_per_second": 6.076,
      "step": 53000
    },
    {
      "epoch": 5.709826714024325,
      "grad_norm": 0.7013212442398071,
      "learning_rate": 0.00025782997521818767,
      "loss": 3.3419,
      "step": 53050
    },
    {
      "epoch": 5.715208266063933,
      "grad_norm": 0.7109099626541138,
      "learning_rate": 0.00025750673418812627,
      "loss": 3.3527,
      "step": 53100
    },
    {
      "epoch": 5.720589818103541,
      "grad_norm": 0.6711696982383728,
      "learning_rate": 0.00025718349315806486,
      "loss": 3.345,
      "step": 53150
    },
    {
      "epoch": 5.725971370143149,
      "grad_norm": 0.7037459015846252,
      "learning_rate": 0.00025686025212800346,
      "loss": 3.3283,
      "step": 53200
    },
    {
      "epoch": 5.731352922182758,
      "grad_norm": 0.6624120473861694,
      "learning_rate": 0.000256537011097942,
      "loss": 3.353,
      "step": 53250
    },
    {
      "epoch": 5.736734474222366,
      "grad_norm": 0.6593768000602722,
      "learning_rate": 0.0002562137700678806,
      "loss": 3.3437,
      "step": 53300
    },
    {
      "epoch": 5.742116026261974,
      "grad_norm": 0.7452617287635803,
      "learning_rate": 0.0002558905290378192,
      "loss": 3.3512,
      "step": 53350
    },
    {
      "epoch": 5.747497578301582,
      "grad_norm": 0.7954784035682678,
      "learning_rate": 0.0002555672880077578,
      "loss": 3.348,
      "step": 53400
    },
    {
      "epoch": 5.75287913034119,
      "grad_norm": 0.6386750936508179,
      "learning_rate": 0.0002552440469776963,
      "loss": 3.3461,
      "step": 53450
    },
    {
      "epoch": 5.758260682380799,
      "grad_norm": 0.6605433821678162,
      "learning_rate": 0.00025492080594763497,
      "loss": 3.3528,
      "step": 53500
    },
    {
      "epoch": 5.763642234420407,
      "grad_norm": 0.7153132557868958,
      "learning_rate": 0.0002545975649175735,
      "loss": 3.3327,
      "step": 53550
    },
    {
      "epoch": 5.769023786460015,
      "grad_norm": 0.6868748068809509,
      "learning_rate": 0.0002542743238875121,
      "loss": 3.3354,
      "step": 53600
    },
    {
      "epoch": 5.774405338499624,
      "grad_norm": 0.7212463617324829,
      "learning_rate": 0.0002539510828574507,
      "loss": 3.3429,
      "step": 53650
    },
    {
      "epoch": 5.779786890539231,
      "grad_norm": 0.6269412040710449,
      "learning_rate": 0.00025362784182738924,
      "loss": 3.3577,
      "step": 53700
    },
    {
      "epoch": 5.78516844257884,
      "grad_norm": 0.7045172452926636,
      "learning_rate": 0.00025330460079732783,
      "loss": 3.3459,
      "step": 53750
    },
    {
      "epoch": 5.790549994618448,
      "grad_norm": 0.7062743902206421,
      "learning_rate": 0.00025298135976726643,
      "loss": 3.3327,
      "step": 53800
    },
    {
      "epoch": 5.795931546658056,
      "grad_norm": 0.6376510858535767,
      "learning_rate": 0.000252658118737205,
      "loss": 3.3316,
      "step": 53850
    },
    {
      "epoch": 5.801313098697665,
      "grad_norm": 0.676217257976532,
      "learning_rate": 0.00025233487770714356,
      "loss": 3.3367,
      "step": 53900
    },
    {
      "epoch": 5.806694650737272,
      "grad_norm": 0.6494348049163818,
      "learning_rate": 0.0002520116366770822,
      "loss": 3.3419,
      "step": 53950
    },
    {
      "epoch": 5.812076202776881,
      "grad_norm": 0.6735350489616394,
      "learning_rate": 0.00025168839564702075,
      "loss": 3.3548,
      "step": 54000
    },
    {
      "epoch": 5.812076202776881,
      "eval_accuracy": 0.3825783352008954,
      "eval_loss": 3.392937660217285,
      "eval_runtime": 185.5082,
      "eval_samples_per_second": 97.09,
      "eval_steps_per_second": 6.07,
      "step": 54000
    },
    {
      "epoch": 5.817457754816489,
      "grad_norm": 0.7118306159973145,
      "learning_rate": 0.00025136515461695935,
      "loss": 3.3457,
      "step": 54050
    },
    {
      "epoch": 5.822839306856097,
      "grad_norm": 0.7467418313026428,
      "learning_rate": 0.00025104191358689794,
      "loss": 3.3411,
      "step": 54100
    },
    {
      "epoch": 5.828220858895706,
      "grad_norm": 0.6742581129074097,
      "learning_rate": 0.00025071867255683654,
      "loss": 3.3503,
      "step": 54150
    },
    {
      "epoch": 5.833602410935313,
      "grad_norm": 0.6745625138282776,
      "learning_rate": 0.00025039543152677513,
      "loss": 3.3431,
      "step": 54200
    },
    {
      "epoch": 5.838983962974922,
      "grad_norm": 0.6575548052787781,
      "learning_rate": 0.0002500721904967137,
      "loss": 3.3253,
      "step": 54250
    },
    {
      "epoch": 5.84436551501453,
      "grad_norm": 0.6717263460159302,
      "learning_rate": 0.00024974894946665227,
      "loss": 3.3446,
      "step": 54300
    },
    {
      "epoch": 5.849747067054138,
      "grad_norm": 0.6477599740028381,
      "learning_rate": 0.00024942570843659086,
      "loss": 3.3369,
      "step": 54350
    },
    {
      "epoch": 5.855128619093747,
      "grad_norm": 0.7447801232337952,
      "learning_rate": 0.00024910246740652946,
      "loss": 3.356,
      "step": 54400
    },
    {
      "epoch": 5.860510171133355,
      "grad_norm": 0.686426043510437,
      "learning_rate": 0.000248779226376468,
      "loss": 3.3363,
      "step": 54450
    },
    {
      "epoch": 5.865891723172963,
      "grad_norm": 0.659697413444519,
      "learning_rate": 0.00024845598534640665,
      "loss": 3.3381,
      "step": 54500
    },
    {
      "epoch": 5.871273275212571,
      "grad_norm": 0.6772626638412476,
      "learning_rate": 0.0002481327443163452,
      "loss": 3.3386,
      "step": 54550
    },
    {
      "epoch": 5.87665482725218,
      "grad_norm": 0.6307031512260437,
      "learning_rate": 0.0002478095032862838,
      "loss": 3.3289,
      "step": 54600
    },
    {
      "epoch": 5.882036379291788,
      "grad_norm": 0.7684736251831055,
      "learning_rate": 0.0002474862622562224,
      "loss": 3.3493,
      "step": 54650
    },
    {
      "epoch": 5.887417931331396,
      "grad_norm": 0.7073497772216797,
      "learning_rate": 0.00024716302122616097,
      "loss": 3.3683,
      "step": 54700
    },
    {
      "epoch": 5.892799483371004,
      "grad_norm": 0.7363914847373962,
      "learning_rate": 0.0002468397801960995,
      "loss": 3.3416,
      "step": 54750
    },
    {
      "epoch": 5.898181035410612,
      "grad_norm": 0.6807088255882263,
      "learning_rate": 0.0002465165391660381,
      "loss": 3.3376,
      "step": 54800
    },
    {
      "epoch": 5.903562587450221,
      "grad_norm": 0.6982871294021606,
      "learning_rate": 0.0002461932981359767,
      "loss": 3.3612,
      "step": 54850
    },
    {
      "epoch": 5.9089441394898286,
      "grad_norm": 0.6537452936172485,
      "learning_rate": 0.0002458700571059153,
      "loss": 3.3438,
      "step": 54900
    },
    {
      "epoch": 5.914325691529437,
      "grad_norm": 0.6711610555648804,
      "learning_rate": 0.0002455468160758539,
      "loss": 3.3588,
      "step": 54950
    },
    {
      "epoch": 5.919707243569046,
      "grad_norm": 0.6602628231048584,
      "learning_rate": 0.00024522357504579243,
      "loss": 3.3568,
      "step": 55000
    },
    {
      "epoch": 5.919707243569046,
      "eval_accuracy": 0.38311388528493295,
      "eval_loss": 3.3890511989593506,
      "eval_runtime": 185.2821,
      "eval_samples_per_second": 97.209,
      "eval_steps_per_second": 6.077,
      "step": 55000
    },
    {
      "epoch": 5.925088795608653,
      "grad_norm": 0.67746502161026,
      "learning_rate": 0.000244900334015731,
      "loss": 3.3457,
      "step": 55050
    },
    {
      "epoch": 5.930470347648262,
      "grad_norm": 0.6979883313179016,
      "learning_rate": 0.0002445770929856696,
      "loss": 3.3437,
      "step": 55100
    },
    {
      "epoch": 5.93585189968787,
      "grad_norm": 0.7038848400115967,
      "learning_rate": 0.0002442538519556082,
      "loss": 3.3511,
      "step": 55150
    },
    {
      "epoch": 5.941233451727478,
      "grad_norm": 0.6588810682296753,
      "learning_rate": 0.00024393707574614802,
      "loss": 3.3561,
      "step": 55200
    },
    {
      "epoch": 5.946615003767087,
      "grad_norm": Infinity,
      "learning_rate": 0.00024362029953668785,
      "loss": 3.3511,
      "step": 55250
    },
    {
      "epoch": 5.951996555806694,
      "grad_norm": 0.7010414600372314,
      "learning_rate": 0.00024329705850662642,
      "loss": 3.3628,
      "step": 55300
    },
    {
      "epoch": 5.957378107846303,
      "grad_norm": 0.7299208045005798,
      "learning_rate": 0.000242973817476565,
      "loss": 3.37,
      "step": 55350
    },
    {
      "epoch": 5.962759659885911,
      "grad_norm": 0.710688054561615,
      "learning_rate": 0.00024265057644650358,
      "loss": 3.3335,
      "step": 55400
    },
    {
      "epoch": 5.968141211925519,
      "grad_norm": 0.6901038885116577,
      "learning_rate": 0.00024232733541644218,
      "loss": 3.3469,
      "step": 55450
    },
    {
      "epoch": 5.973522763965128,
      "grad_norm": 0.7379507422447205,
      "learning_rate": 0.00024200409438638075,
      "loss": 3.357,
      "step": 55500
    },
    {
      "epoch": 5.978904316004736,
      "grad_norm": 0.7134886980056763,
      "learning_rate": 0.00024168085335631934,
      "loss": 3.3426,
      "step": 55550
    },
    {
      "epoch": 5.984285868044344,
      "grad_norm": 0.6999474763870239,
      "learning_rate": 0.0002413576123262579,
      "loss": 3.3432,
      "step": 55600
    },
    {
      "epoch": 5.989667420083952,
      "grad_norm": 0.6996995210647583,
      "learning_rate": 0.0002410343712961965,
      "loss": 3.341,
      "step": 55650
    },
    {
      "epoch": 5.995048972123561,
      "grad_norm": 0.747147798538208,
      "learning_rate": 0.0002407111302661351,
      "loss": 3.346,
      "step": 55700
    },
    {
      "epoch": 6.000430524163169,
      "grad_norm": 0.6988963484764099,
      "learning_rate": 0.00024038788923607367,
      "loss": 3.3414,
      "step": 55750
    },
    {
      "epoch": 6.005812076202777,
      "grad_norm": 0.6728678345680237,
      "learning_rate": 0.0002400646482060123,
      "loss": 3.2391,
      "step": 55800
    },
    {
      "epoch": 6.011193628242385,
      "grad_norm": 0.7063876390457153,
      "learning_rate": 0.00023974140717595085,
      "loss": 3.2604,
      "step": 55850
    },
    {
      "epoch": 6.016575180281993,
      "grad_norm": 0.7127146124839783,
      "learning_rate": 0.00023941816614588942,
      "loss": 3.2628,
      "step": 55900
    },
    {
      "epoch": 6.021956732321602,
      "grad_norm": 0.6635514497756958,
      "learning_rate": 0.00023909492511582802,
      "loss": 3.2628,
      "step": 55950
    },
    {
      "epoch": 6.0273382843612096,
      "grad_norm": 0.7230376601219177,
      "learning_rate": 0.0002387716840857666,
      "loss": 3.2596,
      "step": 56000
    },
    {
      "epoch": 6.0273382843612096,
      "eval_accuracy": 0.38299686612475675,
      "eval_loss": 3.39104962348938,
      "eval_runtime": 185.2126,
      "eval_samples_per_second": 97.245,
      "eval_steps_per_second": 6.08,
      "step": 56000
    },
    {
      "epoch": 6.032719836400818,
      "grad_norm": 0.702446460723877,
      "learning_rate": 0.00023844844305570518,
      "loss": 3.279,
      "step": 56050
    },
    {
      "epoch": 6.038101388440427,
      "grad_norm": 0.6905210614204407,
      "learning_rate": 0.00023812520202564377,
      "loss": 3.2642,
      "step": 56100
    },
    {
      "epoch": 6.043482940480034,
      "grad_norm": 0.6880192756652832,
      "learning_rate": 0.00023780196099558234,
      "loss": 3.2609,
      "step": 56150
    },
    {
      "epoch": 6.048864492519643,
      "grad_norm": 0.7522834539413452,
      "learning_rate": 0.00023747871996552094,
      "loss": 3.2529,
      "step": 56200
    },
    {
      "epoch": 6.0542460445592505,
      "grad_norm": 0.6889835596084595,
      "learning_rate": 0.00023715547893545953,
      "loss": 3.2631,
      "step": 56250
    },
    {
      "epoch": 6.059627596598859,
      "grad_norm": 0.7528659701347351,
      "learning_rate": 0.0002368322379053981,
      "loss": 3.2823,
      "step": 56300
    },
    {
      "epoch": 6.065009148638468,
      "grad_norm": 0.7965409159660339,
      "learning_rate": 0.00023650899687533667,
      "loss": 3.2735,
      "step": 56350
    },
    {
      "epoch": 6.070390700678075,
      "grad_norm": 0.7661939263343811,
      "learning_rate": 0.0002361857558452753,
      "loss": 3.2754,
      "step": 56400
    },
    {
      "epoch": 6.075772252717684,
      "grad_norm": 0.6804177165031433,
      "learning_rate": 0.00023586251481521386,
      "loss": 3.2684,
      "step": 56450
    },
    {
      "epoch": 6.081153804757292,
      "grad_norm": 0.7249098420143127,
      "learning_rate": 0.00023553927378515242,
      "loss": 3.2715,
      "step": 56500
    },
    {
      "epoch": 6.0865353567969,
      "grad_norm": 0.7100164890289307,
      "learning_rate": 0.00023521603275509104,
      "loss": 3.2686,
      "step": 56550
    },
    {
      "epoch": 6.091916908836509,
      "grad_norm": 0.6656424403190613,
      "learning_rate": 0.0002348927917250296,
      "loss": 3.2634,
      "step": 56600
    },
    {
      "epoch": 6.097298460876116,
      "grad_norm": 0.6703857183456421,
      "learning_rate": 0.00023456955069496818,
      "loss": 3.2929,
      "step": 56650
    },
    {
      "epoch": 6.102680012915725,
      "grad_norm": 0.6990338563919067,
      "learning_rate": 0.00023424630966490677,
      "loss": 3.2827,
      "step": 56700
    },
    {
      "epoch": 6.108061564955333,
      "grad_norm": 0.6904770731925964,
      "learning_rate": 0.00023392306863484537,
      "loss": 3.2696,
      "step": 56750
    },
    {
      "epoch": 6.113443116994941,
      "grad_norm": 0.7322364449501038,
      "learning_rate": 0.00023359982760478396,
      "loss": 3.2717,
      "step": 56800
    },
    {
      "epoch": 6.11882466903455,
      "grad_norm": 0.7013688683509827,
      "learning_rate": 0.00023327658657472253,
      "loss": 3.2955,
      "step": 56850
    },
    {
      "epoch": 6.124206221074158,
      "grad_norm": 0.7182783484458923,
      "learning_rate": 0.0002329533455446611,
      "loss": 3.2746,
      "step": 56900
    },
    {
      "epoch": 6.129587773113766,
      "grad_norm": 0.720745325088501,
      "learning_rate": 0.00023263010451459972,
      "loss": 3.2681,
      "step": 56950
    },
    {
      "epoch": 6.134969325153374,
      "grad_norm": 0.7033182382583618,
      "learning_rate": 0.0002323068634845383,
      "loss": 3.28,
      "step": 57000
    },
    {
      "epoch": 6.134969325153374,
      "eval_accuracy": 0.3832887077814265,
      "eval_loss": 3.3910751342773438,
      "eval_runtime": 184.9937,
      "eval_samples_per_second": 97.36,
      "eval_steps_per_second": 6.087,
      "step": 57000
    },
    {
      "epoch": 6.140350877192983,
      "grad_norm": 0.6933819055557251,
      "learning_rate": 0.00023198362245447686,
      "loss": 3.2857,
      "step": 57050
    },
    {
      "epoch": 6.1457324292325906,
      "grad_norm": 0.7143245339393616,
      "learning_rate": 0.00023166038142441548,
      "loss": 3.2772,
      "step": 57100
    },
    {
      "epoch": 6.151113981272199,
      "grad_norm": 0.6586877107620239,
      "learning_rate": 0.00023133714039435404,
      "loss": 3.2824,
      "step": 57150
    },
    {
      "epoch": 6.156495533311807,
      "grad_norm": 0.667927622795105,
      "learning_rate": 0.0002310138993642926,
      "loss": 3.2786,
      "step": 57200
    },
    {
      "epoch": 6.161877085351415,
      "grad_norm": 0.6979237198829651,
      "learning_rate": 0.0002306906583342312,
      "loss": 3.267,
      "step": 57250
    },
    {
      "epoch": 6.167258637391024,
      "grad_norm": 0.7055776119232178,
      "learning_rate": 0.00023036741730416977,
      "loss": 3.2706,
      "step": 57300
    },
    {
      "epoch": 6.1726401894306315,
      "grad_norm": 0.6994844675064087,
      "learning_rate": 0.00023004417627410837,
      "loss": 3.2844,
      "step": 57350
    },
    {
      "epoch": 6.17802174147024,
      "grad_norm": 0.7888532876968384,
      "learning_rate": 0.00022972093524404696,
      "loss": 3.2702,
      "step": 57400
    },
    {
      "epoch": 6.183403293509849,
      "grad_norm": 0.6941890716552734,
      "learning_rate": 0.00022939769421398553,
      "loss": 3.2751,
      "step": 57450
    },
    {
      "epoch": 6.188784845549456,
      "grad_norm": 0.739536702632904,
      "learning_rate": 0.0002290744531839241,
      "loss": 3.2927,
      "step": 57500
    },
    {
      "epoch": 6.194166397589065,
      "grad_norm": 0.7241300940513611,
      "learning_rate": 0.00022875121215386272,
      "loss": 3.2802,
      "step": 57550
    },
    {
      "epoch": 6.1995479496286725,
      "grad_norm": 0.7131288647651672,
      "learning_rate": 0.0002284279711238013,
      "loss": 3.292,
      "step": 57600
    },
    {
      "epoch": 6.204929501668281,
      "grad_norm": 0.6874881982803345,
      "learning_rate": 0.00022810473009373986,
      "loss": 3.2903,
      "step": 57650
    },
    {
      "epoch": 6.21031105370789,
      "grad_norm": 0.6826223134994507,
      "learning_rate": 0.00022778148906367848,
      "loss": 3.278,
      "step": 57700
    },
    {
      "epoch": 6.215692605747497,
      "grad_norm": 0.6940621137619019,
      "learning_rate": 0.00022746471285421828,
      "loss": 3.2893,
      "step": 57750
    },
    {
      "epoch": 6.221074157787106,
      "grad_norm": 0.7043401598930359,
      "learning_rate": 0.00022714147182415685,
      "loss": 3.2715,
      "step": 57800
    },
    {
      "epoch": 6.226455709826714,
      "grad_norm": 0.722109317779541,
      "learning_rate": 0.00022681823079409545,
      "loss": 3.3039,
      "step": 57850
    },
    {
      "epoch": 6.231837261866322,
      "grad_norm": 0.7242453098297119,
      "learning_rate": 0.00022649498976403404,
      "loss": 3.285,
      "step": 57900
    },
    {
      "epoch": 6.237218813905931,
      "grad_norm": 0.7554165720939636,
      "learning_rate": 0.0002261717487339726,
      "loss": 3.2744,
      "step": 57950
    },
    {
      "epoch": 6.242600365945538,
      "grad_norm": 0.8292284607887268,
      "learning_rate": 0.00022584850770391118,
      "loss": 3.2973,
      "step": 58000
    },
    {
      "epoch": 6.242600365945538,
      "eval_accuracy": 0.3837654766531564,
      "eval_loss": 3.386949062347412,
      "eval_runtime": 185.535,
      "eval_samples_per_second": 97.076,
      "eval_steps_per_second": 6.069,
      "step": 58000
    },
    {
      "epoch": 6.247981917985147,
      "grad_norm": 0.7098879218101501,
      "learning_rate": 0.0002255252666738498,
      "loss": 3.2958,
      "step": 58050
    },
    {
      "epoch": 6.253363470024755,
      "grad_norm": 0.7486851811408997,
      "learning_rate": 0.00022520202564378837,
      "loss": 3.2775,
      "step": 58100
    },
    {
      "epoch": 6.258745022064363,
      "grad_norm": 0.6934053301811218,
      "learning_rate": 0.00022487878461372693,
      "loss": 3.2729,
      "step": 58150
    },
    {
      "epoch": 6.264126574103972,
      "grad_norm": 0.6911810636520386,
      "learning_rate": 0.00022455554358366555,
      "loss": 3.3006,
      "step": 58200
    },
    {
      "epoch": 6.26950812614358,
      "grad_norm": 0.6725990772247314,
      "learning_rate": 0.00022423230255360412,
      "loss": 3.2913,
      "step": 58250
    },
    {
      "epoch": 6.274889678183188,
      "grad_norm": 0.7402995824813843,
      "learning_rate": 0.0002239090615235427,
      "loss": 3.3098,
      "step": 58300
    },
    {
      "epoch": 6.280271230222796,
      "grad_norm": 0.7451393008232117,
      "learning_rate": 0.00022358582049348128,
      "loss": 3.2933,
      "step": 58350
    },
    {
      "epoch": 6.285652782262405,
      "grad_norm": 0.6835018396377563,
      "learning_rate": 0.00022326257946341988,
      "loss": 3.2727,
      "step": 58400
    },
    {
      "epoch": 6.2910343343020125,
      "grad_norm": 0.7079624533653259,
      "learning_rate": 0.00022293933843335845,
      "loss": 3.2796,
      "step": 58450
    },
    {
      "epoch": 6.296415886341621,
      "grad_norm": 0.7219333648681641,
      "learning_rate": 0.00022261609740329704,
      "loss": 3.2822,
      "step": 58500
    },
    {
      "epoch": 6.301797438381229,
      "grad_norm": 0.6832009553909302,
      "learning_rate": 0.0002222928563732356,
      "loss": 3.2912,
      "step": 58550
    },
    {
      "epoch": 6.307178990420837,
      "grad_norm": 0.688037633895874,
      "learning_rate": 0.00022196961534317423,
      "loss": 3.2903,
      "step": 58600
    },
    {
      "epoch": 6.312560542460446,
      "grad_norm": 0.7042102813720703,
      "learning_rate": 0.0002216463743131128,
      "loss": 3.2941,
      "step": 58650
    },
    {
      "epoch": 6.3179420945000535,
      "grad_norm": 0.7006003856658936,
      "learning_rate": 0.00022132313328305137,
      "loss": 3.2949,
      "step": 58700
    },
    {
      "epoch": 6.323323646539662,
      "grad_norm": 0.6868375539779663,
      "learning_rate": 0.00022099989225299,
      "loss": 3.3073,
      "step": 58750
    },
    {
      "epoch": 6.328705198579271,
      "grad_norm": 0.7184757590293884,
      "learning_rate": 0.00022067665122292856,
      "loss": 3.2762,
      "step": 58800
    },
    {
      "epoch": 6.334086750618878,
      "grad_norm": 0.7562534213066101,
      "learning_rate": 0.00022035341019286712,
      "loss": 3.285,
      "step": 58850
    },
    {
      "epoch": 6.339468302658487,
      "grad_norm": 0.7731133103370667,
      "learning_rate": 0.00022003016916280572,
      "loss": 3.3064,
      "step": 58900
    },
    {
      "epoch": 6.344849854698095,
      "grad_norm": 0.7119162678718567,
      "learning_rate": 0.00021970692813274428,
      "loss": 3.2866,
      "step": 58950
    },
    {
      "epoch": 6.350231406737703,
      "grad_norm": 0.7308032512664795,
      "learning_rate": 0.00021938368710268288,
      "loss": 3.2905,
      "step": 59000
    },
    {
      "epoch": 6.350231406737703,
      "eval_accuracy": 0.38389651203586067,
      "eval_loss": 3.3842954635620117,
      "eval_runtime": 185.0331,
      "eval_samples_per_second": 97.339,
      "eval_steps_per_second": 6.085,
      "step": 59000
    },
    {
      "epoch": 6.355612958777312,
      "grad_norm": 0.7241775393486023,
      "learning_rate": 0.00021906044607262147,
      "loss": 3.2861,
      "step": 59050
    },
    {
      "epoch": 6.360994510816919,
      "grad_norm": 0.7252910137176514,
      "learning_rate": 0.00021873720504256004,
      "loss": 3.2944,
      "step": 59100
    },
    {
      "epoch": 6.366376062856528,
      "grad_norm": 0.7173689603805542,
      "learning_rate": 0.0002184139640124986,
      "loss": 3.2991,
      "step": 59150
    },
    {
      "epoch": 6.371757614896136,
      "grad_norm": 0.7135725617408752,
      "learning_rate": 0.00021809072298243723,
      "loss": 3.2793,
      "step": 59200
    },
    {
      "epoch": 6.377139166935744,
      "grad_norm": 0.6765625476837158,
      "learning_rate": 0.0002177674819523758,
      "loss": 3.3045,
      "step": 59250
    },
    {
      "epoch": 6.382520718975353,
      "grad_norm": 0.7070035934448242,
      "learning_rate": 0.00021744424092231437,
      "loss": 3.2847,
      "step": 59300
    },
    {
      "epoch": 6.387902271014961,
      "grad_norm": 0.7414677739143372,
      "learning_rate": 0.000217120999892253,
      "loss": 3.2974,
      "step": 59350
    },
    {
      "epoch": 6.393283823054569,
      "grad_norm": 0.6972620487213135,
      "learning_rate": 0.00021679775886219156,
      "loss": 3.2893,
      "step": 59400
    },
    {
      "epoch": 6.398665375094177,
      "grad_norm": 0.7405353784561157,
      "learning_rate": 0.00021647451783213012,
      "loss": 3.2922,
      "step": 59450
    },
    {
      "epoch": 6.404046927133785,
      "grad_norm": 0.7180378437042236,
      "learning_rate": 0.00021615127680206872,
      "loss": 3.3001,
      "step": 59500
    },
    {
      "epoch": 6.4094284791733935,
      "grad_norm": 0.7183099985122681,
      "learning_rate": 0.0002158280357720073,
      "loss": 3.3021,
      "step": 59550
    },
    {
      "epoch": 6.414810031213002,
      "grad_norm": 0.7518490552902222,
      "learning_rate": 0.0002155047947419459,
      "loss": 3.3136,
      "step": 59600
    },
    {
      "epoch": 6.42019158325261,
      "grad_norm": 0.7282083630561829,
      "learning_rate": 0.00021518155371188447,
      "loss": 3.2914,
      "step": 59650
    },
    {
      "epoch": 6.425573135292218,
      "grad_norm": 0.7133340239524841,
      "learning_rate": 0.00021485831268182304,
      "loss": 3.3061,
      "step": 59700
    },
    {
      "epoch": 6.430954687331827,
      "grad_norm": 0.7224932312965393,
      "learning_rate": 0.00021453507165176166,
      "loss": 3.3086,
      "step": 59750
    },
    {
      "epoch": 6.4363362393714345,
      "grad_norm": 0.7358588576316833,
      "learning_rate": 0.00021421183062170023,
      "loss": 3.281,
      "step": 59800
    },
    {
      "epoch": 6.441717791411043,
      "grad_norm": 0.73481684923172,
      "learning_rate": 0.0002138885895916388,
      "loss": 3.2774,
      "step": 59850
    },
    {
      "epoch": 6.447099343450651,
      "grad_norm": 0.7021733522415161,
      "learning_rate": 0.00021356534856157742,
      "loss": 3.2916,
      "step": 59900
    },
    {
      "epoch": 6.452480895490259,
      "grad_norm": 0.6721026301383972,
      "learning_rate": 0.000213242107531516,
      "loss": 3.2931,
      "step": 59950
    },
    {
      "epoch": 6.457862447529868,
      "grad_norm": 0.7121398448944092,
      "learning_rate": 0.00021291886650145456,
      "loss": 3.2984,
      "step": 60000
    },
    {
      "epoch": 6.457862447529868,
      "eval_accuracy": 0.3844197843435752,
      "eval_loss": 3.379528760910034,
      "eval_runtime": 185.6608,
      "eval_samples_per_second": 97.01,
      "eval_steps_per_second": 6.065,
      "step": 60000
    },
    {
      "epoch": 6.4632439995694755,
      "grad_norm": 0.7491379380226135,
      "learning_rate": 0.00021259562547139315,
      "loss": 3.2939,
      "step": 60050
    },
    {
      "epoch": 6.468625551609084,
      "grad_norm": 0.7895132303237915,
      "learning_rate": 0.00021227238444133175,
      "loss": 3.284,
      "step": 60100
    },
    {
      "epoch": 6.474007103648693,
      "grad_norm": 0.7313899993896484,
      "learning_rate": 0.0002119491434112703,
      "loss": 3.2885,
      "step": 60150
    },
    {
      "epoch": 6.4793886556883,
      "grad_norm": 0.7327179312705994,
      "learning_rate": 0.0002116259023812089,
      "loss": 3.2818,
      "step": 60200
    },
    {
      "epoch": 6.484770207727909,
      "grad_norm": 0.8201233148574829,
      "learning_rate": 0.00021130266135114747,
      "loss": 3.2989,
      "step": 60250
    },
    {
      "epoch": 6.490151759767517,
      "grad_norm": 0.7711613178253174,
      "learning_rate": 0.00021097942032108607,
      "loss": 3.2888,
      "step": 60300
    },
    {
      "epoch": 6.495533311807125,
      "grad_norm": 0.7210743427276611,
      "learning_rate": 0.00021066264411162588,
      "loss": 3.2962,
      "step": 60350
    },
    {
      "epoch": 6.500914863846734,
      "grad_norm": 0.7693405151367188,
      "learning_rate": 0.00021033940308156447,
      "loss": 3.2894,
      "step": 60400
    },
    {
      "epoch": 6.506296415886341,
      "grad_norm": 0.6683956980705261,
      "learning_rate": 0.00021001616205150307,
      "loss": 3.2994,
      "step": 60450
    },
    {
      "epoch": 6.51167796792595,
      "grad_norm": 0.7586647272109985,
      "learning_rate": 0.00020969292102144163,
      "loss": 3.2885,
      "step": 60500
    },
    {
      "epoch": 6.517059519965558,
      "grad_norm": 0.7385610342025757,
      "learning_rate": 0.00020936967999138023,
      "loss": 3.2939,
      "step": 60550
    },
    {
      "epoch": 6.522441072005166,
      "grad_norm": 0.6777862906455994,
      "learning_rate": 0.0002090464389613188,
      "loss": 3.2998,
      "step": 60600
    },
    {
      "epoch": 6.5278226240447745,
      "grad_norm": 0.7657138109207153,
      "learning_rate": 0.0002087231979312574,
      "loss": 3.282,
      "step": 60650
    },
    {
      "epoch": 6.533204176084383,
      "grad_norm": 0.740245521068573,
      "learning_rate": 0.00020839995690119598,
      "loss": 3.2956,
      "step": 60700
    },
    {
      "epoch": 6.538585728123991,
      "grad_norm": 0.7346565127372742,
      "learning_rate": 0.00020807671587113455,
      "loss": 3.3257,
      "step": 60750
    },
    {
      "epoch": 6.543967280163599,
      "grad_norm": 0.7704212665557861,
      "learning_rate": 0.00020775347484107312,
      "loss": 3.2889,
      "step": 60800
    },
    {
      "epoch": 6.549348832203208,
      "grad_norm": 0.7644638419151306,
      "learning_rate": 0.00020743023381101174,
      "loss": 3.3044,
      "step": 60850
    },
    {
      "epoch": 6.5547303842428155,
      "grad_norm": 0.6896412968635559,
      "learning_rate": 0.0002071069927809503,
      "loss": 3.3038,
      "step": 60900
    },
    {
      "epoch": 6.560111936282424,
      "grad_norm": 0.7370016574859619,
      "learning_rate": 0.00020678375175088888,
      "loss": 3.3193,
      "step": 60950
    },
    {
      "epoch": 6.565493488322032,
      "grad_norm": 0.7837535738945007,
      "learning_rate": 0.0002064605107208275,
      "loss": 3.2829,
      "step": 61000
    },
    {
      "epoch": 6.565493488322032,
      "eval_accuracy": 0.3846488246310881,
      "eval_loss": 3.3760488033294678,
      "eval_runtime": 184.8377,
      "eval_samples_per_second": 97.442,
      "eval_steps_per_second": 6.092,
      "step": 61000
    },
    {
      "epoch": 6.57087504036164,
      "grad_norm": 0.7083461880683899,
      "learning_rate": 0.00020613726969076607,
      "loss": 3.313,
      "step": 61050
    },
    {
      "epoch": 6.576256592401249,
      "grad_norm": 0.7646893858909607,
      "learning_rate": 0.00020581402866070463,
      "loss": 3.297,
      "step": 61100
    },
    {
      "epoch": 6.5816381444408565,
      "grad_norm": 0.765580415725708,
      "learning_rate": 0.00020549078763064323,
      "loss": 3.299,
      "step": 61150
    },
    {
      "epoch": 6.587019696480465,
      "grad_norm": 0.7021796107292175,
      "learning_rate": 0.00020516754660058182,
      "loss": 3.2948,
      "step": 61200
    },
    {
      "epoch": 6.592401248520073,
      "grad_norm": 0.8181710243225098,
      "learning_rate": 0.0002048443055705204,
      "loss": 3.2931,
      "step": 61250
    },
    {
      "epoch": 6.597782800559681,
      "grad_norm": 0.7198193669319153,
      "learning_rate": 0.00020452106454045898,
      "loss": 3.2982,
      "step": 61300
    },
    {
      "epoch": 6.60316435259929,
      "grad_norm": 0.7290482521057129,
      "learning_rate": 0.00020419782351039755,
      "loss": 3.3045,
      "step": 61350
    },
    {
      "epoch": 6.608545904638898,
      "grad_norm": 0.7291182279586792,
      "learning_rate": 0.00020387458248033617,
      "loss": 3.3004,
      "step": 61400
    },
    {
      "epoch": 6.613927456678506,
      "grad_norm": 0.7434918880462646,
      "learning_rate": 0.00020355134145027474,
      "loss": 3.2776,
      "step": 61450
    },
    {
      "epoch": 6.619309008718115,
      "grad_norm": 0.7196027040481567,
      "learning_rate": 0.0002032281004202133,
      "loss": 3.2941,
      "step": 61500
    },
    {
      "epoch": 6.624690560757722,
      "grad_norm": 0.7437618374824524,
      "learning_rate": 0.00020290485939015193,
      "loss": 3.3099,
      "step": 61550
    },
    {
      "epoch": 6.630072112797331,
      "grad_norm": 0.6985304355621338,
      "learning_rate": 0.0002025816183600905,
      "loss": 3.3231,
      "step": 61600
    },
    {
      "epoch": 6.635453664836939,
      "grad_norm": 0.7338959574699402,
      "learning_rate": 0.00020225837733002907,
      "loss": 3.2847,
      "step": 61650
    },
    {
      "epoch": 6.640835216876547,
      "grad_norm": 0.7830687165260315,
      "learning_rate": 0.00020193513629996766,
      "loss": 3.286,
      "step": 61700
    },
    {
      "epoch": 6.6462167689161555,
      "grad_norm": 0.7383367419242859,
      "learning_rate": 0.00020161189526990626,
      "loss": 3.2845,
      "step": 61750
    },
    {
      "epoch": 6.651598320955763,
      "grad_norm": 0.7406599521636963,
      "learning_rate": 0.00020128865423984482,
      "loss": 3.3024,
      "step": 61800
    },
    {
      "epoch": 6.656979872995372,
      "grad_norm": 0.7393373250961304,
      "learning_rate": 0.00020096541320978342,
      "loss": 3.2814,
      "step": 61850
    },
    {
      "epoch": 6.66236142503498,
      "grad_norm": 0.7196076512336731,
      "learning_rate": 0.00020064217217972199,
      "loss": 3.2986,
      "step": 61900
    },
    {
      "epoch": 6.667742977074588,
      "grad_norm": 0.774937093257904,
      "learning_rate": 0.00020031893114966058,
      "loss": 3.3034,
      "step": 61950
    },
    {
      "epoch": 6.6731245291141965,
      "grad_norm": 0.739497721195221,
      "learning_rate": 0.00019999569011959917,
      "loss": 3.2952,
      "step": 62000
    },
    {
      "epoch": 6.6731245291141965,
      "eval_accuracy": 0.385301937139741,
      "eval_loss": 3.3732869625091553,
      "eval_runtime": 185.3712,
      "eval_samples_per_second": 97.162,
      "eval_steps_per_second": 6.074,
      "step": 62000
    },
    {
      "epoch": 6.678506081153805,
      "grad_norm": 0.7534037828445435,
      "learning_rate": 0.00019967244908953774,
      "loss": 3.3124,
      "step": 62050
    },
    {
      "epoch": 6.683887633193413,
      "grad_norm": 0.733932375907898,
      "learning_rate": 0.0001993492080594763,
      "loss": 3.3007,
      "step": 62100
    },
    {
      "epoch": 6.689269185233021,
      "grad_norm": 0.7124557495117188,
      "learning_rate": 0.00019902596702941493,
      "loss": 3.2984,
      "step": 62150
    },
    {
      "epoch": 6.69465073727263,
      "grad_norm": 0.7213685512542725,
      "learning_rate": 0.0001987027259993535,
      "loss": 3.2998,
      "step": 62200
    },
    {
      "epoch": 6.7000322893122375,
      "grad_norm": 0.7234432101249695,
      "learning_rate": 0.00019837948496929207,
      "loss": 3.2971,
      "step": 62250
    },
    {
      "epoch": 6.705413841351846,
      "grad_norm": 0.7089570164680481,
      "learning_rate": 0.00019805624393923066,
      "loss": 3.3171,
      "step": 62300
    },
    {
      "epoch": 6.710795393391454,
      "grad_norm": 0.739058256149292,
      "learning_rate": 0.00019773300290916926,
      "loss": 3.3051,
      "step": 62350
    },
    {
      "epoch": 6.716176945431062,
      "grad_norm": 0.7353187203407288,
      "learning_rate": 0.00019740976187910785,
      "loss": 3.304,
      "step": 62400
    },
    {
      "epoch": 6.721558497470671,
      "grad_norm": 0.7043803334236145,
      "learning_rate": 0.00019708652084904642,
      "loss": 3.2963,
      "step": 62450
    },
    {
      "epoch": 6.7269400495102785,
      "grad_norm": 0.7873364090919495,
      "learning_rate": 0.00019676327981898499,
      "loss": 3.2826,
      "step": 62500
    },
    {
      "epoch": 6.732321601549887,
      "grad_norm": 0.7169027328491211,
      "learning_rate": 0.0001964400387889236,
      "loss": 3.2933,
      "step": 62550
    },
    {
      "epoch": 6.737703153589496,
      "grad_norm": 0.7438593506813049,
      "learning_rate": 0.00019611679775886217,
      "loss": 3.3092,
      "step": 62600
    },
    {
      "epoch": 6.743084705629103,
      "grad_norm": 0.726280927658081,
      "learning_rate": 0.00019579355672880074,
      "loss": 3.3103,
      "step": 62650
    },
    {
      "epoch": 6.748466257668712,
      "grad_norm": 0.7109901309013367,
      "learning_rate": 0.00019547031569873936,
      "loss": 3.2905,
      "step": 62700
    },
    {
      "epoch": 6.75384780970832,
      "grad_norm": 0.7438680529594421,
      "learning_rate": 0.00019514707466867793,
      "loss": 3.2913,
      "step": 62750
    },
    {
      "epoch": 6.759229361747928,
      "grad_norm": 0.7096954584121704,
      "learning_rate": 0.0001948238336386165,
      "loss": 3.2881,
      "step": 62800
    },
    {
      "epoch": 6.7646109137875365,
      "grad_norm": 0.7945369482040405,
      "learning_rate": 0.0001945005926085551,
      "loss": 3.2984,
      "step": 62850
    },
    {
      "epoch": 6.769992465827144,
      "grad_norm": 0.7154929041862488,
      "learning_rate": 0.0001941773515784937,
      "loss": 3.2864,
      "step": 62900
    },
    {
      "epoch": 6.775374017866753,
      "grad_norm": 0.7511465549468994,
      "learning_rate": 0.00019385411054843226,
      "loss": 3.277,
      "step": 62950
    },
    {
      "epoch": 6.780755569906361,
      "grad_norm": 0.7981410026550293,
      "learning_rate": 0.00019353086951837085,
      "loss": 3.2994,
      "step": 63000
    },
    {
      "epoch": 6.780755569906361,
      "eval_accuracy": 0.3857326371870376,
      "eval_loss": 3.367685079574585,
      "eval_runtime": 185.1684,
      "eval_samples_per_second": 97.268,
      "eval_steps_per_second": 6.081,
      "step": 63000
    },
    {
      "epoch": 6.786137121945969,
      "grad_norm": 0.7183791995048523,
      "learning_rate": 0.00019320762848830942,
      "loss": 3.2943,
      "step": 63050
    },
    {
      "epoch": 6.7915186739855775,
      "grad_norm": 0.7264687418937683,
      "learning_rate": 0.000192884387458248,
      "loss": 3.294,
      "step": 63100
    },
    {
      "epoch": 6.796900226025185,
      "grad_norm": 0.7436566352844238,
      "learning_rate": 0.0001925611464281866,
      "loss": 3.2973,
      "step": 63150
    },
    {
      "epoch": 6.802281778064794,
      "grad_norm": 0.7772260904312134,
      "learning_rate": 0.00019223790539812518,
      "loss": 3.2924,
      "step": 63200
    },
    {
      "epoch": 6.807663330104402,
      "grad_norm": 0.7376819252967834,
      "learning_rate": 0.00019191466436806374,
      "loss": 3.2918,
      "step": 63250
    },
    {
      "epoch": 6.813044882144011,
      "grad_norm": 0.7349845170974731,
      "learning_rate": 0.00019159142333800236,
      "loss": 3.3034,
      "step": 63300
    },
    {
      "epoch": 6.8184264341836185,
      "grad_norm": 0.7907821536064148,
      "learning_rate": 0.00019126818230794093,
      "loss": 3.2982,
      "step": 63350
    },
    {
      "epoch": 6.823807986223227,
      "grad_norm": 0.7795349359512329,
      "learning_rate": 0.00019094494127787953,
      "loss": 3.2983,
      "step": 63400
    },
    {
      "epoch": 6.829189538262835,
      "grad_norm": 0.7741080522537231,
      "learning_rate": 0.00019062170024781812,
      "loss": 3.3109,
      "step": 63450
    },
    {
      "epoch": 6.834571090302443,
      "grad_norm": 0.7275488376617432,
      "learning_rate": 0.0001902984592177567,
      "loss": 3.3103,
      "step": 63500
    },
    {
      "epoch": 6.839952642342052,
      "grad_norm": 0.7318737506866455,
      "learning_rate": 0.00018997521818769528,
      "loss": 3.291,
      "step": 63550
    },
    {
      "epoch": 6.8453341943816595,
      "grad_norm": 0.7193735837936401,
      "learning_rate": 0.00018965197715763385,
      "loss": 3.2967,
      "step": 63600
    },
    {
      "epoch": 6.850715746421268,
      "grad_norm": 0.7699055671691895,
      "learning_rate": 0.00018932873612757245,
      "loss": 3.3189,
      "step": 63650
    },
    {
      "epoch": 6.856097298460876,
      "grad_norm": 0.7138307094573975,
      "learning_rate": 0.00018900549509751104,
      "loss": 3.2882,
      "step": 63700
    },
    {
      "epoch": 6.861478850500484,
      "grad_norm": 0.7493799924850464,
      "learning_rate": 0.0001886822540674496,
      "loss": 3.2743,
      "step": 63750
    },
    {
      "epoch": 6.866860402540093,
      "grad_norm": 0.7282101511955261,
      "learning_rate": 0.00018835901303738818,
      "loss": 3.3036,
      "step": 63800
    },
    {
      "epoch": 6.8722419545797,
      "grad_norm": 0.741081714630127,
      "learning_rate": 0.0001880357720073268,
      "loss": 3.3124,
      "step": 63850
    },
    {
      "epoch": 6.877623506619309,
      "grad_norm": 0.7544240951538086,
      "learning_rate": 0.00018771253097726537,
      "loss": 3.2888,
      "step": 63900
    },
    {
      "epoch": 6.8830050586589175,
      "grad_norm": 0.7214697003364563,
      "learning_rate": 0.00018738928994720393,
      "loss": 3.2907,
      "step": 63950
    },
    {
      "epoch": 6.888386610698525,
      "grad_norm": 0.7240142226219177,
      "learning_rate": 0.00018706604891714255,
      "loss": 3.2935,
      "step": 64000
    },
    {
      "epoch": 6.888386610698525,
      "eval_accuracy": 0.3858205373732795,
      "eval_loss": 3.364917516708374,
      "eval_runtime": 185.4017,
      "eval_samples_per_second": 97.146,
      "eval_steps_per_second": 6.073,
      "step": 64000
    },
    {
      "epoch": 6.893768162738134,
      "grad_norm": 0.7279296517372131,
      "learning_rate": 0.00018674280788708112,
      "loss": 3.3074,
      "step": 64050
    },
    {
      "epoch": 6.899149714777742,
      "grad_norm": 0.7326666712760925,
      "learning_rate": 0.0001864195668570197,
      "loss": 3.289,
      "step": 64100
    },
    {
      "epoch": 6.90453126681735,
      "grad_norm": 0.809907078742981,
      "learning_rate": 0.00018609632582695828,
      "loss": 3.2873,
      "step": 64150
    },
    {
      "epoch": 6.9099128188569585,
      "grad_norm": 0.8244600296020508,
      "learning_rate": 0.00018577308479689685,
      "loss": 3.2871,
      "step": 64200
    },
    {
      "epoch": 6.915294370896566,
      "grad_norm": 0.7873415350914001,
      "learning_rate": 0.00018544984376683545,
      "loss": 3.3071,
      "step": 64250
    },
    {
      "epoch": 6.920675922936175,
      "grad_norm": 0.7484003901481628,
      "learning_rate": 0.00018513306755737525,
      "loss": 3.3141,
      "step": 64300
    },
    {
      "epoch": 6.926057474975783,
      "grad_norm": 0.7384822368621826,
      "learning_rate": 0.00018480982652731387,
      "loss": 3.2974,
      "step": 64350
    },
    {
      "epoch": 6.931439027015391,
      "grad_norm": 0.7646717429161072,
      "learning_rate": 0.00018448658549725244,
      "loss": 3.3105,
      "step": 64400
    },
    {
      "epoch": 6.9368205790549995,
      "grad_norm": 0.750130295753479,
      "learning_rate": 0.000184163344467191,
      "loss": 3.2869,
      "step": 64450
    },
    {
      "epoch": 6.942202131094608,
      "grad_norm": 0.7026450037956238,
      "learning_rate": 0.0001838401034371296,
      "loss": 3.2835,
      "step": 64500
    },
    {
      "epoch": 6.947583683134216,
      "grad_norm": 0.801478385925293,
      "learning_rate": 0.0001835168624070682,
      "loss": 3.3015,
      "step": 64550
    },
    {
      "epoch": 6.952965235173824,
      "grad_norm": 0.7395054697990417,
      "learning_rate": 0.00018319362137700677,
      "loss": 3.292,
      "step": 64600
    },
    {
      "epoch": 6.958346787213433,
      "grad_norm": 0.7504779100418091,
      "learning_rate": 0.00018287038034694536,
      "loss": 3.3003,
      "step": 64650
    },
    {
      "epoch": 6.9637283392530405,
      "grad_norm": 0.8172934651374817,
      "learning_rate": 0.00018254713931688393,
      "loss": 3.3101,
      "step": 64700
    },
    {
      "epoch": 6.969109891292649,
      "grad_norm": 0.7465377449989319,
      "learning_rate": 0.00018222389828682252,
      "loss": 3.2886,
      "step": 64750
    },
    {
      "epoch": 6.974491443332257,
      "grad_norm": 0.7690287232398987,
      "learning_rate": 0.00018190065725676112,
      "loss": 3.2974,
      "step": 64800
    },
    {
      "epoch": 6.979872995371865,
      "grad_norm": 0.7986275553703308,
      "learning_rate": 0.00018157741622669969,
      "loss": 3.3128,
      "step": 64850
    },
    {
      "epoch": 6.985254547411474,
      "grad_norm": 0.8031266927719116,
      "learning_rate": 0.00018125417519663825,
      "loss": 3.3073,
      "step": 64900
    },
    {
      "epoch": 6.990636099451081,
      "grad_norm": 0.7595334649085999,
      "learning_rate": 0.00018093093416657687,
      "loss": 3.2897,
      "step": 64950
    },
    {
      "epoch": 6.99601765149069,
      "grad_norm": 0.7986748814582825,
      "learning_rate": 0.00018060769313651544,
      "loss": 3.3002,
      "step": 65000
    },
    {
      "epoch": 6.99601765149069,
      "eval_accuracy": 0.3864837546004991,
      "eval_loss": 3.3594601154327393,
      "eval_runtime": 184.9532,
      "eval_samples_per_second": 97.381,
      "eval_steps_per_second": 6.088,
      "step": 65000
    },
    {
      "epoch": 7.0013992035302985,
      "grad_norm": 0.7555762529373169,
      "learning_rate": 0.000180284452106454,
      "loss": 3.2706,
      "step": 65050
    },
    {
      "epoch": 7.006780755569906,
      "grad_norm": 0.8189263343811035,
      "learning_rate": 0.00017996121107639263,
      "loss": 3.208,
      "step": 65100
    },
    {
      "epoch": 7.012162307609515,
      "grad_norm": 0.7540623545646667,
      "learning_rate": 0.0001796379700463312,
      "loss": 3.2201,
      "step": 65150
    },
    {
      "epoch": 7.017543859649122,
      "grad_norm": 0.7571452856063843,
      "learning_rate": 0.0001793147290162698,
      "loss": 3.2108,
      "step": 65200
    },
    {
      "epoch": 7.022925411688731,
      "grad_norm": 0.7571054100990295,
      "learning_rate": 0.00017899148798620836,
      "loss": 3.2136,
      "step": 65250
    },
    {
      "epoch": 7.0283069637283395,
      "grad_norm": 0.7304456830024719,
      "learning_rate": 0.00017866824695614696,
      "loss": 3.203,
      "step": 65300
    },
    {
      "epoch": 7.033688515767947,
      "grad_norm": 0.7756145000457764,
      "learning_rate": 0.00017834500592608555,
      "loss": 3.2158,
      "step": 65350
    },
    {
      "epoch": 7.039070067807556,
      "grad_norm": 0.7198467254638672,
      "learning_rate": 0.00017802176489602412,
      "loss": 3.2325,
      "step": 65400
    },
    {
      "epoch": 7.044451619847164,
      "grad_norm": 0.7484449148178101,
      "learning_rate": 0.00017769852386596269,
      "loss": 3.2122,
      "step": 65450
    },
    {
      "epoch": 7.049833171886772,
      "grad_norm": 0.7791109681129456,
      "learning_rate": 0.0001773752828359013,
      "loss": 3.2197,
      "step": 65500
    },
    {
      "epoch": 7.0552147239263805,
      "grad_norm": 0.7751621007919312,
      "learning_rate": 0.00017705204180583988,
      "loss": 3.2158,
      "step": 65550
    },
    {
      "epoch": 7.060596275965988,
      "grad_norm": 0.7578132748603821,
      "learning_rate": 0.00017672880077577844,
      "loss": 3.207,
      "step": 65600
    },
    {
      "epoch": 7.065977828005597,
      "grad_norm": 0.7406477332115173,
      "learning_rate": 0.00017640555974571704,
      "loss": 3.226,
      "step": 65650
    },
    {
      "epoch": 7.071359380045205,
      "grad_norm": 0.7612883448600769,
      "learning_rate": 0.00017608231871565563,
      "loss": 3.2372,
      "step": 65700
    },
    {
      "epoch": 7.076740932084813,
      "grad_norm": 0.7592723965644836,
      "learning_rate": 0.0001757590776855942,
      "loss": 3.2271,
      "step": 65750
    },
    {
      "epoch": 7.0821224841244215,
      "grad_norm": 0.7953776121139526,
      "learning_rate": 0.0001754358366555328,
      "loss": 3.2328,
      "step": 65800
    },
    {
      "epoch": 7.08750403616403,
      "grad_norm": 0.8319557309150696,
      "learning_rate": 0.00017511259562547136,
      "loss": 3.2146,
      "step": 65850
    },
    {
      "epoch": 7.092885588203638,
      "grad_norm": 0.7751454710960388,
      "learning_rate": 0.00017478935459540996,
      "loss": 3.2373,
      "step": 65900
    },
    {
      "epoch": 7.098267140243246,
      "grad_norm": 0.750540554523468,
      "learning_rate": 0.00017446611356534855,
      "loss": 3.2333,
      "step": 65950
    },
    {
      "epoch": 7.103648692282855,
      "grad_norm": 0.7501264810562134,
      "learning_rate": 0.00017414287253528712,
      "loss": 3.2116,
      "step": 66000
    },
    {
      "epoch": 7.103648692282855,
      "eval_accuracy": 0.38645626541988204,
      "eval_loss": 3.3664190769195557,
      "eval_runtime": 185.1354,
      "eval_samples_per_second": 97.286,
      "eval_steps_per_second": 6.082,
      "step": 66000
    },
    {
      "epoch": 7.109030244322462,
      "grad_norm": 0.7809863090515137,
      "learning_rate": 0.00017381963150522569,
      "loss": 3.2288,
      "step": 66050
    },
    {
      "epoch": 7.114411796362071,
      "grad_norm": 0.7352522611618042,
      "learning_rate": 0.0001734963904751643,
      "loss": 3.2301,
      "step": 66100
    },
    {
      "epoch": 7.119793348401679,
      "grad_norm": 0.826615571975708,
      "learning_rate": 0.00017317314944510288,
      "loss": 3.224,
      "step": 66150
    },
    {
      "epoch": 7.125174900441287,
      "grad_norm": 0.7712108492851257,
      "learning_rate": 0.00017284990841504147,
      "loss": 3.2378,
      "step": 66200
    },
    {
      "epoch": 7.130556452480896,
      "grad_norm": 0.7894640564918518,
      "learning_rate": 0.00017252666738498007,
      "loss": 3.2173,
      "step": 66250
    },
    {
      "epoch": 7.135938004520503,
      "grad_norm": 0.7729959487915039,
      "learning_rate": 0.00017220342635491863,
      "loss": 3.2145,
      "step": 66300
    },
    {
      "epoch": 7.141319556560112,
      "grad_norm": 0.7718258500099182,
      "learning_rate": 0.00017188018532485723,
      "loss": 3.2167,
      "step": 66350
    },
    {
      "epoch": 7.1467011085997205,
      "grad_norm": 0.7565211057662964,
      "learning_rate": 0.0001715569442947958,
      "loss": 3.2142,
      "step": 66400
    },
    {
      "epoch": 7.152082660639328,
      "grad_norm": 0.7619813084602356,
      "learning_rate": 0.0001712337032647344,
      "loss": 3.228,
      "step": 66450
    },
    {
      "epoch": 7.157464212678937,
      "grad_norm": 0.770587146282196,
      "learning_rate": 0.0001709169270552742,
      "loss": 3.2447,
      "step": 66500
    },
    {
      "epoch": 7.162845764718545,
      "grad_norm": 0.769239604473114,
      "learning_rate": 0.00017059368602521276,
      "loss": 3.2471,
      "step": 66550
    },
    {
      "epoch": 7.168227316758153,
      "grad_norm": 0.7729017734527588,
      "learning_rate": 0.00017027044499515139,
      "loss": 3.2385,
      "step": 66600
    },
    {
      "epoch": 7.1736088687977615,
      "grad_norm": 0.7612881660461426,
      "learning_rate": 0.00016994720396508995,
      "loss": 3.2299,
      "step": 66650
    },
    {
      "epoch": 7.178990420837369,
      "grad_norm": 0.769726037979126,
      "learning_rate": 0.00016962396293502852,
      "loss": 3.248,
      "step": 66700
    },
    {
      "epoch": 7.184371972876978,
      "grad_norm": 0.7710228562355042,
      "learning_rate": 0.00016930072190496714,
      "loss": 3.2452,
      "step": 66750
    },
    {
      "epoch": 7.189753524916586,
      "grad_norm": 0.7368969917297363,
      "learning_rate": 0.0001689774808749057,
      "loss": 3.2405,
      "step": 66800
    },
    {
      "epoch": 7.195135076956194,
      "grad_norm": 0.8005502223968506,
      "learning_rate": 0.00016865423984484428,
      "loss": 3.2305,
      "step": 66850
    },
    {
      "epoch": 7.2005166289958025,
      "grad_norm": 0.7608311772346497,
      "learning_rate": 0.00016833099881478287,
      "loss": 3.2266,
      "step": 66900
    },
    {
      "epoch": 7.205898181035411,
      "grad_norm": 0.8126056790351868,
      "learning_rate": 0.00016800775778472144,
      "loss": 3.2545,
      "step": 66950
    },
    {
      "epoch": 7.211279733075019,
      "grad_norm": 0.8470507860183716,
      "learning_rate": 0.00016768451675466006,
      "loss": 3.2481,
      "step": 67000
    },
    {
      "epoch": 7.211279733075019,
      "eval_accuracy": 0.3865580731757642,
      "eval_loss": 3.363801956176758,
      "eval_runtime": 185.4158,
      "eval_samples_per_second": 97.138,
      "eval_steps_per_second": 6.073,
      "step": 67000
    },
    {
      "epoch": 7.216661285114627,
      "grad_norm": 0.7487084269523621,
      "learning_rate": 0.00016736127572459863,
      "loss": 3.235,
      "step": 67050
    },
    {
      "epoch": 7.222042837154235,
      "grad_norm": 0.7650425434112549,
      "learning_rate": 0.0001670380346945372,
      "loss": 3.232,
      "step": 67100
    },
    {
      "epoch": 7.2274243891938434,
      "grad_norm": 0.8240507245063782,
      "learning_rate": 0.00016671479366447582,
      "loss": 3.2413,
      "step": 67150
    },
    {
      "epoch": 7.232805941233452,
      "grad_norm": 0.744361400604248,
      "learning_rate": 0.00016639155263441439,
      "loss": 3.2331,
      "step": 67200
    },
    {
      "epoch": 7.23818749327306,
      "grad_norm": 0.8013153672218323,
      "learning_rate": 0.00016606831160435295,
      "loss": 3.2352,
      "step": 67250
    },
    {
      "epoch": 7.243569045312668,
      "grad_norm": 0.7801939249038696,
      "learning_rate": 0.00016574507057429155,
      "loss": 3.2468,
      "step": 67300
    },
    {
      "epoch": 7.248950597352277,
      "grad_norm": 0.8023585677146912,
      "learning_rate": 0.00016542182954423014,
      "loss": 3.2463,
      "step": 67350
    },
    {
      "epoch": 7.254332149391884,
      "grad_norm": 0.7868350148200989,
      "learning_rate": 0.0001650985885141687,
      "loss": 3.2318,
      "step": 67400
    },
    {
      "epoch": 7.259713701431493,
      "grad_norm": 0.7524642944335938,
      "learning_rate": 0.0001647753474841073,
      "loss": 3.2487,
      "step": 67450
    },
    {
      "epoch": 7.265095253471101,
      "grad_norm": 0.7617383599281311,
      "learning_rate": 0.00016445210645404587,
      "loss": 3.2455,
      "step": 67500
    },
    {
      "epoch": 7.270476805510709,
      "grad_norm": 0.8090763688087463,
      "learning_rate": 0.00016412886542398447,
      "loss": 3.2386,
      "step": 67550
    },
    {
      "epoch": 7.275858357550318,
      "grad_norm": 0.816100537776947,
      "learning_rate": 0.00016380562439392306,
      "loss": 3.249,
      "step": 67600
    },
    {
      "epoch": 7.281239909589925,
      "grad_norm": 0.7546621561050415,
      "learning_rate": 0.00016348238336386163,
      "loss": 3.235,
      "step": 67650
    },
    {
      "epoch": 7.286621461629534,
      "grad_norm": 0.7477095127105713,
      "learning_rate": 0.0001631591423338002,
      "loss": 3.2628,
      "step": 67700
    },
    {
      "epoch": 7.2920030136691425,
      "grad_norm": 0.8014695644378662,
      "learning_rate": 0.00016283590130373882,
      "loss": 3.2325,
      "step": 67750
    },
    {
      "epoch": 7.29738456570875,
      "grad_norm": 0.8122795820236206,
      "learning_rate": 0.00016251266027367739,
      "loss": 3.243,
      "step": 67800
    },
    {
      "epoch": 7.302766117748359,
      "grad_norm": 0.7973216772079468,
      "learning_rate": 0.00016218941924361595,
      "loss": 3.2389,
      "step": 67850
    },
    {
      "epoch": 7.308147669787967,
      "grad_norm": 0.7753863334655762,
      "learning_rate": 0.00016186617821355458,
      "loss": 3.2439,
      "step": 67900
    },
    {
      "epoch": 7.313529221827575,
      "grad_norm": 0.7927606701850891,
      "learning_rate": 0.00016154293718349314,
      "loss": 3.2214,
      "step": 67950
    },
    {
      "epoch": 7.3189107738671835,
      "grad_norm": 0.7578996419906616,
      "learning_rate": 0.00016121969615343174,
      "loss": 3.2384,
      "step": 68000
    },
    {
      "epoch": 7.3189107738671835,
      "eval_accuracy": 0.38674984552275676,
      "eval_loss": 3.361203193664551,
      "eval_runtime": 184.8998,
      "eval_samples_per_second": 97.41,
      "eval_steps_per_second": 6.09,
      "step": 68000
    },
    {
      "epoch": 7.324292325906791,
      "grad_norm": 0.7464688420295715,
      "learning_rate": 0.0001608964551233703,
      "loss": 3.253,
      "step": 68050
    },
    {
      "epoch": 7.3296738779464,
      "grad_norm": 0.7672775983810425,
      "learning_rate": 0.0001605732140933089,
      "loss": 3.2448,
      "step": 68100
    },
    {
      "epoch": 7.335055429986008,
      "grad_norm": 0.8121482729911804,
      "learning_rate": 0.0001602499730632475,
      "loss": 3.2396,
      "step": 68150
    },
    {
      "epoch": 7.340436982025616,
      "grad_norm": 0.8000307083129883,
      "learning_rate": 0.00015992673203318606,
      "loss": 3.2417,
      "step": 68200
    },
    {
      "epoch": 7.3458185340652244,
      "grad_norm": 0.7772724628448486,
      "learning_rate": 0.00015960349100312463,
      "loss": 3.2447,
      "step": 68250
    },
    {
      "epoch": 7.351200086104833,
      "grad_norm": 0.7894703149795532,
      "learning_rate": 0.00015928024997306325,
      "loss": 3.2363,
      "step": 68300
    },
    {
      "epoch": 7.356581638144441,
      "grad_norm": 0.7807490229606628,
      "learning_rate": 0.00015895700894300182,
      "loss": 3.2406,
      "step": 68350
    },
    {
      "epoch": 7.361963190184049,
      "grad_norm": 0.778313398361206,
      "learning_rate": 0.00015863376791294039,
      "loss": 3.2426,
      "step": 68400
    },
    {
      "epoch": 7.367344742223658,
      "grad_norm": 0.7508113980293274,
      "learning_rate": 0.000158310526882879,
      "loss": 3.2643,
      "step": 68450
    },
    {
      "epoch": 7.372726294263265,
      "grad_norm": 0.8495672941207886,
      "learning_rate": 0.00015798728585281758,
      "loss": 3.2432,
      "step": 68500
    },
    {
      "epoch": 7.378107846302874,
      "grad_norm": 0.7832462191581726,
      "learning_rate": 0.00015766404482275614,
      "loss": 3.2389,
      "step": 68550
    },
    {
      "epoch": 7.383489398342482,
      "grad_norm": 0.787056565284729,
      "learning_rate": 0.00015734080379269474,
      "loss": 3.248,
      "step": 68600
    },
    {
      "epoch": 7.38887095038209,
      "grad_norm": 0.7693321704864502,
      "learning_rate": 0.00015701756276263333,
      "loss": 3.2463,
      "step": 68650
    },
    {
      "epoch": 7.394252502421699,
      "grad_norm": 0.7926502823829651,
      "learning_rate": 0.0001566943217325719,
      "loss": 3.2415,
      "step": 68700
    },
    {
      "epoch": 7.399634054461306,
      "grad_norm": 0.7598545551300049,
      "learning_rate": 0.0001563710807025105,
      "loss": 3.2528,
      "step": 68750
    },
    {
      "epoch": 7.405015606500915,
      "grad_norm": 0.7867632508277893,
      "learning_rate": 0.00015604783967244906,
      "loss": 3.2395,
      "step": 68800
    },
    {
      "epoch": 7.4103971585405235,
      "grad_norm": 0.8383733034133911,
      "learning_rate": 0.00015572459864238763,
      "loss": 3.2502,
      "step": 68850
    },
    {
      "epoch": 7.415778710580131,
      "grad_norm": 0.7990793585777283,
      "learning_rate": 0.00015540135761232625,
      "loss": 3.2436,
      "step": 68900
    },
    {
      "epoch": 7.42116026261974,
      "grad_norm": 0.7842612266540527,
      "learning_rate": 0.00015507811658226482,
      "loss": 3.2389,
      "step": 68950
    },
    {
      "epoch": 7.426541814659347,
      "grad_norm": 0.7659835815429688,
      "learning_rate": 0.00015475487555220344,
      "loss": 3.2446,
      "step": 69000
    },
    {
      "epoch": 7.426541814659347,
      "eval_accuracy": 0.3873940484946088,
      "eval_loss": 3.3569676876068115,
      "eval_runtime": 185.1154,
      "eval_samples_per_second": 97.296,
      "eval_steps_per_second": 6.083,
      "step": 69000
    },
    {
      "epoch": 7.431923366698956,
      "grad_norm": 0.8252059817314148,
      "learning_rate": 0.000154431634522142,
      "loss": 3.2347,
      "step": 69050
    },
    {
      "epoch": 7.4373049187385645,
      "grad_norm": 0.8039933443069458,
      "learning_rate": 0.00015410839349208058,
      "loss": 3.2622,
      "step": 69100
    },
    {
      "epoch": 7.442686470778172,
      "grad_norm": 0.830694854259491,
      "learning_rate": 0.00015378515246201917,
      "loss": 3.2617,
      "step": 69150
    },
    {
      "epoch": 7.448068022817781,
      "grad_norm": 0.7537750005722046,
      "learning_rate": 0.00015346191143195774,
      "loss": 3.2622,
      "step": 69200
    },
    {
      "epoch": 7.453449574857389,
      "grad_norm": 0.8499190807342529,
      "learning_rate": 0.00015313867040189633,
      "loss": 3.2529,
      "step": 69250
    },
    {
      "epoch": 7.458831126896997,
      "grad_norm": 0.7965125441551208,
      "learning_rate": 0.00015281542937183493,
      "loss": 3.2478,
      "step": 69300
    },
    {
      "epoch": 7.4642126789366054,
      "grad_norm": 0.7606722712516785,
      "learning_rate": 0.0001524921883417735,
      "loss": 3.2344,
      "step": 69350
    },
    {
      "epoch": 7.469594230976213,
      "grad_norm": 0.7247071266174316,
      "learning_rate": 0.00015216894731171206,
      "loss": 3.2432,
      "step": 69400
    },
    {
      "epoch": 7.474975783015822,
      "grad_norm": 0.7709882855415344,
      "learning_rate": 0.00015184570628165068,
      "loss": 3.2402,
      "step": 69450
    },
    {
      "epoch": 7.48035733505543,
      "grad_norm": 0.811995804309845,
      "learning_rate": 0.00015152246525158925,
      "loss": 3.2441,
      "step": 69500
    },
    {
      "epoch": 7.485738887095038,
      "grad_norm": 0.8299000263214111,
      "learning_rate": 0.00015119922422152782,
      "loss": 3.2408,
      "step": 69550
    },
    {
      "epoch": 7.491120439134646,
      "grad_norm": 0.7973170280456543,
      "learning_rate": 0.00015087598319146644,
      "loss": 3.2529,
      "step": 69600
    },
    {
      "epoch": 7.496501991174255,
      "grad_norm": 0.7989823222160339,
      "learning_rate": 0.000150552742161405,
      "loss": 3.2438,
      "step": 69650
    },
    {
      "epoch": 7.501883543213863,
      "grad_norm": 0.7758206725120544,
      "learning_rate": 0.00015022950113134358,
      "loss": 3.2342,
      "step": 69700
    },
    {
      "epoch": 7.507265095253471,
      "grad_norm": 0.7893825769424438,
      "learning_rate": 0.00014990626010128217,
      "loss": 3.2423,
      "step": 69750
    },
    {
      "epoch": 7.51264664729308,
      "grad_norm": 0.859170138835907,
      "learning_rate": 0.00014958301907122077,
      "loss": 3.2544,
      "step": 69800
    },
    {
      "epoch": 7.518028199332687,
      "grad_norm": 0.834900975227356,
      "learning_rate": 0.00014925977804115933,
      "loss": 3.2516,
      "step": 69850
    },
    {
      "epoch": 7.523409751372296,
      "grad_norm": 0.7640520930290222,
      "learning_rate": 0.00014893653701109793,
      "loss": 3.2327,
      "step": 69900
    },
    {
      "epoch": 7.528791303411904,
      "grad_norm": 0.7625937461853027,
      "learning_rate": 0.0001486132959810365,
      "loss": 3.2518,
      "step": 69950
    },
    {
      "epoch": 7.534172855451512,
      "grad_norm": 0.7990286946296692,
      "learning_rate": 0.0001482900549509751,
      "loss": 3.2569,
      "step": 70000
    },
    {
      "epoch": 7.534172855451512,
      "eval_accuracy": 0.3877212023398182,
      "eval_loss": 3.35430908203125,
      "eval_runtime": 185.3407,
      "eval_samples_per_second": 97.178,
      "eval_steps_per_second": 6.075,
      "step": 70000
    },
    {
      "epoch": 7.539554407491121,
      "grad_norm": 0.7802683115005493,
      "learning_rate": 0.00014796681392091368,
      "loss": 3.2547,
      "step": 70050
    },
    {
      "epoch": 7.544935959530728,
      "grad_norm": 0.7706187963485718,
      "learning_rate": 0.00014764357289085228,
      "loss": 3.2427,
      "step": 70100
    },
    {
      "epoch": 7.550317511570337,
      "grad_norm": 0.7758929133415222,
      "learning_rate": 0.00014732033186079085,
      "loss": 3.248,
      "step": 70150
    },
    {
      "epoch": 7.5556990636099455,
      "grad_norm": 0.8216681480407715,
      "learning_rate": 0.00014699709083072944,
      "loss": 3.2411,
      "step": 70200
    },
    {
      "epoch": 7.561080615649553,
      "grad_norm": 0.7994145750999451,
      "learning_rate": 0.00014667384980066804,
      "loss": 3.2475,
      "step": 70250
    },
    {
      "epoch": 7.566462167689162,
      "grad_norm": 0.7821169495582581,
      "learning_rate": 0.0001463506087706066,
      "loss": 3.2576,
      "step": 70300
    },
    {
      "epoch": 7.57184371972877,
      "grad_norm": 0.7666969299316406,
      "learning_rate": 0.0001460273677405452,
      "loss": 3.249,
      "step": 70350
    },
    {
      "epoch": 7.577225271768378,
      "grad_norm": 0.7849153280258179,
      "learning_rate": 0.00014570412671048377,
      "loss": 3.2498,
      "step": 70400
    },
    {
      "epoch": 7.5826068238079865,
      "grad_norm": 0.7996466159820557,
      "learning_rate": 0.00014538088568042236,
      "loss": 3.2652,
      "step": 70450
    },
    {
      "epoch": 7.587988375847594,
      "grad_norm": 0.7394452691078186,
      "learning_rate": 0.00014506410947096217,
      "loss": 3.2214,
      "step": 70500
    },
    {
      "epoch": 7.593369927887203,
      "grad_norm": 0.7737628221511841,
      "learning_rate": 0.00014474086844090076,
      "loss": 3.2429,
      "step": 70550
    },
    {
      "epoch": 7.598751479926811,
      "grad_norm": 0.7494852542877197,
      "learning_rate": 0.00014441762741083933,
      "loss": 3.2379,
      "step": 70600
    },
    {
      "epoch": 7.604133031966419,
      "grad_norm": 0.7984319925308228,
      "learning_rate": 0.00014409438638077792,
      "loss": 3.2456,
      "step": 70650
    },
    {
      "epoch": 7.609514584006027,
      "grad_norm": 0.775387167930603,
      "learning_rate": 0.00014377114535071652,
      "loss": 3.237,
      "step": 70700
    },
    {
      "epoch": 7.614896136045635,
      "grad_norm": 0.7979833483695984,
      "learning_rate": 0.0001434479043206551,
      "loss": 3.242,
      "step": 70750
    },
    {
      "epoch": 7.620277688085244,
      "grad_norm": 0.7707479596138,
      "learning_rate": 0.00014312466329059368,
      "loss": 3.2407,
      "step": 70800
    },
    {
      "epoch": 7.625659240124852,
      "grad_norm": 0.8784703612327576,
      "learning_rate": 0.00014280142226053225,
      "loss": 3.2422,
      "step": 70850
    },
    {
      "epoch": 7.63104079216446,
      "grad_norm": 0.7588596940040588,
      "learning_rate": 0.00014247818123047084,
      "loss": 3.2527,
      "step": 70900
    },
    {
      "epoch": 7.636422344204068,
      "grad_norm": 0.7865006923675537,
      "learning_rate": 0.0001421549402004094,
      "loss": 3.2623,
      "step": 70950
    },
    {
      "epoch": 7.641803896243677,
      "grad_norm": 0.8868777751922607,
      "learning_rate": 0.000141831699170348,
      "loss": 3.2258,
      "step": 71000
    },
    {
      "epoch": 7.641803896243677,
      "eval_accuracy": 0.3879447013300526,
      "eval_loss": 3.3508944511413574,
      "eval_runtime": 185.024,
      "eval_samples_per_second": 97.344,
      "eval_steps_per_second": 6.086,
      "step": 71000
    },
    {
      "epoch": 7.647185448283285,
      "grad_norm": 0.8430408239364624,
      "learning_rate": 0.0001415084581402866,
      "loss": 3.2356,
      "step": 71050
    },
    {
      "epoch": 7.652567000322893,
      "grad_norm": 0.7994721531867981,
      "learning_rate": 0.00014118521711022517,
      "loss": 3.2562,
      "step": 71100
    },
    {
      "epoch": 7.657948552362502,
      "grad_norm": 0.8080430030822754,
      "learning_rate": 0.00014086197608016376,
      "loss": 3.2319,
      "step": 71150
    },
    {
      "epoch": 7.663330104402109,
      "grad_norm": 0.78050696849823,
      "learning_rate": 0.00014054519987070357,
      "loss": 3.2483,
      "step": 71200
    },
    {
      "epoch": 7.668711656441718,
      "grad_norm": 0.7682746052742004,
      "learning_rate": 0.00014022195884064216,
      "loss": 3.2671,
      "step": 71250
    },
    {
      "epoch": 7.674093208481326,
      "grad_norm": 0.7863152027130127,
      "learning_rate": 0.00013989871781058076,
      "loss": 3.2429,
      "step": 71300
    },
    {
      "epoch": 7.679474760520934,
      "grad_norm": 0.781204879283905,
      "learning_rate": 0.00013957547678051933,
      "loss": 3.2523,
      "step": 71350
    },
    {
      "epoch": 7.684856312560543,
      "grad_norm": 0.8052288293838501,
      "learning_rate": 0.00013925223575045792,
      "loss": 3.2417,
      "step": 71400
    },
    {
      "epoch": 7.69023786460015,
      "grad_norm": 0.8121232986450195,
      "learning_rate": 0.0001389289947203965,
      "loss": 3.2539,
      "step": 71450
    },
    {
      "epoch": 7.695619416639759,
      "grad_norm": 0.8257960677146912,
      "learning_rate": 0.00013860575369033508,
      "loss": 3.2408,
      "step": 71500
    },
    {
      "epoch": 7.7010009686793675,
      "grad_norm": 0.7864553928375244,
      "learning_rate": 0.00013828251266027365,
      "loss": 3.2432,
      "step": 71550
    },
    {
      "epoch": 7.706382520718975,
      "grad_norm": 0.7899511456489563,
      "learning_rate": 0.00013795927163021224,
      "loss": 3.2443,
      "step": 71600
    },
    {
      "epoch": 7.711764072758584,
      "grad_norm": 0.7889692783355713,
      "learning_rate": 0.00013763603060015084,
      "loss": 3.2511,
      "step": 71650
    },
    {
      "epoch": 7.717145624798192,
      "grad_norm": 0.8366278409957886,
      "learning_rate": 0.0001373127895700894,
      "loss": 3.2603,
      "step": 71700
    },
    {
      "epoch": 7.7225271768378,
      "grad_norm": 0.7575733065605164,
      "learning_rate": 0.000136989548540028,
      "loss": 3.2452,
      "step": 71750
    },
    {
      "epoch": 7.727908728877408,
      "grad_norm": 0.7931903600692749,
      "learning_rate": 0.0001366663075099666,
      "loss": 3.2544,
      "step": 71800
    },
    {
      "epoch": 7.733290280917016,
      "grad_norm": 0.8185277581214905,
      "learning_rate": 0.0001363430664799052,
      "loss": 3.2342,
      "step": 71850
    },
    {
      "epoch": 7.738671832956625,
      "grad_norm": 0.7901844382286072,
      "learning_rate": 0.00013601982544984376,
      "loss": 3.2509,
      "step": 71900
    },
    {
      "epoch": 7.744053384996233,
      "grad_norm": 0.7859436869621277,
      "learning_rate": 0.00013569658441978233,
      "loss": 3.2586,
      "step": 71950
    },
    {
      "epoch": 7.749434937035841,
      "grad_norm": 0.7709542512893677,
      "learning_rate": 0.00013537334338972092,
      "loss": 3.258,
      "step": 72000
    },
    {
      "epoch": 7.749434937035841,
      "eval_accuracy": 0.3885651093194733,
      "eval_loss": 3.345294713973999,
      "eval_runtime": 185.0086,
      "eval_samples_per_second": 97.352,
      "eval_steps_per_second": 6.086,
      "step": 72000
    },
    {
      "epoch": 7.754816489075449,
      "grad_norm": 0.7761699557304382,
      "learning_rate": 0.0001350501023596595,
      "loss": 3.2518,
      "step": 72050
    },
    {
      "epoch": 7.760198041115058,
      "grad_norm": 0.8641573190689087,
      "learning_rate": 0.00013472686132959808,
      "loss": 3.2622,
      "step": 72100
    },
    {
      "epoch": 7.765579593154666,
      "grad_norm": 0.8142861127853394,
      "learning_rate": 0.00013440362029953668,
      "loss": 3.2465,
      "step": 72150
    },
    {
      "epoch": 7.770961145194274,
      "grad_norm": 0.7901236414909363,
      "learning_rate": 0.00013408037926947525,
      "loss": 3.2338,
      "step": 72200
    },
    {
      "epoch": 7.776342697233883,
      "grad_norm": 0.7913640737533569,
      "learning_rate": 0.00013375713823941384,
      "loss": 3.2631,
      "step": 72250
    },
    {
      "epoch": 7.78172424927349,
      "grad_norm": 0.828833818435669,
      "learning_rate": 0.00013343389720935243,
      "loss": 3.254,
      "step": 72300
    },
    {
      "epoch": 7.787105801313099,
      "grad_norm": 0.7752662301063538,
      "learning_rate": 0.00013311065617929103,
      "loss": 3.2609,
      "step": 72350
    },
    {
      "epoch": 7.792487353352707,
      "grad_norm": 0.8014473915100098,
      "learning_rate": 0.0001327874151492296,
      "loss": 3.2365,
      "step": 72400
    },
    {
      "epoch": 7.797868905392315,
      "grad_norm": 0.8082858920097351,
      "learning_rate": 0.0001324641741191682,
      "loss": 3.2367,
      "step": 72450
    },
    {
      "epoch": 7.803250457431924,
      "grad_norm": 0.7893218994140625,
      "learning_rate": 0.00013214093308910676,
      "loss": 3.2591,
      "step": 72500
    },
    {
      "epoch": 7.808632009471531,
      "grad_norm": 0.7564924955368042,
      "learning_rate": 0.00013181769205904535,
      "loss": 3.2503,
      "step": 72550
    },
    {
      "epoch": 7.81401356151114,
      "grad_norm": 0.8294974565505981,
      "learning_rate": 0.00013149445102898392,
      "loss": 3.2371,
      "step": 72600
    },
    {
      "epoch": 7.819395113550748,
      "grad_norm": 0.75737065076828,
      "learning_rate": 0.00013117120999892252,
      "loss": 3.244,
      "step": 72650
    },
    {
      "epoch": 7.824776665590356,
      "grad_norm": 0.8115005493164062,
      "learning_rate": 0.00013084796896886108,
      "loss": 3.2616,
      "step": 72700
    },
    {
      "epoch": 7.830158217629965,
      "grad_norm": 0.8291163444519043,
      "learning_rate": 0.00013052472793879968,
      "loss": 3.2593,
      "step": 72750
    },
    {
      "epoch": 7.835539769669572,
      "grad_norm": 0.8326108455657959,
      "learning_rate": 0.00013020148690873827,
      "loss": 3.2329,
      "step": 72800
    },
    {
      "epoch": 7.840921321709181,
      "grad_norm": 0.8050605654716492,
      "learning_rate": 0.00012987824587867687,
      "loss": 3.2507,
      "step": 72850
    },
    {
      "epoch": 7.846302873748789,
      "grad_norm": 0.8034680485725403,
      "learning_rate": 0.00012955500484861543,
      "loss": 3.249,
      "step": 72900
    },
    {
      "epoch": 7.851684425788397,
      "grad_norm": 0.8438678979873657,
      "learning_rate": 0.00012923176381855403,
      "loss": 3.2675,
      "step": 72950
    },
    {
      "epoch": 7.857065977828006,
      "grad_norm": 0.7695510983467102,
      "learning_rate": 0.00012890852278849262,
      "loss": 3.2484,
      "step": 73000
    },
    {
      "epoch": 7.857065977828006,
      "eval_accuracy": 0.3888908506771411,
      "eval_loss": 3.3417837619781494,
      "eval_runtime": 185.3539,
      "eval_samples_per_second": 97.171,
      "eval_steps_per_second": 6.075,
      "step": 73000
    },
    {
      "epoch": 7.862447529867614,
      "grad_norm": 0.7945556640625,
      "learning_rate": 0.0001285852817584312,
      "loss": 3.2463,
      "step": 73050
    },
    {
      "epoch": 7.867829081907222,
      "grad_norm": 0.799061119556427,
      "learning_rate": 0.0001282620407283698,
      "loss": 3.2555,
      "step": 73100
    },
    {
      "epoch": 7.87321063394683,
      "grad_norm": 0.7596940994262695,
      "learning_rate": 0.00012793879969830835,
      "loss": 3.247,
      "step": 73150
    },
    {
      "epoch": 7.878592185986438,
      "grad_norm": 2.275698661804199,
      "learning_rate": 0.00012761555866824695,
      "loss": 3.2465,
      "step": 73200
    },
    {
      "epoch": 7.883973738026047,
      "grad_norm": 0.8105242848396301,
      "learning_rate": 0.00012729231763818552,
      "loss": 3.2471,
      "step": 73250
    },
    {
      "epoch": 7.889355290065655,
      "grad_norm": 0.7951189279556274,
      "learning_rate": 0.0001269690766081241,
      "loss": 3.2292,
      "step": 73300
    },
    {
      "epoch": 7.894736842105263,
      "grad_norm": 0.7983742356300354,
      "learning_rate": 0.0001266458355780627,
      "loss": 3.2344,
      "step": 73350
    },
    {
      "epoch": 7.900118394144871,
      "grad_norm": 0.8250595927238464,
      "learning_rate": 0.00012632259454800127,
      "loss": 3.2616,
      "step": 73400
    },
    {
      "epoch": 7.90549994618448,
      "grad_norm": 0.7646833658218384,
      "learning_rate": 0.00012599935351793987,
      "loss": 3.253,
      "step": 73450
    },
    {
      "epoch": 7.910881498224088,
      "grad_norm": 0.8301563858985901,
      "learning_rate": 0.00012567611248787846,
      "loss": 3.2567,
      "step": 73500
    },
    {
      "epoch": 7.916263050263696,
      "grad_norm": 0.8209587335586548,
      "learning_rate": 0.00012535287145781703,
      "loss": 3.2604,
      "step": 73550
    },
    {
      "epoch": 7.921644602303305,
      "grad_norm": 0.8052746057510376,
      "learning_rate": 0.00012502963042775562,
      "loss": 3.2516,
      "step": 73600
    },
    {
      "epoch": 7.927026154342912,
      "grad_norm": 0.8003509044647217,
      "learning_rate": 0.00012470638939769422,
      "loss": 3.2537,
      "step": 73650
    },
    {
      "epoch": 7.932407706382521,
      "grad_norm": 0.8209139108657837,
      "learning_rate": 0.0001243831483676328,
      "loss": 3.2355,
      "step": 73700
    },
    {
      "epoch": 7.937789258422129,
      "grad_norm": 0.815592348575592,
      "learning_rate": 0.00012405990733757138,
      "loss": 3.2564,
      "step": 73750
    },
    {
      "epoch": 7.943170810461737,
      "grad_norm": 0.8316048979759216,
      "learning_rate": 0.00012373666630750995,
      "loss": 3.2334,
      "step": 73800
    },
    {
      "epoch": 7.948552362501346,
      "grad_norm": 0.8341267108917236,
      "learning_rate": 0.00012341989009804976,
      "loss": 3.2419,
      "step": 73850
    },
    {
      "epoch": 7.953933914540953,
      "grad_norm": 0.8219509720802307,
      "learning_rate": 0.00012309664906798835,
      "loss": 3.2582,
      "step": 73900
    },
    {
      "epoch": 7.959315466580562,
      "grad_norm": 0.851315438747406,
      "learning_rate": 0.00012277340803792694,
      "loss": 3.2566,
      "step": 73950
    },
    {
      "epoch": 7.96469701862017,
      "grad_norm": 0.7956839799880981,
      "learning_rate": 0.0001224501670078655,
      "loss": 3.252,
      "step": 74000
    },
    {
      "epoch": 7.96469701862017,
      "eval_accuracy": 0.38924940520692886,
      "eval_loss": 3.338170051574707,
      "eval_runtime": 184.8447,
      "eval_samples_per_second": 97.439,
      "eval_steps_per_second": 6.092,
      "step": 74000
    },
    {
      "epoch": 7.970078570659778,
      "grad_norm": 0.7973542213439941,
      "learning_rate": 0.0001221269259778041,
      "loss": 3.2648,
      "step": 74050
    },
    {
      "epoch": 7.975460122699387,
      "grad_norm": 0.8095277547836304,
      "learning_rate": 0.00012180368494774269,
      "loss": 3.2483,
      "step": 74100
    },
    {
      "epoch": 7.980841674738995,
      "grad_norm": 0.8348960280418396,
      "learning_rate": 0.00012148044391768128,
      "loss": 3.2434,
      "step": 74150
    },
    {
      "epoch": 7.986223226778603,
      "grad_norm": 0.822513222694397,
      "learning_rate": 0.00012115720288761985,
      "loss": 3.2565,
      "step": 74200
    },
    {
      "epoch": 7.991604778818211,
      "grad_norm": 0.809738278388977,
      "learning_rate": 0.00012083396185755845,
      "loss": 3.2653,
      "step": 74250
    },
    {
      "epoch": 7.996986330857819,
      "grad_norm": 0.8194822669029236,
      "learning_rate": 0.00012051072082749703,
      "loss": 3.2475,
      "step": 74300
    },
    {
      "epoch": 8.002367882897428,
      "grad_norm": 0.7727041244506836,
      "learning_rate": 0.00012018747979743561,
      "loss": 3.2203,
      "step": 74350
    },
    {
      "epoch": 8.007749434937036,
      "grad_norm": 0.807702898979187,
      "learning_rate": 0.00011986423876737419,
      "loss": 3.1708,
      "step": 74400
    },
    {
      "epoch": 8.013130986976645,
      "grad_norm": 0.8706448674201965,
      "learning_rate": 0.00011954099773731278,
      "loss": 3.1823,
      "step": 74450
    },
    {
      "epoch": 8.018512539016251,
      "grad_norm": 0.8152173161506653,
      "learning_rate": 0.00011921775670725135,
      "loss": 3.1745,
      "step": 74500
    },
    {
      "epoch": 8.02389409105586,
      "grad_norm": 0.8083778619766235,
      "learning_rate": 0.00011889451567718995,
      "loss": 3.1763,
      "step": 74550
    },
    {
      "epoch": 8.029275643095469,
      "grad_norm": 0.7975329756736755,
      "learning_rate": 0.00011857127464712854,
      "loss": 3.1721,
      "step": 74600
    },
    {
      "epoch": 8.034657195135077,
      "grad_norm": 0.8000638484954834,
      "learning_rate": 0.00011824803361706712,
      "loss": 3.1574,
      "step": 74650
    },
    {
      "epoch": 8.040038747174686,
      "grad_norm": 0.7966259717941284,
      "learning_rate": 0.00011792479258700569,
      "loss": 3.1731,
      "step": 74700
    },
    {
      "epoch": 8.045420299214294,
      "grad_norm": 0.8433051109313965,
      "learning_rate": 0.00011760155155694428,
      "loss": 3.1935,
      "step": 74750
    },
    {
      "epoch": 8.050801851253901,
      "grad_norm": 0.8051315546035767,
      "learning_rate": 0.00011727831052688288,
      "loss": 3.167,
      "step": 74800
    },
    {
      "epoch": 8.05618340329351,
      "grad_norm": 0.7908819913864136,
      "learning_rate": 0.00011695506949682145,
      "loss": 3.1756,
      "step": 74850
    },
    {
      "epoch": 8.061564955333118,
      "grad_norm": 0.8369430899620056,
      "learning_rate": 0.00011663182846676004,
      "loss": 3.1801,
      "step": 74900
    },
    {
      "epoch": 8.066946507372727,
      "grad_norm": 0.8394438624382019,
      "learning_rate": 0.00011630858743669862,
      "loss": 3.178,
      "step": 74950
    },
    {
      "epoch": 8.072328059412335,
      "grad_norm": 0.8868573307991028,
      "learning_rate": 0.0001159853464066372,
      "loss": 3.1996,
      "step": 75000
    },
    {
      "epoch": 8.072328059412335,
      "eval_accuracy": 0.3889214907915048,
      "eval_loss": 3.3451037406921387,
      "eval_runtime": 184.9954,
      "eval_samples_per_second": 97.359,
      "eval_steps_per_second": 6.087,
      "step": 75000
    },
    {
      "epoch": 8.077709611451942,
      "grad_norm": 0.8365190625190735,
      "learning_rate": 0.00011566210537657578,
      "loss": 3.1941,
      "step": 75050
    },
    {
      "epoch": 8.08309116349155,
      "grad_norm": 0.8807831406593323,
      "learning_rate": 0.00011533886434651438,
      "loss": 3.1722,
      "step": 75100
    },
    {
      "epoch": 8.088472715531159,
      "grad_norm": 0.8207399845123291,
      "learning_rate": 0.00011501562331645296,
      "loss": 3.1795,
      "step": 75150
    },
    {
      "epoch": 8.093854267570768,
      "grad_norm": 0.8633342981338501,
      "learning_rate": 0.00011469238228639154,
      "loss": 3.1748,
      "step": 75200
    },
    {
      "epoch": 8.099235819610376,
      "grad_norm": 0.8366163372993469,
      "learning_rate": 0.00011436914125633012,
      "loss": 3.1891,
      "step": 75250
    },
    {
      "epoch": 8.104617371649983,
      "grad_norm": 0.782554566860199,
      "learning_rate": 0.00011404590022626872,
      "loss": 3.1679,
      "step": 75300
    },
    {
      "epoch": 8.109998923689592,
      "grad_norm": 0.8265233039855957,
      "learning_rate": 0.00011372265919620728,
      "loss": 3.2064,
      "step": 75350
    },
    {
      "epoch": 8.1153804757292,
      "grad_norm": 0.838972270488739,
      "learning_rate": 0.00011339941816614588,
      "loss": 3.1863,
      "step": 75400
    },
    {
      "epoch": 8.120762027768809,
      "grad_norm": 0.8085297346115112,
      "learning_rate": 0.00011307617713608447,
      "loss": 3.1928,
      "step": 75450
    },
    {
      "epoch": 8.126143579808417,
      "grad_norm": 0.8338049650192261,
      "learning_rate": 0.00011275293610602305,
      "loss": 3.1964,
      "step": 75500
    },
    {
      "epoch": 8.131525131848026,
      "grad_norm": 0.8223406672477722,
      "learning_rate": 0.00011242969507596164,
      "loss": 3.1807,
      "step": 75550
    },
    {
      "epoch": 8.136906683887632,
      "grad_norm": 0.7878255248069763,
      "learning_rate": 0.00011210645404590022,
      "loss": 3.1909,
      "step": 75600
    },
    {
      "epoch": 8.142288235927241,
      "grad_norm": 0.8164534568786621,
      "learning_rate": 0.00011178321301583881,
      "loss": 3.1731,
      "step": 75650
    },
    {
      "epoch": 8.14766978796685,
      "grad_norm": 0.8075200319290161,
      "learning_rate": 0.00011145997198577738,
      "loss": 3.1762,
      "step": 75700
    },
    {
      "epoch": 8.153051340006458,
      "grad_norm": 0.8299018740653992,
      "learning_rate": 0.00011113673095571597,
      "loss": 3.1801,
      "step": 75750
    },
    {
      "epoch": 8.158432892046067,
      "grad_norm": 0.8242932558059692,
      "learning_rate": 0.00011081348992565455,
      "loss": 3.1716,
      "step": 75800
    },
    {
      "epoch": 8.163814444085673,
      "grad_norm": 0.8051115870475769,
      "learning_rate": 0.00011049024889559314,
      "loss": 3.1899,
      "step": 75850
    },
    {
      "epoch": 8.169195996125282,
      "grad_norm": 0.8188024759292603,
      "learning_rate": 0.00011016700786553172,
      "loss": 3.1748,
      "step": 75900
    },
    {
      "epoch": 8.17457754816489,
      "grad_norm": 0.8170576691627502,
      "learning_rate": 0.00010984376683547031,
      "loss": 3.1903,
      "step": 75950
    },
    {
      "epoch": 8.1799591002045,
      "grad_norm": 0.8629337549209595,
      "learning_rate": 0.00010952052580540889,
      "loss": 3.1725,
      "step": 76000
    },
    {
      "epoch": 8.1799591002045,
      "eval_accuracy": 0.3895448324088965,
      "eval_loss": 3.342695713043213,
      "eval_runtime": 185.307,
      "eval_samples_per_second": 97.195,
      "eval_steps_per_second": 6.076,
      "step": 76000
    },
    {
      "epoch": 8.185340652244108,
      "grad_norm": 0.8078259229660034,
      "learning_rate": 0.00010919728477534747,
      "loss": 3.1766,
      "step": 76050
    },
    {
      "epoch": 8.190722204283716,
      "grad_norm": 0.8017242550849915,
      "learning_rate": 0.00010887404374528605,
      "loss": 3.1857,
      "step": 76100
    },
    {
      "epoch": 8.196103756323323,
      "grad_norm": 0.8516344428062439,
      "learning_rate": 0.00010855080271522465,
      "loss": 3.1847,
      "step": 76150
    },
    {
      "epoch": 8.201485308362932,
      "grad_norm": 0.8172838687896729,
      "learning_rate": 0.00010822756168516322,
      "loss": 3.1916,
      "step": 76200
    },
    {
      "epoch": 8.20686686040254,
      "grad_norm": 0.8483747839927673,
      "learning_rate": 0.00010790432065510181,
      "loss": 3.1919,
      "step": 76250
    },
    {
      "epoch": 8.212248412442149,
      "grad_norm": 0.8249405026435852,
      "learning_rate": 0.0001075810796250404,
      "loss": 3.1763,
      "step": 76300
    },
    {
      "epoch": 8.217629964481757,
      "grad_norm": 0.8060430288314819,
      "learning_rate": 0.00010725783859497897,
      "loss": 3.1807,
      "step": 76350
    },
    {
      "epoch": 8.223011516521364,
      "grad_norm": 0.8632802963256836,
      "learning_rate": 0.00010693459756491757,
      "loss": 3.1873,
      "step": 76400
    },
    {
      "epoch": 8.228393068560973,
      "grad_norm": 0.8333990573883057,
      "learning_rate": 0.00010661135653485615,
      "loss": 3.1846,
      "step": 76450
    },
    {
      "epoch": 8.233774620600581,
      "grad_norm": 0.7996773719787598,
      "learning_rate": 0.00010628811550479474,
      "loss": 3.1755,
      "step": 76500
    },
    {
      "epoch": 8.23915617264019,
      "grad_norm": 0.7984475493431091,
      "learning_rate": 0.00010596487447473331,
      "loss": 3.1737,
      "step": 76550
    },
    {
      "epoch": 8.244537724679798,
      "grad_norm": 0.8256939649581909,
      "learning_rate": 0.0001056416334446719,
      "loss": 3.2028,
      "step": 76600
    },
    {
      "epoch": 8.249919276719407,
      "grad_norm": 0.8198964595794678,
      "learning_rate": 0.00010531839241461049,
      "loss": 3.1867,
      "step": 76650
    },
    {
      "epoch": 8.255300828759013,
      "grad_norm": 0.8413407206535339,
      "learning_rate": 0.00010499515138454907,
      "loss": 3.1989,
      "step": 76700
    },
    {
      "epoch": 8.260682380798622,
      "grad_norm": 0.8061610460281372,
      "learning_rate": 0.00010467191035448765,
      "loss": 3.1919,
      "step": 76750
    },
    {
      "epoch": 8.26606393283823,
      "grad_norm": 0.8221855759620667,
      "learning_rate": 0.00010434866932442624,
      "loss": 3.1774,
      "step": 76800
    },
    {
      "epoch": 8.27144548487784,
      "grad_norm": 0.8132125735282898,
      "learning_rate": 0.00010402542829436481,
      "loss": 3.2062,
      "step": 76850
    },
    {
      "epoch": 8.276827036917448,
      "grad_norm": 0.8057106137275696,
      "learning_rate": 0.0001037021872643034,
      "loss": 3.1941,
      "step": 76900
    },
    {
      "epoch": 8.282208588957054,
      "grad_norm": 0.8262264728546143,
      "learning_rate": 0.00010337894623424199,
      "loss": 3.2045,
      "step": 76950
    },
    {
      "epoch": 8.287590140996663,
      "grad_norm": 0.8277431130409241,
      "learning_rate": 0.00010305570520418058,
      "loss": 3.2013,
      "step": 77000
    },
    {
      "epoch": 8.287590140996663,
      "eval_accuracy": 0.38962567015743044,
      "eval_loss": 3.3404276371002197,
      "eval_runtime": 184.8309,
      "eval_samples_per_second": 97.446,
      "eval_steps_per_second": 6.092,
      "step": 77000
    },
    {
      "epoch": 8.292971693036272,
      "grad_norm": 0.7952620983123779,
      "learning_rate": 0.00010273246417411915,
      "loss": 3.2074,
      "step": 77050
    },
    {
      "epoch": 8.29835324507588,
      "grad_norm": 0.8601248264312744,
      "learning_rate": 0.00010240922314405774,
      "loss": 3.1882,
      "step": 77100
    },
    {
      "epoch": 8.303734797115489,
      "grad_norm": 0.8398854732513428,
      "learning_rate": 0.00010208598211399634,
      "loss": 3.1704,
      "step": 77150
    },
    {
      "epoch": 8.309116349155097,
      "grad_norm": 0.7851851582527161,
      "learning_rate": 0.0001017627410839349,
      "loss": 3.1598,
      "step": 77200
    },
    {
      "epoch": 8.314497901194704,
      "grad_norm": 0.8402460217475891,
      "learning_rate": 0.0001014395000538735,
      "loss": 3.1974,
      "step": 77250
    },
    {
      "epoch": 8.319879453234313,
      "grad_norm": 0.8805944919586182,
      "learning_rate": 0.00010111625902381208,
      "loss": 3.2049,
      "step": 77300
    },
    {
      "epoch": 8.325261005273921,
      "grad_norm": 0.808759331703186,
      "learning_rate": 0.00010079301799375066,
      "loss": 3.1966,
      "step": 77350
    },
    {
      "epoch": 8.33064255731353,
      "grad_norm": 0.8168617486953735,
      "learning_rate": 0.00010046977696368924,
      "loss": 3.189,
      "step": 77400
    },
    {
      "epoch": 8.336024109353138,
      "grad_norm": 0.836647629737854,
      "learning_rate": 0.00010014653593362784,
      "loss": 3.1907,
      "step": 77450
    },
    {
      "epoch": 8.341405661392745,
      "grad_norm": 0.8394541144371033,
      "learning_rate": 9.982329490356642e-05,
      "loss": 3.1942,
      "step": 77500
    },
    {
      "epoch": 8.346787213432354,
      "grad_norm": 0.8457735776901245,
      "learning_rate": 9.9500053873505e-05,
      "loss": 3.1805,
      "step": 77550
    },
    {
      "epoch": 8.352168765471962,
      "grad_norm": 0.81809401512146,
      "learning_rate": 9.917681284344358e-05,
      "loss": 3.203,
      "step": 77600
    },
    {
      "epoch": 8.35755031751157,
      "grad_norm": 0.8279471397399902,
      "learning_rate": 9.885357181338218e-05,
      "loss": 3.1928,
      "step": 77650
    },
    {
      "epoch": 8.36293186955118,
      "grad_norm": 0.852434515953064,
      "learning_rate": 9.853033078332074e-05,
      "loss": 3.1887,
      "step": 77700
    },
    {
      "epoch": 8.368313421590786,
      "grad_norm": 0.8403735756874084,
      "learning_rate": 9.820708975325934e-05,
      "loss": 3.1961,
      "step": 77750
    },
    {
      "epoch": 8.373694973630395,
      "grad_norm": 0.8798422813415527,
      "learning_rate": 9.788384872319793e-05,
      "loss": 3.1994,
      "step": 77800
    },
    {
      "epoch": 8.379076525670003,
      "grad_norm": 0.8449347019195557,
      "learning_rate": 9.756707251373773e-05,
      "loss": 3.191,
      "step": 77850
    },
    {
      "epoch": 8.384458077709612,
      "grad_norm": 0.7985642552375793,
      "learning_rate": 9.724383148367632e-05,
      "loss": 3.1959,
      "step": 77900
    },
    {
      "epoch": 8.38983962974922,
      "grad_norm": 0.8204814195632935,
      "learning_rate": 9.692059045361492e-05,
      "loss": 3.1965,
      "step": 77950
    },
    {
      "epoch": 8.395221181788829,
      "grad_norm": 0.816024124622345,
      "learning_rate": 9.659734942355348e-05,
      "loss": 3.1951,
      "step": 78000
    },
    {
      "epoch": 8.395221181788829,
      "eval_accuracy": 0.3899985668684097,
      "eval_loss": 3.33735728263855,
      "eval_runtime": 185.0,
      "eval_samples_per_second": 97.357,
      "eval_steps_per_second": 6.086,
      "step": 78000
    },
    {
      "epoch": 8.400602733828435,
      "grad_norm": 0.8745648860931396,
      "learning_rate": 9.627410839349208e-05,
      "loss": 3.2011,
      "step": 78050
    },
    {
      "epoch": 8.405984285868044,
      "grad_norm": 0.8315465450286865,
      "learning_rate": 9.595086736343066e-05,
      "loss": 3.1825,
      "step": 78100
    },
    {
      "epoch": 8.411365837907653,
      "grad_norm": 0.8296406269073486,
      "learning_rate": 9.562762633336923e-05,
      "loss": 3.1923,
      "step": 78150
    },
    {
      "epoch": 8.416747389947261,
      "grad_norm": 0.8062648773193359,
      "learning_rate": 9.530438530330782e-05,
      "loss": 3.1981,
      "step": 78200
    },
    {
      "epoch": 8.42212894198687,
      "grad_norm": 0.8563634157180786,
      "learning_rate": 9.498114427324642e-05,
      "loss": 3.2034,
      "step": 78250
    },
    {
      "epoch": 8.427510494026476,
      "grad_norm": 0.7938030958175659,
      "learning_rate": 9.4657903243185e-05,
      "loss": 3.1934,
      "step": 78300
    },
    {
      "epoch": 8.432892046066085,
      "grad_norm": 0.8524783253669739,
      "learning_rate": 9.433466221312358e-05,
      "loss": 3.2016,
      "step": 78350
    },
    {
      "epoch": 8.438273598105694,
      "grad_norm": 0.8547974824905396,
      "learning_rate": 9.401142118306216e-05,
      "loss": 3.2003,
      "step": 78400
    },
    {
      "epoch": 8.443655150145302,
      "grad_norm": 0.8259228467941284,
      "learning_rate": 9.368818015300075e-05,
      "loss": 3.2012,
      "step": 78450
    },
    {
      "epoch": 8.44903670218491,
      "grad_norm": 0.8110027313232422,
      "learning_rate": 9.336493912293932e-05,
      "loss": 3.2025,
      "step": 78500
    },
    {
      "epoch": 8.45441825422452,
      "grad_norm": 0.8531944155693054,
      "learning_rate": 9.304169809287792e-05,
      "loss": 3.1994,
      "step": 78550
    },
    {
      "epoch": 8.459799806264126,
      "grad_norm": 0.8652747273445129,
      "learning_rate": 9.27184570628165e-05,
      "loss": 3.2063,
      "step": 78600
    },
    {
      "epoch": 8.465181358303735,
      "grad_norm": 0.816554069519043,
      "learning_rate": 9.239521603275508e-05,
      "loss": 3.2037,
      "step": 78650
    },
    {
      "epoch": 8.470562910343343,
      "grad_norm": 0.8304002285003662,
      "learning_rate": 9.207197500269366e-05,
      "loss": 3.1934,
      "step": 78700
    },
    {
      "epoch": 8.475944462382952,
      "grad_norm": 0.8607257604598999,
      "learning_rate": 9.174873397263225e-05,
      "loss": 3.1963,
      "step": 78750
    },
    {
      "epoch": 8.48132601442256,
      "grad_norm": 0.8093053102493286,
      "learning_rate": 9.142549294257085e-05,
      "loss": 3.1927,
      "step": 78800
    },
    {
      "epoch": 8.486707566462167,
      "grad_norm": 0.8152172565460205,
      "learning_rate": 9.110225191250942e-05,
      "loss": 3.1925,
      "step": 78850
    },
    {
      "epoch": 8.492089118501776,
      "grad_norm": 0.8187666535377502,
      "learning_rate": 9.077901088244801e-05,
      "loss": 3.2055,
      "step": 78900
    },
    {
      "epoch": 8.497470670541384,
      "grad_norm": 0.8896129727363586,
      "learning_rate": 9.045576985238659e-05,
      "loss": 3.1898,
      "step": 78950
    },
    {
      "epoch": 8.502852222580993,
      "grad_norm": 0.853837788105011,
      "learning_rate": 9.013252882232517e-05,
      "loss": 3.1936,
      "step": 79000
    },
    {
      "epoch": 8.502852222580993,
      "eval_accuracy": 0.39012786380490894,
      "eval_loss": 3.33323335647583,
      "eval_runtime": 185.2565,
      "eval_samples_per_second": 97.222,
      "eval_steps_per_second": 6.078,
      "step": 79000
    },
    {
      "epoch": 8.508233774620601,
      "grad_norm": 0.7914009690284729,
      "learning_rate": 8.980928779226375e-05,
      "loss": 3.1733,
      "step": 79050
    },
    {
      "epoch": 8.513615326660208,
      "grad_norm": 0.8368597030639648,
      "learning_rate": 8.948604676220235e-05,
      "loss": 3.2029,
      "step": 79100
    },
    {
      "epoch": 8.518996878699816,
      "grad_norm": 0.7963562607765198,
      "learning_rate": 8.916280573214092e-05,
      "loss": 3.1999,
      "step": 79150
    },
    {
      "epoch": 8.524378430739425,
      "grad_norm": 0.8621082305908203,
      "learning_rate": 8.883956470207951e-05,
      "loss": 3.2058,
      "step": 79200
    },
    {
      "epoch": 8.529759982779034,
      "grad_norm": 0.826045036315918,
      "learning_rate": 8.851632367201809e-05,
      "loss": 3.1977,
      "step": 79250
    },
    {
      "epoch": 8.535141534818642,
      "grad_norm": 0.8315468430519104,
      "learning_rate": 8.819308264195669e-05,
      "loss": 3.1943,
      "step": 79300
    },
    {
      "epoch": 8.54052308685825,
      "grad_norm": 0.7864601016044617,
      "learning_rate": 8.786984161189526e-05,
      "loss": 3.187,
      "step": 79350
    },
    {
      "epoch": 8.545904638897857,
      "grad_norm": 0.8104679584503174,
      "learning_rate": 8.754660058183385e-05,
      "loss": 3.203,
      "step": 79400
    },
    {
      "epoch": 8.551286190937466,
      "grad_norm": 0.8404777646064758,
      "learning_rate": 8.722335955177243e-05,
      "loss": 3.2129,
      "step": 79450
    },
    {
      "epoch": 8.556667742977075,
      "grad_norm": 0.8269492387771606,
      "learning_rate": 8.690011852171101e-05,
      "loss": 3.2071,
      "step": 79500
    },
    {
      "epoch": 8.562049295016683,
      "grad_norm": 0.8499221205711365,
      "learning_rate": 8.657687749164959e-05,
      "loss": 3.2035,
      "step": 79550
    },
    {
      "epoch": 8.567430847056292,
      "grad_norm": 0.8582130670547485,
      "learning_rate": 8.625363646158819e-05,
      "loss": 3.2073,
      "step": 79600
    },
    {
      "epoch": 8.572812399095898,
      "grad_norm": 0.8317336440086365,
      "learning_rate": 8.593039543152676e-05,
      "loss": 3.1962,
      "step": 79650
    },
    {
      "epoch": 8.578193951135507,
      "grad_norm": 0.8151216506958008,
      "learning_rate": 8.560715440146535e-05,
      "loss": 3.1991,
      "step": 79700
    },
    {
      "epoch": 8.583575503175116,
      "grad_norm": 0.8646751642227173,
      "learning_rate": 8.528391337140394e-05,
      "loss": 3.1884,
      "step": 79750
    },
    {
      "epoch": 8.588957055214724,
      "grad_norm": 0.8245811462402344,
      "learning_rate": 8.496067234134253e-05,
      "loss": 3.2016,
      "step": 79800
    },
    {
      "epoch": 8.594338607254333,
      "grad_norm": 0.8000688552856445,
      "learning_rate": 8.46374313112811e-05,
      "loss": 3.1916,
      "step": 79850
    },
    {
      "epoch": 8.599720159293941,
      "grad_norm": 0.8498730659484863,
      "learning_rate": 8.432065510182093e-05,
      "loss": 3.2075,
      "step": 79900
    },
    {
      "epoch": 8.605101711333548,
      "grad_norm": 0.8368560671806335,
      "learning_rate": 8.39974140717595e-05,
      "loss": 3.2049,
      "step": 79950
    },
    {
      "epoch": 8.610483263373157,
      "grad_norm": 0.8408549427986145,
      "learning_rate": 8.367417304169809e-05,
      "loss": 3.2063,
      "step": 80000
    },
    {
      "epoch": 8.610483263373157,
      "eval_accuracy": 0.3907614187937552,
      "eval_loss": 3.3295724391937256,
      "eval_runtime": 185.4133,
      "eval_samples_per_second": 97.14,
      "eval_steps_per_second": 6.073,
      "step": 80000
    },
    {
      "epoch": 8.615864815412765,
      "grad_norm": 0.8205835223197937,
      "learning_rate": 8.335093201163667e-05,
      "loss": 3.1972,
      "step": 80050
    },
    {
      "epoch": 8.621246367452374,
      "grad_norm": 0.8415143489837646,
      "learning_rate": 8.302769098157526e-05,
      "loss": 3.2082,
      "step": 80100
    },
    {
      "epoch": 8.626627919491982,
      "grad_norm": 0.8902560472488403,
      "learning_rate": 8.270444995151383e-05,
      "loss": 3.2102,
      "step": 80150
    },
    {
      "epoch": 8.632009471531589,
      "grad_norm": 0.8608832359313965,
      "learning_rate": 8.238120892145243e-05,
      "loss": 3.202,
      "step": 80200
    },
    {
      "epoch": 8.637391023571197,
      "grad_norm": 0.8526662588119507,
      "learning_rate": 8.205796789139101e-05,
      "loss": 3.1839,
      "step": 80250
    },
    {
      "epoch": 8.642772575610806,
      "grad_norm": 0.8512690663337708,
      "learning_rate": 8.173472686132959e-05,
      "loss": 3.2087,
      "step": 80300
    },
    {
      "epoch": 8.648154127650415,
      "grad_norm": 0.7982252240180969,
      "learning_rate": 8.141148583126817e-05,
      "loss": 3.1831,
      "step": 80350
    },
    {
      "epoch": 8.653535679690023,
      "grad_norm": 0.841755211353302,
      "learning_rate": 8.108824480120676e-05,
      "loss": 3.1955,
      "step": 80400
    },
    {
      "epoch": 8.658917231729632,
      "grad_norm": 0.8930681347846985,
      "learning_rate": 8.076500377114533e-05,
      "loss": 3.2072,
      "step": 80450
    },
    {
      "epoch": 8.664298783769238,
      "grad_norm": 0.8469364643096924,
      "learning_rate": 8.044176274108393e-05,
      "loss": 3.2032,
      "step": 80500
    },
    {
      "epoch": 8.669680335808847,
      "grad_norm": 0.7756960391998291,
      "learning_rate": 8.011852171102252e-05,
      "loss": 3.1958,
      "step": 80550
    },
    {
      "epoch": 8.675061887848456,
      "grad_norm": 0.8554670214653015,
      "learning_rate": 7.97952806809611e-05,
      "loss": 3.2107,
      "step": 80600
    },
    {
      "epoch": 8.680443439888064,
      "grad_norm": 0.8408672213554382,
      "learning_rate": 7.947203965089967e-05,
      "loss": 3.192,
      "step": 80650
    },
    {
      "epoch": 8.685824991927673,
      "grad_norm": 0.8197949528694153,
      "learning_rate": 7.914879862083827e-05,
      "loss": 3.1885,
      "step": 80700
    },
    {
      "epoch": 8.69120654396728,
      "grad_norm": 0.8987782001495361,
      "learning_rate": 7.882555759077686e-05,
      "loss": 3.1843,
      "step": 80750
    },
    {
      "epoch": 8.696588096006888,
      "grad_norm": 0.8135241866111755,
      "learning_rate": 7.850231656071543e-05,
      "loss": 3.1832,
      "step": 80800
    },
    {
      "epoch": 8.701969648046497,
      "grad_norm": 0.8692455887794495,
      "learning_rate": 7.817907553065402e-05,
      "loss": 3.2195,
      "step": 80850
    },
    {
      "epoch": 8.707351200086105,
      "grad_norm": 0.8524072766304016,
      "learning_rate": 7.78558345005926e-05,
      "loss": 3.2032,
      "step": 80900
    },
    {
      "epoch": 8.712732752125714,
      "grad_norm": 0.9154481291770935,
      "learning_rate": 7.753259347053118e-05,
      "loss": 3.1999,
      "step": 80950
    },
    {
      "epoch": 8.718114304165322,
      "grad_norm": 0.8274575471878052,
      "learning_rate": 7.720935244046977e-05,
      "loss": 3.2116,
      "step": 81000
    },
    {
      "epoch": 8.718114304165322,
      "eval_accuracy": 0.39098220146179424,
      "eval_loss": 3.326021432876587,
      "eval_runtime": 185.0212,
      "eval_samples_per_second": 97.346,
      "eval_steps_per_second": 6.086,
      "step": 81000
    },
    {
      "epoch": 8.723495856204929,
      "grad_norm": 0.8735546469688416,
      "learning_rate": 7.688611141040836e-05,
      "loss": 3.2068,
      "step": 81050
    },
    {
      "epoch": 8.728877408244538,
      "grad_norm": 0.9155970215797424,
      "learning_rate": 7.656287038034694e-05,
      "loss": 3.1886,
      "step": 81100
    },
    {
      "epoch": 8.734258960284146,
      "grad_norm": 0.8984163403511047,
      "learning_rate": 7.623962935028552e-05,
      "loss": 3.1908,
      "step": 81150
    },
    {
      "epoch": 8.739640512323755,
      "grad_norm": 0.8430924415588379,
      "learning_rate": 7.59163883202241e-05,
      "loss": 3.1843,
      "step": 81200
    },
    {
      "epoch": 8.745022064363363,
      "grad_norm": 0.8486507534980774,
      "learning_rate": 7.55931472901627e-05,
      "loss": 3.2055,
      "step": 81250
    },
    {
      "epoch": 8.75040361640297,
      "grad_norm": 0.8695037364959717,
      "learning_rate": 7.526990626010127e-05,
      "loss": 3.199,
      "step": 81300
    },
    {
      "epoch": 8.755785168442578,
      "grad_norm": 0.8437666893005371,
      "learning_rate": 7.494666523003986e-05,
      "loss": 3.205,
      "step": 81350
    },
    {
      "epoch": 8.761166720482187,
      "grad_norm": 0.8271080255508423,
      "learning_rate": 7.462342419997844e-05,
      "loss": 3.1903,
      "step": 81400
    },
    {
      "epoch": 8.766548272521796,
      "grad_norm": 0.8238117694854736,
      "learning_rate": 7.430018316991704e-05,
      "loss": 3.1899,
      "step": 81450
    },
    {
      "epoch": 8.771929824561404,
      "grad_norm": 0.8370108604431152,
      "learning_rate": 7.397694213985562e-05,
      "loss": 3.1947,
      "step": 81500
    },
    {
      "epoch": 8.777311376601011,
      "grad_norm": 0.845049262046814,
      "learning_rate": 7.36537011097942e-05,
      "loss": 3.1849,
      "step": 81550
    },
    {
      "epoch": 8.78269292864062,
      "grad_norm": 0.8397005796432495,
      "learning_rate": 7.333046007973278e-05,
      "loss": 3.2066,
      "step": 81600
    },
    {
      "epoch": 8.788074480680228,
      "grad_norm": 0.7932242155075073,
      "learning_rate": 7.300721904967136e-05,
      "loss": 3.192,
      "step": 81650
    },
    {
      "epoch": 8.793456032719837,
      "grad_norm": 0.800791323184967,
      "learning_rate": 7.268397801960996e-05,
      "loss": 3.1952,
      "step": 81700
    },
    {
      "epoch": 8.798837584759445,
      "grad_norm": 0.8420032858848572,
      "learning_rate": 7.236073698954854e-05,
      "loss": 3.1857,
      "step": 81750
    },
    {
      "epoch": 8.804219136799054,
      "grad_norm": 0.8216664791107178,
      "learning_rate": 7.203749595948712e-05,
      "loss": 3.1978,
      "step": 81800
    },
    {
      "epoch": 8.80960068883866,
      "grad_norm": 0.8337118029594421,
      "learning_rate": 7.17142549294257e-05,
      "loss": 3.2037,
      "step": 81850
    },
    {
      "epoch": 8.814982240878269,
      "grad_norm": 0.8795151114463806,
      "learning_rate": 7.139101389936428e-05,
      "loss": 3.1844,
      "step": 81900
    },
    {
      "epoch": 8.820363792917878,
      "grad_norm": 0.8091270923614502,
      "learning_rate": 7.10742376899041e-05,
      "loss": 3.2282,
      "step": 81950
    },
    {
      "epoch": 8.825745344957486,
      "grad_norm": 0.8393868207931519,
      "learning_rate": 7.075099665984268e-05,
      "loss": 3.1955,
      "step": 82000
    },
    {
      "epoch": 8.825745344957486,
      "eval_accuracy": 0.3912356886490654,
      "eval_loss": 3.3253026008605957,
      "eval_runtime": 185.2394,
      "eval_samples_per_second": 97.231,
      "eval_steps_per_second": 6.079,
      "step": 82000
    },
    {
      "epoch": 8.831126896997095,
      "grad_norm": 0.8434122800827026,
      "learning_rate": 7.042775562978126e-05,
      "loss": 3.2119,
      "step": 82050
    },
    {
      "epoch": 8.836508449036701,
      "grad_norm": 0.8267486691474915,
      "learning_rate": 7.010451459971986e-05,
      "loss": 3.201,
      "step": 82100
    },
    {
      "epoch": 8.84189000107631,
      "grad_norm": 0.8528499007225037,
      "learning_rate": 6.978127356965844e-05,
      "loss": 3.1917,
      "step": 82150
    },
    {
      "epoch": 8.847271553115919,
      "grad_norm": 0.8093975782394409,
      "learning_rate": 6.945803253959702e-05,
      "loss": 3.185,
      "step": 82200
    },
    {
      "epoch": 8.852653105155527,
      "grad_norm": 0.8355822563171387,
      "learning_rate": 6.91347915095356e-05,
      "loss": 3.2022,
      "step": 82250
    },
    {
      "epoch": 8.858034657195136,
      "grad_norm": 0.8534575700759888,
      "learning_rate": 6.881155047947418e-05,
      "loss": 3.1889,
      "step": 82300
    },
    {
      "epoch": 8.863416209234742,
      "grad_norm": 0.8227938413619995,
      "learning_rate": 6.848830944941278e-05,
      "loss": 3.203,
      "step": 82350
    },
    {
      "epoch": 8.868797761274351,
      "grad_norm": 0.8066578507423401,
      "learning_rate": 6.816506841935136e-05,
      "loss": 3.2034,
      "step": 82400
    },
    {
      "epoch": 8.87417931331396,
      "grad_norm": 0.8343899250030518,
      "learning_rate": 6.784182738928994e-05,
      "loss": 3.218,
      "step": 82450
    },
    {
      "epoch": 8.879560865353568,
      "grad_norm": 0.8471869230270386,
      "learning_rate": 6.751858635922853e-05,
      "loss": 3.2025,
      "step": 82500
    },
    {
      "epoch": 8.884942417393177,
      "grad_norm": 0.866264820098877,
      "learning_rate": 6.719534532916711e-05,
      "loss": 3.2164,
      "step": 82550
    },
    {
      "epoch": 8.890323969432785,
      "grad_norm": 0.8603134155273438,
      "learning_rate": 6.68721042991057e-05,
      "loss": 3.2102,
      "step": 82600
    },
    {
      "epoch": 8.895705521472392,
      "grad_norm": 0.8576210141181946,
      "learning_rate": 6.654886326904428e-05,
      "loss": 3.2167,
      "step": 82650
    },
    {
      "epoch": 8.901087073512,
      "grad_norm": 0.8593220114707947,
      "learning_rate": 6.622562223898286e-05,
      "loss": 3.197,
      "step": 82700
    },
    {
      "epoch": 8.906468625551609,
      "grad_norm": 0.8577572703361511,
      "learning_rate": 6.590238120892145e-05,
      "loss": 3.1969,
      "step": 82750
    },
    {
      "epoch": 8.911850177591218,
      "grad_norm": 0.884865939617157,
      "learning_rate": 6.557914017886003e-05,
      "loss": 3.2166,
      "step": 82800
    },
    {
      "epoch": 8.917231729630826,
      "grad_norm": 0.8057446479797363,
      "learning_rate": 6.525589914879861e-05,
      "loss": 3.1971,
      "step": 82850
    },
    {
      "epoch": 8.922613281670433,
      "grad_norm": 0.9048945307731628,
      "learning_rate": 6.49326581187372e-05,
      "loss": 3.1883,
      "step": 82900
    },
    {
      "epoch": 8.927994833710041,
      "grad_norm": 0.8496543169021606,
      "learning_rate": 6.460941708867578e-05,
      "loss": 3.2132,
      "step": 82950
    },
    {
      "epoch": 8.93337638574965,
      "grad_norm": 0.8496195673942566,
      "learning_rate": 6.428617605861437e-05,
      "loss": 3.1902,
      "step": 83000
    },
    {
      "epoch": 8.93337638574965,
      "eval_accuracy": 0.3917239747269037,
      "eval_loss": 3.319549083709717,
      "eval_runtime": 185.2789,
      "eval_samples_per_second": 97.21,
      "eval_steps_per_second": 6.077,
      "step": 83000
    },
    {
      "epoch": 8.938757937789259,
      "grad_norm": 0.8542683720588684,
      "learning_rate": 6.396293502855295e-05,
      "loss": 3.199,
      "step": 83050
    },
    {
      "epoch": 8.944139489828867,
      "grad_norm": 0.8481775522232056,
      "learning_rate": 6.363969399849153e-05,
      "loss": 3.1911,
      "step": 83100
    },
    {
      "epoch": 8.949521041868476,
      "grad_norm": 0.9178012013435364,
      "learning_rate": 6.331645296843011e-05,
      "loss": 3.2024,
      "step": 83150
    },
    {
      "epoch": 8.954902593908082,
      "grad_norm": 0.8295817375183105,
      "learning_rate": 6.29932119383687e-05,
      "loss": 3.1905,
      "step": 83200
    },
    {
      "epoch": 8.960284145947691,
      "grad_norm": 0.8027358055114746,
      "learning_rate": 6.266997090830729e-05,
      "loss": 3.2092,
      "step": 83250
    },
    {
      "epoch": 8.9656656979873,
      "grad_norm": 0.8591536283493042,
      "learning_rate": 6.234672987824587e-05,
      "loss": 3.2045,
      "step": 83300
    },
    {
      "epoch": 8.971047250026908,
      "grad_norm": 0.8455834984779358,
      "learning_rate": 6.202348884818447e-05,
      "loss": 3.2058,
      "step": 83350
    },
    {
      "epoch": 8.976428802066517,
      "grad_norm": 0.8751072287559509,
      "learning_rate": 6.170024781812305e-05,
      "loss": 3.1905,
      "step": 83400
    },
    {
      "epoch": 8.981810354106123,
      "grad_norm": 0.8248118162155151,
      "learning_rate": 6.137700678806163e-05,
      "loss": 3.2143,
      "step": 83450
    },
    {
      "epoch": 8.987191906145732,
      "grad_norm": 0.8677883148193359,
      "learning_rate": 6.105376575800021e-05,
      "loss": 3.1857,
      "step": 83500
    },
    {
      "epoch": 8.99257345818534,
      "grad_norm": 0.826434314250946,
      "learning_rate": 6.073052472793879e-05,
      "loss": 3.2074,
      "step": 83550
    },
    {
      "epoch": 8.997955010224949,
      "grad_norm": 0.8779363632202148,
      "learning_rate": 6.0407283697877384e-05,
      "loss": 3.2023,
      "step": 83600
    },
    {
      "epoch": 9.003336562264558,
      "grad_norm": 0.8303587436676025,
      "learning_rate": 6.0084042667815966e-05,
      "loss": 3.1637,
      "step": 83650
    },
    {
      "epoch": 9.008718114304166,
      "grad_norm": 0.8080105185508728,
      "learning_rate": 5.976080163775455e-05,
      "loss": 3.1326,
      "step": 83700
    },
    {
      "epoch": 9.014099666343773,
      "grad_norm": 0.859494149684906,
      "learning_rate": 5.9437560607693135e-05,
      "loss": 3.1379,
      "step": 83750
    },
    {
      "epoch": 9.019481218383381,
      "grad_norm": 0.8405566215515137,
      "learning_rate": 5.9114319577631716e-05,
      "loss": 3.1357,
      "step": 83800
    },
    {
      "epoch": 9.02486277042299,
      "grad_norm": 0.8935002684593201,
      "learning_rate": 5.8791078547570304e-05,
      "loss": 3.1303,
      "step": 83850
    },
    {
      "epoch": 9.030244322462599,
      "grad_norm": 0.8550703525543213,
      "learning_rate": 5.8467837517508885e-05,
      "loss": 3.1363,
      "step": 83900
    },
    {
      "epoch": 9.035625874502207,
      "grad_norm": 0.8307282328605652,
      "learning_rate": 5.81510613080487e-05,
      "loss": 3.1405,
      "step": 83950
    },
    {
      "epoch": 9.041007426541814,
      "grad_norm": 0.8739628195762634,
      "learning_rate": 5.782782027798728e-05,
      "loss": 3.1168,
      "step": 84000
    },
    {
      "epoch": 9.041007426541814,
      "eval_accuracy": 0.3916440062014722,
      "eval_loss": 3.323110818862915,
      "eval_runtime": 185.4558,
      "eval_samples_per_second": 97.117,
      "eval_steps_per_second": 6.072,
      "step": 84000
    },
    {
      "epoch": 9.046388978581422,
      "grad_norm": 0.8511670827865601,
      "learning_rate": 5.750457924792587e-05,
      "loss": 3.1457,
      "step": 84050
    },
    {
      "epoch": 9.051770530621031,
      "grad_norm": 0.8276234865188599,
      "learning_rate": 5.718133821786445e-05,
      "loss": 3.1498,
      "step": 84100
    },
    {
      "epoch": 9.05715208266064,
      "grad_norm": 0.8410462737083435,
      "learning_rate": 5.6858097187803036e-05,
      "loss": 3.1457,
      "step": 84150
    },
    {
      "epoch": 9.062533634700248,
      "grad_norm": 0.8362277150154114,
      "learning_rate": 5.653485615774162e-05,
      "loss": 3.1568,
      "step": 84200
    },
    {
      "epoch": 9.067915186739857,
      "grad_norm": 0.8473198413848877,
      "learning_rate": 5.62116151276802e-05,
      "loss": 3.1339,
      "step": 84250
    },
    {
      "epoch": 9.073296738779463,
      "grad_norm": 0.8880166411399841,
      "learning_rate": 5.5888374097618786e-05,
      "loss": 3.1448,
      "step": 84300
    },
    {
      "epoch": 9.078678290819072,
      "grad_norm": 0.8744847774505615,
      "learning_rate": 5.556513306755737e-05,
      "loss": 3.141,
      "step": 84350
    },
    {
      "epoch": 9.08405984285868,
      "grad_norm": 0.822683572769165,
      "learning_rate": 5.5241892037495955e-05,
      "loss": 3.1471,
      "step": 84400
    },
    {
      "epoch": 9.089441394898289,
      "grad_norm": 0.9234932065010071,
      "learning_rate": 5.4918651007434536e-05,
      "loss": 3.1444,
      "step": 84450
    },
    {
      "epoch": 9.094822946937898,
      "grad_norm": 0.8440639972686768,
      "learning_rate": 5.459540997737312e-05,
      "loss": 3.1578,
      "step": 84500
    },
    {
      "epoch": 9.100204498977504,
      "grad_norm": 0.8613219857215881,
      "learning_rate": 5.427216894731171e-05,
      "loss": 3.1325,
      "step": 84550
    },
    {
      "epoch": 9.105586051017113,
      "grad_norm": 0.8684327006340027,
      "learning_rate": 5.394892791725029e-05,
      "loss": 3.1439,
      "step": 84600
    },
    {
      "epoch": 9.110967603056721,
      "grad_norm": 0.8598358035087585,
      "learning_rate": 5.362568688718888e-05,
      "loss": 3.1329,
      "step": 84650
    },
    {
      "epoch": 9.11634915509633,
      "grad_norm": 0.8764355182647705,
      "learning_rate": 5.330244585712746e-05,
      "loss": 3.1284,
      "step": 84700
    },
    {
      "epoch": 9.121730707135939,
      "grad_norm": 0.8224071860313416,
      "learning_rate": 5.297920482706604e-05,
      "loss": 3.1437,
      "step": 84750
    },
    {
      "epoch": 9.127112259175545,
      "grad_norm": 0.8611595630645752,
      "learning_rate": 5.265596379700463e-05,
      "loss": 3.1353,
      "step": 84800
    },
    {
      "epoch": 9.132493811215154,
      "grad_norm": 0.82994145154953,
      "learning_rate": 5.233272276694321e-05,
      "loss": 3.1475,
      "step": 84850
    },
    {
      "epoch": 9.137875363254762,
      "grad_norm": 0.83476322889328,
      "learning_rate": 5.20094817368818e-05,
      "loss": 3.1451,
      "step": 84900
    },
    {
      "epoch": 9.143256915294371,
      "grad_norm": 0.8830773234367371,
      "learning_rate": 5.168624070682038e-05,
      "loss": 3.1484,
      "step": 84950
    },
    {
      "epoch": 9.14863846733398,
      "grad_norm": 0.865684986114502,
      "learning_rate": 5.136299967675896e-05,
      "loss": 3.1455,
      "step": 85000
    },
    {
      "epoch": 9.14863846733398,
      "eval_accuracy": 0.39201940192887125,
      "eval_loss": 3.320784568786621,
      "eval_runtime": 184.9784,
      "eval_samples_per_second": 97.368,
      "eval_steps_per_second": 6.087,
      "step": 85000
    },
    {
      "epoch": 9.154020019373588,
      "grad_norm": 0.8420153856277466,
      "learning_rate": 5.103975864669755e-05,
      "loss": 3.1572,
      "step": 85050
    },
    {
      "epoch": 9.159401571413195,
      "grad_norm": 0.8855005502700806,
      "learning_rate": 5.071651761663613e-05,
      "loss": 3.1463,
      "step": 85100
    },
    {
      "epoch": 9.164783123452803,
      "grad_norm": 0.8882396817207336,
      "learning_rate": 5.039327658657472e-05,
      "loss": 3.1547,
      "step": 85150
    },
    {
      "epoch": 9.170164675492412,
      "grad_norm": 0.8707055449485779,
      "learning_rate": 5.00700355565133e-05,
      "loss": 3.1286,
      "step": 85200
    },
    {
      "epoch": 9.17554622753202,
      "grad_norm": 0.8751034140586853,
      "learning_rate": 4.974679452645188e-05,
      "loss": 3.1454,
      "step": 85250
    },
    {
      "epoch": 9.180927779571629,
      "grad_norm": 0.8328869342803955,
      "learning_rate": 4.942355349639047e-05,
      "loss": 3.1395,
      "step": 85300
    },
    {
      "epoch": 9.186309331611236,
      "grad_norm": 0.8701635003089905,
      "learning_rate": 4.910031246632905e-05,
      "loss": 3.1374,
      "step": 85350
    },
    {
      "epoch": 9.191690883650844,
      "grad_norm": 0.914456844329834,
      "learning_rate": 4.8777071436267645e-05,
      "loss": 3.119,
      "step": 85400
    },
    {
      "epoch": 9.197072435690453,
      "grad_norm": 0.8553349375724792,
      "learning_rate": 4.8453830406206226e-05,
      "loss": 3.1368,
      "step": 85450
    },
    {
      "epoch": 9.202453987730062,
      "grad_norm": 0.8387361764907837,
      "learning_rate": 4.813058937614481e-05,
      "loss": 3.1521,
      "step": 85500
    },
    {
      "epoch": 9.20783553976967,
      "grad_norm": 0.8626623749732971,
      "learning_rate": 4.7807348346083395e-05,
      "loss": 3.1416,
      "step": 85550
    },
    {
      "epoch": 9.213217091809279,
      "grad_norm": 0.8619423508644104,
      "learning_rate": 4.7484107316021976e-05,
      "loss": 3.1318,
      "step": 85600
    },
    {
      "epoch": 9.218598643848885,
      "grad_norm": 0.8979377150535583,
      "learning_rate": 4.7160866285960564e-05,
      "loss": 3.1459,
      "step": 85650
    },
    {
      "epoch": 9.223980195888494,
      "grad_norm": 0.8428804874420166,
      "learning_rate": 4.6837625255899145e-05,
      "loss": 3.1573,
      "step": 85700
    },
    {
      "epoch": 9.229361747928102,
      "grad_norm": 0.9066821932792664,
      "learning_rate": 4.6514384225837726e-05,
      "loss": 3.1385,
      "step": 85750
    },
    {
      "epoch": 9.234743299967711,
      "grad_norm": 0.8633950352668762,
      "learning_rate": 4.6191143195776314e-05,
      "loss": 3.1528,
      "step": 85800
    },
    {
      "epoch": 9.24012485200732,
      "grad_norm": 0.8855082392692566,
      "learning_rate": 4.5867902165714895e-05,
      "loss": 3.1415,
      "step": 85850
    },
    {
      "epoch": 9.245506404046926,
      "grad_norm": 0.8459654450416565,
      "learning_rate": 4.554466113565348e-05,
      "loss": 3.1465,
      "step": 85900
    },
    {
      "epoch": 9.250887956086535,
      "grad_norm": 0.8651683926582336,
      "learning_rate": 4.5221420105592064e-05,
      "loss": 3.1422,
      "step": 85950
    },
    {
      "epoch": 9.256269508126143,
      "grad_norm": 0.9303048253059387,
      "learning_rate": 4.4898179075530645e-05,
      "loss": 3.1348,
      "step": 86000
    },
    {
      "epoch": 9.256269508126143,
      "eval_accuracy": 0.3921100184373085,
      "eval_loss": 3.3196284770965576,
      "eval_runtime": 185.1194,
      "eval_samples_per_second": 97.294,
      "eval_steps_per_second": 6.083,
      "step": 86000
    },
    {
      "epoch": 9.261651060165752,
      "grad_norm": 0.8690763115882874,
      "learning_rate": 4.458140286607046e-05,
      "loss": 3.1529,
      "step": 86050
    },
    {
      "epoch": 9.26703261220536,
      "grad_norm": 0.8387214541435242,
      "learning_rate": 4.4258161836009046e-05,
      "loss": 3.1518,
      "step": 86100
    },
    {
      "epoch": 9.272414164244967,
      "grad_norm": 0.9187512397766113,
      "learning_rate": 4.393492080594763e-05,
      "loss": 3.157,
      "step": 86150
    },
    {
      "epoch": 9.277795716284576,
      "grad_norm": 0.8413276076316833,
      "learning_rate": 4.3611679775886215e-05,
      "loss": 3.1444,
      "step": 86200
    },
    {
      "epoch": 9.283177268324184,
      "grad_norm": 0.8400559425354004,
      "learning_rate": 4.3288438745824797e-05,
      "loss": 3.1445,
      "step": 86250
    },
    {
      "epoch": 9.288558820363793,
      "grad_norm": 0.8541315793991089,
      "learning_rate": 4.296519771576338e-05,
      "loss": 3.1512,
      "step": 86300
    },
    {
      "epoch": 9.293940372403402,
      "grad_norm": 0.8747986555099487,
      "learning_rate": 4.264195668570197e-05,
      "loss": 3.1626,
      "step": 86350
    },
    {
      "epoch": 9.29932192444301,
      "grad_norm": 0.8762104511260986,
      "learning_rate": 4.231871565564055e-05,
      "loss": 3.1537,
      "step": 86400
    },
    {
      "epoch": 9.304703476482617,
      "grad_norm": 0.8534042835235596,
      "learning_rate": 4.199547462557914e-05,
      "loss": 3.1451,
      "step": 86450
    },
    {
      "epoch": 9.310085028522225,
      "grad_norm": 0.884857177734375,
      "learning_rate": 4.167223359551772e-05,
      "loss": 3.1561,
      "step": 86500
    },
    {
      "epoch": 9.315466580561834,
      "grad_norm": 0.8563604950904846,
      "learning_rate": 4.1348992565456303e-05,
      "loss": 3.1379,
      "step": 86550
    },
    {
      "epoch": 9.320848132601443,
      "grad_norm": 0.8903757333755493,
      "learning_rate": 4.102575153539489e-05,
      "loss": 3.1312,
      "step": 86600
    },
    {
      "epoch": 9.326229684641051,
      "grad_norm": 0.8586431741714478,
      "learning_rate": 4.070251050533347e-05,
      "loss": 3.1623,
      "step": 86650
    },
    {
      "epoch": 9.331611236680658,
      "grad_norm": 0.8478698134422302,
      "learning_rate": 4.037926947527206e-05,
      "loss": 3.1268,
      "step": 86700
    },
    {
      "epoch": 9.336992788720266,
      "grad_norm": 0.8289090394973755,
      "learning_rate": 4.005602844521064e-05,
      "loss": 3.1583,
      "step": 86750
    },
    {
      "epoch": 9.342374340759875,
      "grad_norm": 0.8865683674812317,
      "learning_rate": 3.973278741514922e-05,
      "loss": 3.1462,
      "step": 86800
    },
    {
      "epoch": 9.347755892799483,
      "grad_norm": 0.8730096220970154,
      "learning_rate": 3.940954638508781e-05,
      "loss": 3.1595,
      "step": 86850
    },
    {
      "epoch": 9.353137444839092,
      "grad_norm": 0.8786698579788208,
      "learning_rate": 3.908630535502639e-05,
      "loss": 3.1427,
      "step": 86900
    },
    {
      "epoch": 9.3585189968787,
      "grad_norm": 0.8628304600715637,
      "learning_rate": 3.876306432496498e-05,
      "loss": 3.1439,
      "step": 86950
    },
    {
      "epoch": 9.363900548918307,
      "grad_norm": 0.8740919828414917,
      "learning_rate": 3.843982329490356e-05,
      "loss": 3.1499,
      "step": 87000
    },
    {
      "epoch": 9.363900548918307,
      "eval_accuracy": 0.39247237581816985,
      "eval_loss": 3.3164050579071045,
      "eval_runtime": 185.1969,
      "eval_samples_per_second": 97.253,
      "eval_steps_per_second": 6.08,
      "step": 87000
    },
    {
      "epoch": 9.369282100957916,
      "grad_norm": 0.8209638595581055,
      "learning_rate": 3.811658226484214e-05,
      "loss": 3.1194,
      "step": 87050
    },
    {
      "epoch": 9.374663652997524,
      "grad_norm": 0.8611879348754883,
      "learning_rate": 3.7793341234780736e-05,
      "loss": 3.1531,
      "step": 87100
    },
    {
      "epoch": 9.380045205037133,
      "grad_norm": 0.9188125133514404,
      "learning_rate": 3.747010020471931e-05,
      "loss": 3.1505,
      "step": 87150
    },
    {
      "epoch": 9.385426757076742,
      "grad_norm": 0.8500015735626221,
      "learning_rate": 3.71468591746579e-05,
      "loss": 3.1521,
      "step": 87200
    },
    {
      "epoch": 9.390808309116348,
      "grad_norm": 0.909965455532074,
      "learning_rate": 3.6823618144596486e-05,
      "loss": 3.1485,
      "step": 87250
    },
    {
      "epoch": 9.396189861155957,
      "grad_norm": 0.8830810785293579,
      "learning_rate": 3.650037711453507e-05,
      "loss": 3.1587,
      "step": 87300
    },
    {
      "epoch": 9.401571413195565,
      "grad_norm": 0.827194094657898,
      "learning_rate": 3.6177136084473655e-05,
      "loss": 3.1416,
      "step": 87350
    },
    {
      "epoch": 9.406952965235174,
      "grad_norm": 0.8261989951133728,
      "learning_rate": 3.5853895054412236e-05,
      "loss": 3.1527,
      "step": 87400
    },
    {
      "epoch": 9.412334517274783,
      "grad_norm": 0.894660472869873,
      "learning_rate": 3.553065402435082e-05,
      "loss": 3.148,
      "step": 87450
    },
    {
      "epoch": 9.417716069314391,
      "grad_norm": 0.9027609825134277,
      "learning_rate": 3.5207412994289405e-05,
      "loss": 3.1464,
      "step": 87500
    },
    {
      "epoch": 9.423097621353998,
      "grad_norm": 0.8734874725341797,
      "learning_rate": 3.488417196422799e-05,
      "loss": 3.1492,
      "step": 87550
    },
    {
      "epoch": 9.428479173393606,
      "grad_norm": 0.8822287321090698,
      "learning_rate": 3.4560930934166574e-05,
      "loss": 3.1465,
      "step": 87600
    },
    {
      "epoch": 9.433860725433215,
      "grad_norm": 0.9253379106521606,
      "learning_rate": 3.4237689904105156e-05,
      "loss": 3.1426,
      "step": 87650
    },
    {
      "epoch": 9.439242277472824,
      "grad_norm": 0.897438108921051,
      "learning_rate": 3.3914448874043743e-05,
      "loss": 3.1424,
      "step": 87700
    },
    {
      "epoch": 9.444623829512432,
      "grad_norm": 0.8358360528945923,
      "learning_rate": 3.3591207843982325e-05,
      "loss": 3.1345,
      "step": 87750
    },
    {
      "epoch": 9.450005381552039,
      "grad_norm": 0.8533014059066772,
      "learning_rate": 3.326796681392091e-05,
      "loss": 3.1366,
      "step": 87800
    },
    {
      "epoch": 9.455386933591647,
      "grad_norm": 0.8577408194541931,
      "learning_rate": 3.2944725783859494e-05,
      "loss": 3.1442,
      "step": 87850
    },
    {
      "epoch": 9.460768485631256,
      "grad_norm": 0.872809886932373,
      "learning_rate": 3.2621484753798075e-05,
      "loss": 3.1695,
      "step": 87900
    },
    {
      "epoch": 9.466150037670864,
      "grad_norm": 0.8611479997634888,
      "learning_rate": 3.229824372373666e-05,
      "loss": 3.169,
      "step": 87950
    },
    {
      "epoch": 9.471531589710473,
      "grad_norm": 0.8318737149238586,
      "learning_rate": 3.197500269367525e-05,
      "loss": 3.1286,
      "step": 88000
    },
    {
      "epoch": 9.471531589710473,
      "eval_accuracy": 0.3927649780450542,
      "eval_loss": 3.3146228790283203,
      "eval_runtime": 185.1936,
      "eval_samples_per_second": 97.255,
      "eval_steps_per_second": 6.08,
      "step": 88000
    },
    {
      "epoch": 9.476913141750082,
      "grad_norm": 0.8772470355033875,
      "learning_rate": 3.165822648421506e-05,
      "loss": 3.1413,
      "step": 88050
    },
    {
      "epoch": 9.482294693789688,
      "grad_norm": 0.8867277503013611,
      "learning_rate": 3.1334985454153645e-05,
      "loss": 3.1374,
      "step": 88100
    },
    {
      "epoch": 9.487676245829297,
      "grad_norm": 0.8771072626113892,
      "learning_rate": 3.101174442409223e-05,
      "loss": 3.1464,
      "step": 88150
    },
    {
      "epoch": 9.493057797868905,
      "grad_norm": 0.9186044335365295,
      "learning_rate": 3.0688503394030814e-05,
      "loss": 3.1616,
      "step": 88200
    },
    {
      "epoch": 9.498439349908514,
      "grad_norm": 0.8836739659309387,
      "learning_rate": 3.0365262363969395e-05,
      "loss": 3.1529,
      "step": 88250
    },
    {
      "epoch": 9.503820901948123,
      "grad_norm": 0.8628858327865601,
      "learning_rate": 3.0042021333907983e-05,
      "loss": 3.1528,
      "step": 88300
    },
    {
      "epoch": 9.50920245398773,
      "grad_norm": 0.8870872259140015,
      "learning_rate": 2.9718780303846567e-05,
      "loss": 3.1257,
      "step": 88350
    },
    {
      "epoch": 9.514584006027338,
      "grad_norm": 0.8661220073699951,
      "learning_rate": 2.9395539273785152e-05,
      "loss": 3.1566,
      "step": 88400
    },
    {
      "epoch": 9.519965558066946,
      "grad_norm": 0.861082911491394,
      "learning_rate": 2.9072298243723733e-05,
      "loss": 3.1516,
      "step": 88450
    },
    {
      "epoch": 9.525347110106555,
      "grad_norm": 0.8526957035064697,
      "learning_rate": 2.8749057213662317e-05,
      "loss": 3.1565,
      "step": 88500
    },
    {
      "epoch": 9.530728662146164,
      "grad_norm": 0.8312302231788635,
      "learning_rate": 2.8425816183600902e-05,
      "loss": 3.1381,
      "step": 88550
    },
    {
      "epoch": 9.536110214185772,
      "grad_norm": 0.8517558574676514,
      "learning_rate": 2.8102575153539486e-05,
      "loss": 3.1478,
      "step": 88600
    },
    {
      "epoch": 9.541491766225379,
      "grad_norm": 0.8223721981048584,
      "learning_rate": 2.777933412347807e-05,
      "loss": 3.1441,
      "step": 88650
    },
    {
      "epoch": 9.546873318264987,
      "grad_norm": 0.8433859944343567,
      "learning_rate": 2.7456093093416652e-05,
      "loss": 3.147,
      "step": 88700
    },
    {
      "epoch": 9.552254870304596,
      "grad_norm": 0.8294910788536072,
      "learning_rate": 2.713285206335524e-05,
      "loss": 3.1237,
      "step": 88750
    },
    {
      "epoch": 9.557636422344205,
      "grad_norm": 0.9038265347480774,
      "learning_rate": 2.6809611033293824e-05,
      "loss": 3.1468,
      "step": 88800
    },
    {
      "epoch": 9.563017974383813,
      "grad_norm": 0.8614657521247864,
      "learning_rate": 2.648637000323241e-05,
      "loss": 3.135,
      "step": 88850
    },
    {
      "epoch": 9.56839952642342,
      "grad_norm": 0.8738344311714172,
      "learning_rate": 2.6163128973170993e-05,
      "loss": 3.153,
      "step": 88900
    },
    {
      "epoch": 9.573781078463028,
      "grad_norm": 0.8263546824455261,
      "learning_rate": 2.5839887943109574e-05,
      "loss": 3.1531,
      "step": 88950
    },
    {
      "epoch": 9.579162630502637,
      "grad_norm": 0.8386525511741638,
      "learning_rate": 2.551664691304816e-05,
      "loss": 3.1561,
      "step": 89000
    },
    {
      "epoch": 9.579162630502637,
      "eval_accuracy": 0.3930701840069251,
      "eval_loss": 3.3113958835601807,
      "eval_runtime": 185.3911,
      "eval_samples_per_second": 97.151,
      "eval_steps_per_second": 6.074,
      "step": 89000
    },
    {
      "epoch": 9.584544182542245,
      "grad_norm": 0.8265148997306824,
      "learning_rate": 2.5193405882986743e-05,
      "loss": 3.142,
      "step": 89050
    },
    {
      "epoch": 9.589925734581854,
      "grad_norm": 0.8971044421195984,
      "learning_rate": 2.487016485292533e-05,
      "loss": 3.1412,
      "step": 89100
    },
    {
      "epoch": 9.59530728662146,
      "grad_norm": 0.8778616189956665,
      "learning_rate": 2.4546923822863916e-05,
      "loss": 3.1479,
      "step": 89150
    },
    {
      "epoch": 9.60068883866107,
      "grad_norm": 0.9554949402809143,
      "learning_rate": 2.4223682792802497e-05,
      "loss": 3.1433,
      "step": 89200
    },
    {
      "epoch": 9.606070390700678,
      "grad_norm": 0.886187732219696,
      "learning_rate": 2.390044176274108e-05,
      "loss": 3.1483,
      "step": 89250
    },
    {
      "epoch": 9.611451942740286,
      "grad_norm": 0.8735513091087341,
      "learning_rate": 2.3577200732679666e-05,
      "loss": 3.1512,
      "step": 89300
    },
    {
      "epoch": 9.616833494779895,
      "grad_norm": 0.8736987709999084,
      "learning_rate": 2.325395970261825e-05,
      "loss": 3.145,
      "step": 89350
    },
    {
      "epoch": 9.622215046819504,
      "grad_norm": 0.8718150854110718,
      "learning_rate": 2.2930718672556835e-05,
      "loss": 3.156,
      "step": 89400
    },
    {
      "epoch": 9.62759659885911,
      "grad_norm": 0.8639832735061646,
      "learning_rate": 2.2607477642495416e-05,
      "loss": 3.1494,
      "step": 89450
    },
    {
      "epoch": 9.632978150898719,
      "grad_norm": 0.8477591276168823,
      "learning_rate": 2.2284236612434e-05,
      "loss": 3.1627,
      "step": 89500
    },
    {
      "epoch": 9.638359702938327,
      "grad_norm": 0.8431605696678162,
      "learning_rate": 2.196099558237259e-05,
      "loss": 3.1586,
      "step": 89550
    },
    {
      "epoch": 9.643741254977936,
      "grad_norm": 0.8924245834350586,
      "learning_rate": 2.1637754552311173e-05,
      "loss": 3.1481,
      "step": 89600
    },
    {
      "epoch": 9.649122807017545,
      "grad_norm": 0.8491957187652588,
      "learning_rate": 2.1314513522249757e-05,
      "loss": 3.137,
      "step": 89650
    },
    {
      "epoch": 9.654504359057151,
      "grad_norm": 0.9181621670722961,
      "learning_rate": 2.099127249218834e-05,
      "loss": 3.1332,
      "step": 89700
    },
    {
      "epoch": 9.65988591109676,
      "grad_norm": 0.8941819071769714,
      "learning_rate": 2.0668031462126923e-05,
      "loss": 3.1455,
      "step": 89750
    },
    {
      "epoch": 9.665267463136368,
      "grad_norm": 0.8638315796852112,
      "learning_rate": 2.0344790432065507e-05,
      "loss": 3.1528,
      "step": 89800
    },
    {
      "epoch": 9.670649015175977,
      "grad_norm": 0.862423837184906,
      "learning_rate": 2.0021549402004092e-05,
      "loss": 3.1628,
      "step": 89850
    },
    {
      "epoch": 9.676030567215586,
      "grad_norm": 0.8840934634208679,
      "learning_rate": 1.969830837194268e-05,
      "loss": 3.1667,
      "step": 89900
    },
    {
      "epoch": 9.681412119255192,
      "grad_norm": 0.859711229801178,
      "learning_rate": 1.9375067341881258e-05,
      "loss": 3.1488,
      "step": 89950
    },
    {
      "epoch": 9.6867936712948,
      "grad_norm": 0.8524650931358337,
      "learning_rate": 1.9051826311819845e-05,
      "loss": 3.1599,
      "step": 90000
    },
    {
      "epoch": 9.6867936712948,
      "eval_accuracy": 0.3932824917497146,
      "eval_loss": 3.3091650009155273,
      "eval_runtime": 185.0723,
      "eval_samples_per_second": 97.319,
      "eval_steps_per_second": 6.084,
      "step": 90000
    },
    {
      "epoch": 9.69217522333441,
      "grad_norm": 0.9002635478973389,
      "learning_rate": 1.872858528175843e-05,
      "loss": 3.1515,
      "step": 90050
    },
    {
      "epoch": 9.697556775374018,
      "grad_norm": 0.8586625456809998,
      "learning_rate": 1.8411809072298243e-05,
      "loss": 3.1484,
      "step": 90100
    },
    {
      "epoch": 9.702938327413626,
      "grad_norm": 0.869149386882782,
      "learning_rate": 1.8088568042236828e-05,
      "loss": 3.1254,
      "step": 90150
    },
    {
      "epoch": 9.708319879453235,
      "grad_norm": 0.8365402817726135,
      "learning_rate": 1.776532701217541e-05,
      "loss": 3.1502,
      "step": 90200
    },
    {
      "epoch": 9.713701431492842,
      "grad_norm": 0.8755061626434326,
      "learning_rate": 1.7442085982113997e-05,
      "loss": 3.16,
      "step": 90250
    },
    {
      "epoch": 9.71908298353245,
      "grad_norm": 0.8689913749694824,
      "learning_rate": 1.7118844952052578e-05,
      "loss": 3.1509,
      "step": 90300
    },
    {
      "epoch": 9.724464535572059,
      "grad_norm": 0.821652889251709,
      "learning_rate": 1.6795603921991162e-05,
      "loss": 3.145,
      "step": 90350
    },
    {
      "epoch": 9.729846087611667,
      "grad_norm": 0.820548415184021,
      "learning_rate": 1.6472362891929747e-05,
      "loss": 3.1524,
      "step": 90400
    },
    {
      "epoch": 9.735227639651276,
      "grad_norm": 0.8812232613563538,
      "learning_rate": 1.614912186186833e-05,
      "loss": 3.1431,
      "step": 90450
    },
    {
      "epoch": 9.740609191690883,
      "grad_norm": 0.8736178278923035,
      "learning_rate": 1.5825880831806916e-05,
      "loss": 3.1475,
      "step": 90500
    },
    {
      "epoch": 9.745990743730491,
      "grad_norm": 0.8866865634918213,
      "learning_rate": 1.55026398017455e-05,
      "loss": 3.149,
      "step": 90550
    },
    {
      "epoch": 9.7513722957701,
      "grad_norm": 0.8401597738265991,
      "learning_rate": 1.5179398771684085e-05,
      "loss": 3.1419,
      "step": 90600
    },
    {
      "epoch": 9.756753847809708,
      "grad_norm": 0.8055000901222229,
      "learning_rate": 1.485615774162267e-05,
      "loss": 3.1493,
      "step": 90650
    },
    {
      "epoch": 9.762135399849317,
      "grad_norm": 0.861444354057312,
      "learning_rate": 1.4532916711561252e-05,
      "loss": 3.1296,
      "step": 90700
    },
    {
      "epoch": 9.767516951888926,
      "grad_norm": 0.8934671878814697,
      "learning_rate": 1.4209675681499837e-05,
      "loss": 3.1428,
      "step": 90750
    },
    {
      "epoch": 9.772898503928532,
      "grad_norm": 0.8632842898368835,
      "learning_rate": 1.3886434651438423e-05,
      "loss": 3.1515,
      "step": 90800
    },
    {
      "epoch": 9.77828005596814,
      "grad_norm": 0.8558943867683411,
      "learning_rate": 1.3563193621377006e-05,
      "loss": 3.1351,
      "step": 90850
    },
    {
      "epoch": 9.78366160800775,
      "grad_norm": 0.8853519558906555,
      "learning_rate": 1.323995259131559e-05,
      "loss": 3.1491,
      "step": 90900
    },
    {
      "epoch": 9.789043160047358,
      "grad_norm": 0.8454321622848511,
      "learning_rate": 1.2916711561254173e-05,
      "loss": 3.146,
      "step": 90950
    },
    {
      "epoch": 9.794424712086967,
      "grad_norm": 0.8885436654090881,
      "learning_rate": 1.2593470531192759e-05,
      "loss": 3.1425,
      "step": 91000
    },
    {
      "epoch": 9.794424712086967,
      "eval_accuracy": 0.39353271935035133,
      "eval_loss": 3.307443618774414,
      "eval_runtime": 185.0801,
      "eval_samples_per_second": 97.315,
      "eval_steps_per_second": 6.084,
      "step": 91000
    },
    {
      "epoch": 9.799806264126573,
      "grad_norm": 0.8688841462135315,
      "learning_rate": 1.2270229501131344e-05,
      "loss": 3.1401,
      "step": 91050
    },
    {
      "epoch": 9.805187816166182,
      "grad_norm": 0.8617741465568542,
      "learning_rate": 1.1946988471069926e-05,
      "loss": 3.1498,
      "step": 91100
    },
    {
      "epoch": 9.81056936820579,
      "grad_norm": 0.8525540828704834,
      "learning_rate": 1.162374744100851e-05,
      "loss": 3.1314,
      "step": 91150
    },
    {
      "epoch": 9.815950920245399,
      "grad_norm": 0.8676708936691284,
      "learning_rate": 1.1300506410947095e-05,
      "loss": 3.1511,
      "step": 91200
    },
    {
      "epoch": 9.821332472285007,
      "grad_norm": 0.8501707911491394,
      "learning_rate": 1.097726538088568e-05,
      "loss": 3.137,
      "step": 91250
    },
    {
      "epoch": 9.826714024324616,
      "grad_norm": 0.8348538279533386,
      "learning_rate": 1.0654024350824264e-05,
      "loss": 3.1289,
      "step": 91300
    },
    {
      "epoch": 9.832095576364223,
      "grad_norm": 0.861853837966919,
      "learning_rate": 1.0330783320762847e-05,
      "loss": 3.1511,
      "step": 91350
    },
    {
      "epoch": 9.837477128403831,
      "grad_norm": 0.892373263835907,
      "learning_rate": 1.0007542290701433e-05,
      "loss": 3.1341,
      "step": 91400
    },
    {
      "epoch": 9.84285868044344,
      "grad_norm": 0.8584460616111755,
      "learning_rate": 9.684301260640016e-06,
      "loss": 3.1489,
      "step": 91450
    },
    {
      "epoch": 9.848240232483048,
      "grad_norm": 0.8298298120498657,
      "learning_rate": 9.3610602305786e-06,
      "loss": 3.1456,
      "step": 91500
    },
    {
      "epoch": 9.853621784522657,
      "grad_norm": 0.8852813839912415,
      "learning_rate": 9.037819200517185e-06,
      "loss": 3.1499,
      "step": 91550
    },
    {
      "epoch": 9.859003336562264,
      "grad_norm": 0.8735877871513367,
      "learning_rate": 8.71457817045577e-06,
      "loss": 3.1435,
      "step": 91600
    },
    {
      "epoch": 9.864384888601872,
      "grad_norm": 0.9014072418212891,
      "learning_rate": 8.391337140394354e-06,
      "loss": 3.1516,
      "step": 91650
    },
    {
      "epoch": 9.869766440641481,
      "grad_norm": 0.8518697619438171,
      "learning_rate": 8.068096110332939e-06,
      "loss": 3.1537,
      "step": 91700
    },
    {
      "epoch": 9.87514799268109,
      "grad_norm": 0.8987732529640198,
      "learning_rate": 7.744855080271521e-06,
      "loss": 3.1417,
      "step": 91750
    },
    {
      "epoch": 9.880529544720698,
      "grad_norm": 0.9037292003631592,
      "learning_rate": 7.421614050210106e-06,
      "loss": 3.1522,
      "step": 91800
    },
    {
      "epoch": 9.885911096760307,
      "grad_norm": 0.8912764191627502,
      "learning_rate": 7.09837302014869e-06,
      "loss": 3.1575,
      "step": 91850
    },
    {
      "epoch": 9.891292648799913,
      "grad_norm": 0.8350512385368347,
      "learning_rate": 6.775131990087275e-06,
      "loss": 3.1404,
      "step": 91900
    },
    {
      "epoch": 9.896674200839522,
      "grad_norm": 0.8333302140235901,
      "learning_rate": 6.451890960025859e-06,
      "loss": 3.1496,
      "step": 91950
    },
    {
      "epoch": 9.90205575287913,
      "grad_norm": 0.8613854050636292,
      "learning_rate": 6.128649929964443e-06,
      "loss": 3.1375,
      "step": 92000
    },
    {
      "epoch": 9.90205575287913,
      "eval_accuracy": 0.3936786401786862,
      "eval_loss": 3.306281805038452,
      "eval_runtime": 185.4477,
      "eval_samples_per_second": 97.122,
      "eval_steps_per_second": 6.072,
      "step": 92000
    },
    {
      "epoch": 9.907437304918739,
      "grad_norm": 0.8688348531723022,
      "learning_rate": 5.805408899903027e-06,
      "loss": 3.1366,
      "step": 92050
    },
    {
      "epoch": 9.912818856958348,
      "grad_norm": 0.8763224482536316,
      "learning_rate": 5.482167869841611e-06,
      "loss": 3.1443,
      "step": 92100
    },
    {
      "epoch": 9.918200408997954,
      "grad_norm": 0.8729822635650635,
      "learning_rate": 5.1653916603814235e-06,
      "loss": 3.1436,
      "step": 92150
    },
    {
      "epoch": 9.923581961037563,
      "grad_norm": 0.8494361639022827,
      "learning_rate": 4.842150630320008e-06,
      "loss": 3.1407,
      "step": 92200
    },
    {
      "epoch": 9.928963513077171,
      "grad_norm": 0.8006011843681335,
      "learning_rate": 4.5189096002585925e-06,
      "loss": 3.1325,
      "step": 92250
    },
    {
      "epoch": 9.93434506511678,
      "grad_norm": 0.8903690576553345,
      "learning_rate": 4.195668570197177e-06,
      "loss": 3.1455,
      "step": 92300
    },
    {
      "epoch": 9.939726617156388,
      "grad_norm": 0.8697976469993591,
      "learning_rate": 3.872427540135761e-06,
      "loss": 3.1424,
      "step": 92350
    },
    {
      "epoch": 9.945108169195997,
      "grad_norm": 0.875449538230896,
      "learning_rate": 3.549186510074345e-06,
      "loss": 3.1471,
      "step": 92400
    },
    {
      "epoch": 9.950489721235604,
      "grad_norm": 0.8682908415794373,
      "learning_rate": 3.2259454800129297e-06,
      "loss": 3.136,
      "step": 92450
    },
    {
      "epoch": 9.955871273275212,
      "grad_norm": 0.86710524559021,
      "learning_rate": 2.9027044499515133e-06,
      "loss": 3.14,
      "step": 92500
    },
    {
      "epoch": 9.961252825314821,
      "grad_norm": 0.8853285312652588,
      "learning_rate": 2.5794634198900982e-06,
      "loss": 3.1493,
      "step": 92550
    },
    {
      "epoch": 9.96663437735443,
      "grad_norm": 0.8496193885803223,
      "learning_rate": 2.256222389828682e-06,
      "loss": 3.1199,
      "step": 92600
    },
    {
      "epoch": 9.972015929394038,
      "grad_norm": 0.8478330969810486,
      "learning_rate": 1.9329813597672664e-06,
      "loss": 3.14,
      "step": 92650
    },
    {
      "epoch": 9.977397481433645,
      "grad_norm": 0.8554485440254211,
      "learning_rate": 1.6097403297058506e-06,
      "loss": 3.1306,
      "step": 92700
    },
    {
      "epoch": 9.982779033473253,
      "grad_norm": 0.8549627065658569,
      "learning_rate": 1.292964120245663e-06,
      "loss": 3.1536,
      "step": 92750
    },
    {
      "epoch": 9.988160585512862,
      "grad_norm": 0.8293390274047852,
      "learning_rate": 9.697230901842474e-07,
      "loss": 3.1401,
      "step": 92800
    },
    {
      "epoch": 9.99354213755247,
      "grad_norm": 0.9368473887443542,
      "learning_rate": 6.464820601228315e-07,
      "loss": 3.1696,
      "step": 92850
    },
    {
      "epoch": 9.998923689592079,
      "grad_norm": 0.8504949808120728,
      "learning_rate": 3.2324103006141577e-07,
      "loss": 3.1536,
      "step": 92900
    }
  ],
  "logging_steps": 50,
  "max_steps": 92910,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 10000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.7681859821568e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}