{
  "best_metric": 3.308566093444824,
  "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_1208/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": 8.489361763000488,
      "learning_rate": 0.0003,
      "loss": 8.5006,
      "step": 50
    },
    {
      "epoch": 0.010763104079216447,
      "grad_norm": 1.9615029096603394,
      "learning_rate": 0.0006,
      "loss": 6.8831,
      "step": 100
    },
    {
      "epoch": 0.01614465611882467,
      "grad_norm": 1.9948406219482422,
      "learning_rate": 0.0005996767589699385,
      "loss": 6.4033,
      "step": 150
    },
    {
      "epoch": 0.021526208158432893,
      "grad_norm": 2.2949752807617188,
      "learning_rate": 0.0005993535179398771,
      "loss": 6.2058,
      "step": 200
    },
    {
      "epoch": 0.026907760198041114,
      "grad_norm": 1.0898317098617554,
      "learning_rate": 0.0005990302769098158,
      "loss": 6.0385,
      "step": 250
    },
    {
      "epoch": 0.03228931223764934,
      "grad_norm": 1.9062098264694214,
      "learning_rate": 0.0005987070358797543,
      "loss": 5.972,
      "step": 300
    },
    {
      "epoch": 0.03767086427725756,
      "grad_norm": 1.7062020301818848,
      "learning_rate": 0.0005983837948496929,
      "loss": 5.8878,
      "step": 350
    },
    {
      "epoch": 0.04305241631686579,
      "grad_norm": 1.0963656902313232,
      "learning_rate": 0.0005980605538196314,
      "loss": 5.7736,
      "step": 400
    },
    {
      "epoch": 0.048433968356474004,
      "grad_norm": 1.9562582969665527,
      "learning_rate": 0.0005977373127895701,
      "loss": 5.7226,
      "step": 450
    },
    {
      "epoch": 0.05381552039608223,
      "grad_norm": 1.2749587297439575,
      "learning_rate": 0.0005974140717595086,
      "loss": 5.6243,
      "step": 500
    },
    {
      "epoch": 0.05919707243569045,
      "grad_norm": 1.6608922481536865,
      "learning_rate": 0.0005970908307294472,
      "loss": 5.5516,
      "step": 550
    },
    {
      "epoch": 0.06457862447529868,
      "grad_norm": 1.5849041938781738,
      "learning_rate": 0.0005967675896993858,
      "loss": 5.4851,
      "step": 600
    },
    {
      "epoch": 0.0699601765149069,
      "grad_norm": 1.1615055799484253,
      "learning_rate": 0.0005964443486693243,
      "loss": 5.4325,
      "step": 650
    },
    {
      "epoch": 0.07534172855451512,
      "grad_norm": 1.5066120624542236,
      "learning_rate": 0.000596121107639263,
      "loss": 5.3674,
      "step": 700
    },
    {
      "epoch": 0.08072328059412334,
      "grad_norm": 1.1698895692825317,
      "learning_rate": 0.0005957978666092015,
      "loss": 5.2951,
      "step": 750
    },
    {
      "epoch": 0.08610483263373157,
      "grad_norm": 1.969469428062439,
      "learning_rate": 0.0005954746255791401,
      "loss": 5.256,
      "step": 800
    },
    {
      "epoch": 0.09148638467333979,
      "grad_norm": 1.214189052581787,
      "learning_rate": 0.0005951513845490787,
      "loss": 5.1884,
      "step": 850
    },
    {
      "epoch": 0.09686793671294801,
      "grad_norm": 1.052661657333374,
      "learning_rate": 0.0005948281435190174,
      "loss": 5.1527,
      "step": 900
    },
    {
      "epoch": 0.10224948875255624,
      "grad_norm": 1.2582467794418335,
      "learning_rate": 0.0005945049024889559,
      "loss": 5.1205,
      "step": 950
    },
    {
      "epoch": 0.10763104079216446,
      "grad_norm": 0.8772917985916138,
      "learning_rate": 0.0005941816614588944,
      "loss": 5.0934,
      "step": 1000
    },
    {
      "epoch": 0.10763104079216446,
      "eval_accuracy": 0.22830242575177748,
      "eval_loss": 5.011170387268066,
      "eval_runtime": 184.392,
      "eval_samples_per_second": 97.678,
      "eval_steps_per_second": 6.107,
      "step": 1000
    },
    {
      "epoch": 0.11301259283177269,
      "grad_norm": 1.2470630407333374,
      "learning_rate": 0.000593858420428833,
      "loss": 5.0303,
      "step": 1050
    },
    {
      "epoch": 0.1183941448713809,
      "grad_norm": 1.3724963665008545,
      "learning_rate": 0.0005935351793987716,
      "loss": 5.0012,
      "step": 1100
    },
    {
      "epoch": 0.12377569691098914,
      "grad_norm": 1.0659003257751465,
      "learning_rate": 0.0005932119383687103,
      "loss": 5.0045,
      "step": 1150
    },
    {
      "epoch": 0.12915724895059735,
      "grad_norm": 1.0786244869232178,
      "learning_rate": 0.0005928886973386488,
      "loss": 4.9516,
      "step": 1200
    },
    {
      "epoch": 0.13453880099020557,
      "grad_norm": 0.7411894798278809,
      "learning_rate": 0.0005925654563085874,
      "loss": 4.9101,
      "step": 1250
    },
    {
      "epoch": 0.1399203530298138,
      "grad_norm": 0.8150563836097717,
      "learning_rate": 0.000592242215278526,
      "loss": 4.8947,
      "step": 1300
    },
    {
      "epoch": 0.14530190506942203,
      "grad_norm": 0.8947958946228027,
      "learning_rate": 0.0005919189742484645,
      "loss": 4.8786,
      "step": 1350
    },
    {
      "epoch": 0.15068345710903025,
      "grad_norm": 1.3871676921844482,
      "learning_rate": 0.0005915957332184032,
      "loss": 4.8246,
      "step": 1400
    },
    {
      "epoch": 0.15606500914863847,
      "grad_norm": 0.9359856247901917,
      "learning_rate": 0.0005912724921883417,
      "loss": 4.8174,
      "step": 1450
    },
    {
      "epoch": 0.16144656118824668,
      "grad_norm": 1.100040078163147,
      "learning_rate": 0.0005909492511582803,
      "loss": 4.7943,
      "step": 1500
    },
    {
      "epoch": 0.1668281132278549,
      "grad_norm": 0.9592791795730591,
      "learning_rate": 0.0005906260101282189,
      "loss": 4.7722,
      "step": 1550
    },
    {
      "epoch": 0.17220966526746315,
      "grad_norm": 0.9244337677955627,
      "learning_rate": 0.0005903027690981575,
      "loss": 4.744,
      "step": 1600
    },
    {
      "epoch": 0.17759121730707136,
      "grad_norm": 1.0609703063964844,
      "learning_rate": 0.000589979528068096,
      "loss": 4.704,
      "step": 1650
    },
    {
      "epoch": 0.18297276934667958,
      "grad_norm": 1.3286961317062378,
      "learning_rate": 0.0005896562870380347,
      "loss": 4.6926,
      "step": 1700
    },
    {
      "epoch": 0.1883543213862878,
      "grad_norm": 0.9569286108016968,
      "learning_rate": 0.0005893330460079732,
      "loss": 4.6732,
      "step": 1750
    },
    {
      "epoch": 0.19373587342589602,
      "grad_norm": 0.9205374121665955,
      "learning_rate": 0.0005890098049779118,
      "loss": 4.6424,
      "step": 1800
    },
    {
      "epoch": 0.19911742546550426,
      "grad_norm": 1.1031231880187988,
      "learning_rate": 0.0005886865639478504,
      "loss": 4.6293,
      "step": 1850
    },
    {
      "epoch": 0.20449897750511248,
      "grad_norm": 1.0414140224456787,
      "learning_rate": 0.0005883633229177889,
      "loss": 4.6116,
      "step": 1900
    },
    {
      "epoch": 0.2098805295447207,
      "grad_norm": 0.8618186712265015,
      "learning_rate": 0.0005880400818877276,
      "loss": 4.5951,
      "step": 1950
    },
    {
      "epoch": 0.2152620815843289,
      "grad_norm": 0.8570399880409241,
      "learning_rate": 0.0005877168408576662,
      "loss": 4.5741,
      "step": 2000
    },
    {
      "epoch": 0.2152620815843289,
      "eval_accuracy": 0.27234704688513356,
      "eval_loss": 4.498940467834473,
      "eval_runtime": 184.347,
      "eval_samples_per_second": 97.702,
      "eval_steps_per_second": 6.108,
      "step": 2000
    },
    {
      "epoch": 0.22064363362393713,
      "grad_norm": 0.8821980357170105,
      "learning_rate": 0.0005873935998276048,
      "loss": 4.5753,
      "step": 2050
    },
    {
      "epoch": 0.22602518566354537,
      "grad_norm": 0.8663649559020996,
      "learning_rate": 0.0005870703587975433,
      "loss": 4.5247,
      "step": 2100
    },
    {
      "epoch": 0.2314067377031536,
      "grad_norm": 0.9565311670303345,
      "learning_rate": 0.0005867471177674818,
      "loss": 4.5213,
      "step": 2150
    },
    {
      "epoch": 0.2367882897427618,
      "grad_norm": 0.880970299243927,
      "learning_rate": 0.0005864238767374205,
      "loss": 4.5089,
      "step": 2200
    },
    {
      "epoch": 0.24216984178237003,
      "grad_norm": 0.9244751930236816,
      "learning_rate": 0.0005861006357073591,
      "loss": 4.4794,
      "step": 2250
    },
    {
      "epoch": 0.24755139382197827,
      "grad_norm": 0.824410617351532,
      "learning_rate": 0.0005857773946772977,
      "loss": 4.4594,
      "step": 2300
    },
    {
      "epoch": 0.2529329458615865,
      "grad_norm": 0.7963064908981323,
      "learning_rate": 0.0005854541536472362,
      "loss": 4.4382,
      "step": 2350
    },
    {
      "epoch": 0.2583144979011947,
      "grad_norm": 0.9127968549728394,
      "learning_rate": 0.0005851309126171749,
      "loss": 4.4464,
      "step": 2400
    },
    {
      "epoch": 0.2636960499408029,
      "grad_norm": 0.7743198275566101,
      "learning_rate": 0.0005848076715871134,
      "loss": 4.4167,
      "step": 2450
    },
    {
      "epoch": 0.26907760198041114,
      "grad_norm": 1.0491642951965332,
      "learning_rate": 0.000584484430557052,
      "loss": 4.4227,
      "step": 2500
    },
    {
      "epoch": 0.27445915402001936,
      "grad_norm": 0.9228188395500183,
      "learning_rate": 0.0005841611895269906,
      "loss": 4.4175,
      "step": 2550
    },
    {
      "epoch": 0.2798407060596276,
      "grad_norm": 0.837192177772522,
      "learning_rate": 0.0005838379484969291,
      "loss": 4.3804,
      "step": 2600
    },
    {
      "epoch": 0.2852222580992358,
      "grad_norm": 0.7594670653343201,
      "learning_rate": 0.0005835147074668678,
      "loss": 4.3766,
      "step": 2650
    },
    {
      "epoch": 0.29060381013884407,
      "grad_norm": 0.7400719523429871,
      "learning_rate": 0.0005831914664368063,
      "loss": 4.3444,
      "step": 2700
    },
    {
      "epoch": 0.2959853621784523,
      "grad_norm": 0.8986117243766785,
      "learning_rate": 0.0005828682254067449,
      "loss": 4.353,
      "step": 2750
    },
    {
      "epoch": 0.3013669142180605,
      "grad_norm": 0.8467971682548523,
      "learning_rate": 0.0005825449843766835,
      "loss": 4.3674,
      "step": 2800
    },
    {
      "epoch": 0.3067484662576687,
      "grad_norm": 0.6583292484283447,
      "learning_rate": 0.0005822217433466221,
      "loss": 4.356,
      "step": 2850
    },
    {
      "epoch": 0.31213001829727693,
      "grad_norm": 0.9095297455787659,
      "learning_rate": 0.0005818985023165607,
      "loss": 4.3273,
      "step": 2900
    },
    {
      "epoch": 0.31751157033688515,
      "grad_norm": 0.7011778354644775,
      "learning_rate": 0.0005815752612864992,
      "loss": 4.3155,
      "step": 2950
    },
    {
      "epoch": 0.32289312237649337,
      "grad_norm": 0.8573430776596069,
      "learning_rate": 0.0005812520202564378,
      "loss": 4.2905,
      "step": 3000
    },
    {
      "epoch": 0.32289312237649337,
      "eval_accuracy": 0.2991464989703509,
      "eval_loss": 4.2332539558410645,
      "eval_runtime": 184.5016,
      "eval_samples_per_second": 97.62,
      "eval_steps_per_second": 6.103,
      "step": 3000
    },
    {
      "epoch": 0.3282746744161016,
      "grad_norm": 0.6798554062843323,
      "learning_rate": 0.0005809287792263764,
      "loss": 4.3078,
      "step": 3050
    },
    {
      "epoch": 0.3336562264557098,
      "grad_norm": 1.0313102006912231,
      "learning_rate": 0.0005806055381963151,
      "loss": 4.2915,
      "step": 3100
    },
    {
      "epoch": 0.3390377784953181,
      "grad_norm": 0.7952976822853088,
      "learning_rate": 0.0005802822971662536,
      "loss": 4.267,
      "step": 3150
    },
    {
      "epoch": 0.3444193305349263,
      "grad_norm": 0.6104669570922852,
      "learning_rate": 0.0005799590561361922,
      "loss": 4.2788,
      "step": 3200
    },
    {
      "epoch": 0.3498008825745345,
      "grad_norm": 0.7965825796127319,
      "learning_rate": 0.0005796358151061307,
      "loss": 4.2722,
      "step": 3250
    },
    {
      "epoch": 0.35518243461414273,
      "grad_norm": 0.6662214398384094,
      "learning_rate": 0.0005793125740760694,
      "loss": 4.263,
      "step": 3300
    },
    {
      "epoch": 0.36056398665375095,
      "grad_norm": 0.7256644368171692,
      "learning_rate": 0.0005789893330460079,
      "loss": 4.2343,
      "step": 3350
    },
    {
      "epoch": 0.36594553869335916,
      "grad_norm": 0.6321602463722229,
      "learning_rate": 0.0005786660920159465,
      "loss": 4.2483,
      "step": 3400
    },
    {
      "epoch": 0.3713270907329674,
      "grad_norm": 0.6723715662956238,
      "learning_rate": 0.0005783428509858851,
      "loss": 4.2271,
      "step": 3450
    },
    {
      "epoch": 0.3767086427725756,
      "grad_norm": 0.757610023021698,
      "learning_rate": 0.0005780196099558237,
      "loss": 4.2201,
      "step": 3500
    },
    {
      "epoch": 0.3820901948121838,
      "grad_norm": 0.6822935938835144,
      "learning_rate": 0.0005776963689257623,
      "loss": 4.2124,
      "step": 3550
    },
    {
      "epoch": 0.38747174685179203,
      "grad_norm": 0.6186670660972595,
      "learning_rate": 0.0005773731278957008,
      "loss": 4.2126,
      "step": 3600
    },
    {
      "epoch": 0.3928532988914003,
      "grad_norm": 0.7107723355293274,
      "learning_rate": 0.0005770498868656394,
      "loss": 4.1955,
      "step": 3650
    },
    {
      "epoch": 0.3982348509310085,
      "grad_norm": 0.6134611964225769,
      "learning_rate": 0.000576726645835578,
      "loss": 4.187,
      "step": 3700
    },
    {
      "epoch": 0.40361640297061674,
      "grad_norm": 0.7130006551742554,
      "learning_rate": 0.0005764034048055167,
      "loss": 4.2103,
      "step": 3750
    },
    {
      "epoch": 0.40899795501022496,
      "grad_norm": 0.7081485986709595,
      "learning_rate": 0.0005760801637754552,
      "loss": 4.1975,
      "step": 3800
    },
    {
      "epoch": 0.4143795070498332,
      "grad_norm": 0.8055673241615295,
      "learning_rate": 0.0005757569227453937,
      "loss": 4.1592,
      "step": 3850
    },
    {
      "epoch": 0.4197610590894414,
      "grad_norm": 0.9049285054206848,
      "learning_rate": 0.0005754336817153324,
      "loss": 4.155,
      "step": 3900
    },
    {
      "epoch": 0.4251426111290496,
      "grad_norm": 0.7098596096038818,
      "learning_rate": 0.0005751104406852709,
      "loss": 4.1541,
      "step": 3950
    },
    {
      "epoch": 0.4305241631686578,
      "grad_norm": 0.6744562387466431,
      "learning_rate": 0.0005747871996552096,
      "loss": 4.1453,
      "step": 4000
    },
    {
      "epoch": 0.4305241631686578,
      "eval_accuracy": 0.3129022805263276,
      "eval_loss": 4.090099811553955,
      "eval_runtime": 184.4208,
      "eval_samples_per_second": 97.663,
      "eval_steps_per_second": 6.106,
      "step": 4000
    },
    {
      "epoch": 0.43590571520826604,
      "grad_norm": 0.6521379947662354,
      "learning_rate": 0.0005744639586251481,
      "loss": 4.1527,
      "step": 4050
    },
    {
      "epoch": 0.44128726724787426,
      "grad_norm": 0.6583467721939087,
      "learning_rate": 0.0005741407175950867,
      "loss": 4.1572,
      "step": 4100
    },
    {
      "epoch": 0.44666881928748253,
      "grad_norm": 0.6938270330429077,
      "learning_rate": 0.0005738174765650253,
      "loss": 4.1557,
      "step": 4150
    },
    {
      "epoch": 0.45205037132709075,
      "grad_norm": 0.6435059309005737,
      "learning_rate": 0.0005734942355349638,
      "loss": 4.1395,
      "step": 4200
    },
    {
      "epoch": 0.45743192336669897,
      "grad_norm": 0.6761099696159363,
      "learning_rate": 0.0005731709945049025,
      "loss": 4.1328,
      "step": 4250
    },
    {
      "epoch": 0.4628134754063072,
      "grad_norm": 0.6372486352920532,
      "learning_rate": 0.000572847753474841,
      "loss": 4.138,
      "step": 4300
    },
    {
      "epoch": 0.4681950274459154,
      "grad_norm": 0.7407318949699402,
      "learning_rate": 0.0005725245124447796,
      "loss": 4.1446,
      "step": 4350
    },
    {
      "epoch": 0.4735765794855236,
      "grad_norm": 0.6796604990959167,
      "learning_rate": 0.0005722012714147182,
      "loss": 4.1164,
      "step": 4400
    },
    {
      "epoch": 0.47895813152513184,
      "grad_norm": 0.6598204970359802,
      "learning_rate": 0.0005718780303846568,
      "loss": 4.1197,
      "step": 4450
    },
    {
      "epoch": 0.48433968356474005,
      "grad_norm": 0.6131014823913574,
      "learning_rate": 0.0005715547893545953,
      "loss": 4.106,
      "step": 4500
    },
    {
      "epoch": 0.48972123560434827,
      "grad_norm": 0.6856024265289307,
      "learning_rate": 0.000571231548324534,
      "loss": 4.0877,
      "step": 4550
    },
    {
      "epoch": 0.49510278764395654,
      "grad_norm": 0.6875824928283691,
      "learning_rate": 0.0005709083072944725,
      "loss": 4.0935,
      "step": 4600
    },
    {
      "epoch": 0.5004843396835648,
      "grad_norm": 0.5760625004768372,
      "learning_rate": 0.0005705850662644111,
      "loss": 4.1158,
      "step": 4650
    },
    {
      "epoch": 0.505865891723173,
      "grad_norm": 0.8604832291603088,
      "learning_rate": 0.0005702618252343497,
      "loss": 4.0911,
      "step": 4700
    },
    {
      "epoch": 0.5112474437627812,
      "grad_norm": 0.6754705905914307,
      "learning_rate": 0.0005699385842042882,
      "loss": 4.1016,
      "step": 4750
    },
    {
      "epoch": 0.5166289958023894,
      "grad_norm": 0.5712053775787354,
      "learning_rate": 0.0005696153431742269,
      "loss": 4.0716,
      "step": 4800
    },
    {
      "epoch": 0.5220105478419976,
      "grad_norm": 0.6515219807624817,
      "learning_rate": 0.0005692921021441655,
      "loss": 4.0726,
      "step": 4850
    },
    {
      "epoch": 0.5273920998816058,
      "grad_norm": 0.6000773906707764,
      "learning_rate": 0.0005689688611141041,
      "loss": 4.0808,
      "step": 4900
    },
    {
      "epoch": 0.5327736519212141,
      "grad_norm": 0.6588471531867981,
      "learning_rate": 0.0005686456200840426,
      "loss": 4.0791,
      "step": 4950
    },
    {
      "epoch": 0.5381552039608223,
      "grad_norm": 0.6547403931617737,
      "learning_rate": 0.0005683223790539811,
      "loss": 4.0703,
      "step": 5000
    },
    {
      "epoch": 0.5381552039608223,
      "eval_accuracy": 0.3209675843887965,
      "eval_loss": 3.9936656951904297,
      "eval_runtime": 184.5678,
      "eval_samples_per_second": 97.585,
      "eval_steps_per_second": 6.101,
      "step": 5000
    },
    {
      "epoch": 0.5435367560004305,
      "grad_norm": 0.649213969707489,
      "learning_rate": 0.0005679991380239198,
      "loss": 4.0619,
      "step": 5050
    },
    {
      "epoch": 0.5489183080400387,
      "grad_norm": 0.6468812823295593,
      "learning_rate": 0.0005676758969938584,
      "loss": 4.0439,
      "step": 5100
    },
    {
      "epoch": 0.5542998600796469,
      "grad_norm": 0.6261686682701111,
      "learning_rate": 0.000567352655963797,
      "loss": 4.0405,
      "step": 5150
    },
    {
      "epoch": 0.5596814121192552,
      "grad_norm": 0.7181739211082458,
      "learning_rate": 0.0005670294149337355,
      "loss": 4.0494,
      "step": 5200
    },
    {
      "epoch": 0.5650629641588634,
      "grad_norm": 0.5509181022644043,
      "learning_rate": 0.0005667061739036742,
      "loss": 4.0364,
      "step": 5250
    },
    {
      "epoch": 0.5704445161984716,
      "grad_norm": 0.7940995693206787,
      "learning_rate": 0.0005663829328736127,
      "loss": 4.0348,
      "step": 5300
    },
    {
      "epoch": 0.5758260682380799,
      "grad_norm": 0.6015467047691345,
      "learning_rate": 0.0005660596918435512,
      "loss": 4.0591,
      "step": 5350
    },
    {
      "epoch": 0.5812076202776881,
      "grad_norm": 0.6166031360626221,
      "learning_rate": 0.0005657364508134899,
      "loss": 4.0306,
      "step": 5400
    },
    {
      "epoch": 0.5865891723172963,
      "grad_norm": 0.7661638259887695,
      "learning_rate": 0.0005654132097834284,
      "loss": 4.0307,
      "step": 5450
    },
    {
      "epoch": 0.5919707243569046,
      "grad_norm": 0.6555174589157104,
      "learning_rate": 0.0005650899687533671,
      "loss": 4.024,
      "step": 5500
    },
    {
      "epoch": 0.5973522763965128,
      "grad_norm": 0.595856785774231,
      "learning_rate": 0.0005647667277233056,
      "loss": 4.0182,
      "step": 5550
    },
    {
      "epoch": 0.602733828436121,
      "grad_norm": 0.6452604532241821,
      "learning_rate": 0.0005644434866932442,
      "loss": 4.0241,
      "step": 5600
    },
    {
      "epoch": 0.6081153804757292,
      "grad_norm": 0.6375003457069397,
      "learning_rate": 0.0005641202456631828,
      "loss": 3.9957,
      "step": 5650
    },
    {
      "epoch": 0.6134969325153374,
      "grad_norm": 0.6434177160263062,
      "learning_rate": 0.0005637970046331214,
      "loss": 3.9934,
      "step": 5700
    },
    {
      "epoch": 0.6188784845549457,
      "grad_norm": 0.6278543472290039,
      "learning_rate": 0.00056347376360306,
      "loss": 4.0028,
      "step": 5750
    },
    {
      "epoch": 0.6242600365945539,
      "grad_norm": 0.6161077618598938,
      "learning_rate": 0.0005631505225729985,
      "loss": 4.0068,
      "step": 5800
    },
    {
      "epoch": 0.6296415886341621,
      "grad_norm": 0.5981731414794922,
      "learning_rate": 0.0005628272815429371,
      "loss": 4.0076,
      "step": 5850
    },
    {
      "epoch": 0.6350231406737703,
      "grad_norm": 0.5853173136711121,
      "learning_rate": 0.0005625040405128757,
      "loss": 4.0131,
      "step": 5900
    },
    {
      "epoch": 0.6404046927133785,
      "grad_norm": 0.7181535959243774,
      "learning_rate": 0.0005621807994828143,
      "loss": 3.9811,
      "step": 5950
    },
    {
      "epoch": 0.6457862447529867,
      "grad_norm": 0.7647160887718201,
      "learning_rate": 0.0005618575584527529,
      "loss": 3.9912,
      "step": 6000
    },
    {
      "epoch": 0.6457862447529867,
      "eval_accuracy": 0.32738592777777353,
      "eval_loss": 3.926551342010498,
      "eval_runtime": 184.2978,
      "eval_samples_per_second": 97.728,
      "eval_steps_per_second": 6.11,
      "step": 6000
    },
    {
      "epoch": 0.651167796792595,
      "grad_norm": 0.6108874082565308,
      "learning_rate": 0.0005615343174226915,
      "loss": 3.9942,
      "step": 6050
    },
    {
      "epoch": 0.6565493488322032,
      "grad_norm": 0.539562463760376,
      "learning_rate": 0.00056121107639263,
      "loss": 3.9947,
      "step": 6100
    },
    {
      "epoch": 0.6619309008718114,
      "grad_norm": 0.6353080868721008,
      "learning_rate": 0.0005608878353625687,
      "loss": 3.9713,
      "step": 6150
    },
    {
      "epoch": 0.6673124529114196,
      "grad_norm": 0.6244759559631348,
      "learning_rate": 0.0005605645943325072,
      "loss": 3.9771,
      "step": 6200
    },
    {
      "epoch": 0.6726940049510278,
      "grad_norm": 0.5772109627723694,
      "learning_rate": 0.0005602413533024458,
      "loss": 3.9849,
      "step": 6250
    },
    {
      "epoch": 0.6780755569906362,
      "grad_norm": 0.5281596183776855,
      "learning_rate": 0.0005599181122723844,
      "loss": 3.9597,
      "step": 6300
    },
    {
      "epoch": 0.6834571090302444,
      "grad_norm": 0.5710061192512512,
      "learning_rate": 0.000559594871242323,
      "loss": 3.9619,
      "step": 6350
    },
    {
      "epoch": 0.6888386610698526,
      "grad_norm": 0.605676531791687,
      "learning_rate": 0.0005592716302122616,
      "loss": 3.9618,
      "step": 6400
    },
    {
      "epoch": 0.6942202131094608,
      "grad_norm": 0.6223154067993164,
      "learning_rate": 0.0005589483891822001,
      "loss": 3.9637,
      "step": 6450
    },
    {
      "epoch": 0.699601765149069,
      "grad_norm": 0.776096761226654,
      "learning_rate": 0.0005586251481521387,
      "loss": 3.9487,
      "step": 6500
    },
    {
      "epoch": 0.7049833171886772,
      "grad_norm": 0.5951947569847107,
      "learning_rate": 0.0005583019071220773,
      "loss": 3.9713,
      "step": 6550
    },
    {
      "epoch": 0.7103648692282855,
      "grad_norm": 0.6330660581588745,
      "learning_rate": 0.000557978666092016,
      "loss": 3.9564,
      "step": 6600
    },
    {
      "epoch": 0.7157464212678937,
      "grad_norm": 0.6448531746864319,
      "learning_rate": 0.0005576554250619545,
      "loss": 3.9346,
      "step": 6650
    },
    {
      "epoch": 0.7211279733075019,
      "grad_norm": 0.6924978494644165,
      "learning_rate": 0.000557332184031893,
      "loss": 3.9377,
      "step": 6700
    },
    {
      "epoch": 0.7265095253471101,
      "grad_norm": 0.7174770832061768,
      "learning_rate": 0.0005570089430018317,
      "loss": 3.9641,
      "step": 6750
    },
    {
      "epoch": 0.7318910773867183,
      "grad_norm": 0.5341421365737915,
      "learning_rate": 0.0005566857019717702,
      "loss": 3.9428,
      "step": 6800
    },
    {
      "epoch": 0.7372726294263265,
      "grad_norm": 0.6674601435661316,
      "learning_rate": 0.0005563624609417089,
      "loss": 3.9446,
      "step": 6850
    },
    {
      "epoch": 0.7426541814659348,
      "grad_norm": 0.6375956535339355,
      "learning_rate": 0.0005560392199116474,
      "loss": 3.9471,
      "step": 6900
    },
    {
      "epoch": 0.748035733505543,
      "grad_norm": 0.6670394539833069,
      "learning_rate": 0.000555715978881586,
      "loss": 3.9303,
      "step": 6950
    },
    {
      "epoch": 0.7534172855451512,
      "grad_norm": 0.7189449667930603,
      "learning_rate": 0.0005553927378515246,
      "loss": 3.9195,
      "step": 7000
    },
    {
      "epoch": 0.7534172855451512,
      "eval_accuracy": 0.3326128922518648,
      "eval_loss": 3.872013568878174,
      "eval_runtime": 184.7751,
      "eval_samples_per_second": 97.475,
      "eval_steps_per_second": 6.094,
      "step": 7000
    },
    {
      "epoch": 0.7587988375847594,
      "grad_norm": 0.6056647896766663,
      "learning_rate": 0.0005550694968214631,
      "loss": 3.9423,
      "step": 7050
    },
    {
      "epoch": 0.7641803896243676,
      "grad_norm": 0.6418593525886536,
      "learning_rate": 0.0005547462557914018,
      "loss": 3.9169,
      "step": 7100
    },
    {
      "epoch": 0.7695619416639758,
      "grad_norm": 0.5867264866828918,
      "learning_rate": 0.0005544230147613403,
      "loss": 3.9172,
      "step": 7150
    },
    {
      "epoch": 0.7749434937035841,
      "grad_norm": 0.5679484009742737,
      "learning_rate": 0.0005540997737312789,
      "loss": 3.9287,
      "step": 7200
    },
    {
      "epoch": 0.7803250457431924,
      "grad_norm": 0.6049643754959106,
      "learning_rate": 0.0005537829975218188,
      "loss": 3.9073,
      "step": 7250
    },
    {
      "epoch": 0.7857065977828006,
      "grad_norm": 0.5454378724098206,
      "learning_rate": 0.0005534597564917573,
      "loss": 3.931,
      "step": 7300
    },
    {
      "epoch": 0.7910881498224088,
      "grad_norm": 0.6361756324768066,
      "learning_rate": 0.0005531365154616959,
      "loss": 3.9185,
      "step": 7350
    },
    {
      "epoch": 0.796469701862017,
      "grad_norm": 0.5948432087898254,
      "learning_rate": 0.0005528132744316344,
      "loss": 3.9223,
      "step": 7400
    },
    {
      "epoch": 0.8018512539016253,
      "grad_norm": 0.6710109114646912,
      "learning_rate": 0.0005524900334015731,
      "loss": 3.9103,
      "step": 7450
    },
    {
      "epoch": 0.8072328059412335,
      "grad_norm": 0.6390253305435181,
      "learning_rate": 0.0005521667923715117,
      "loss": 3.8857,
      "step": 7500
    },
    {
      "epoch": 0.8126143579808417,
      "grad_norm": 0.6726111173629761,
      "learning_rate": 0.0005518435513414502,
      "loss": 3.8906,
      "step": 7550
    },
    {
      "epoch": 0.8179959100204499,
      "grad_norm": 0.5953141450881958,
      "learning_rate": 0.0005515203103113888,
      "loss": 3.911,
      "step": 7600
    },
    {
      "epoch": 0.8233774620600581,
      "grad_norm": 0.6084983348846436,
      "learning_rate": 0.0005511970692813274,
      "loss": 3.8937,
      "step": 7650
    },
    {
      "epoch": 0.8287590140996663,
      "grad_norm": 0.6089417338371277,
      "learning_rate": 0.000550873828251266,
      "loss": 3.907,
      "step": 7700
    },
    {
      "epoch": 0.8341405661392746,
      "grad_norm": 0.5650714635848999,
      "learning_rate": 0.0005505505872212045,
      "loss": 3.9024,
      "step": 7750
    },
    {
      "epoch": 0.8395221181788828,
      "grad_norm": 0.556892991065979,
      "learning_rate": 0.0005502273461911432,
      "loss": 3.8743,
      "step": 7800
    },
    {
      "epoch": 0.844903670218491,
      "grad_norm": 0.5664134621620178,
      "learning_rate": 0.0005499041051610817,
      "loss": 3.913,
      "step": 7850
    },
    {
      "epoch": 0.8502852222580992,
      "grad_norm": 0.6096441745758057,
      "learning_rate": 0.0005495808641310204,
      "loss": 3.879,
      "step": 7900
    },
    {
      "epoch": 0.8556667742977074,
      "grad_norm": 0.5594208240509033,
      "learning_rate": 0.0005492576231009589,
      "loss": 3.8784,
      "step": 7950
    },
    {
      "epoch": 0.8610483263373157,
      "grad_norm": 0.538360595703125,
      "learning_rate": 0.0005489343820708974,
      "loss": 3.8902,
      "step": 8000
    },
    {
      "epoch": 0.8610483263373157,
      "eval_accuracy": 0.3369882353912661,
      "eval_loss": 3.822113037109375,
      "eval_runtime": 184.4529,
      "eval_samples_per_second": 97.646,
      "eval_steps_per_second": 6.105,
      "step": 8000
    },
    {
      "epoch": 0.8664298783769239,
      "grad_norm": 0.6012279987335205,
      "learning_rate": 0.0005486111410408361,
      "loss": 3.8879,
      "step": 8050
    },
    {
      "epoch": 0.8718114304165321,
      "grad_norm": 0.5948436260223389,
      "learning_rate": 0.0005482879000107746,
      "loss": 3.8825,
      "step": 8100
    },
    {
      "epoch": 0.8771929824561403,
      "grad_norm": 0.6777280569076538,
      "learning_rate": 0.0005479646589807133,
      "loss": 3.8714,
      "step": 8150
    },
    {
      "epoch": 0.8825745344957485,
      "grad_norm": 0.5959411859512329,
      "learning_rate": 0.0005476414179506518,
      "loss": 3.8722,
      "step": 8200
    },
    {
      "epoch": 0.8879560865353568,
      "grad_norm": 0.5966766476631165,
      "learning_rate": 0.0005473181769205904,
      "loss": 3.877,
      "step": 8250
    },
    {
      "epoch": 0.8933376385749651,
      "grad_norm": 0.5824183225631714,
      "learning_rate": 0.000546994935890529,
      "loss": 3.8932,
      "step": 8300
    },
    {
      "epoch": 0.8987191906145733,
      "grad_norm": 0.565321147441864,
      "learning_rate": 0.0005466716948604677,
      "loss": 3.8744,
      "step": 8350
    },
    {
      "epoch": 0.9041007426541815,
      "grad_norm": 0.6127881407737732,
      "learning_rate": 0.0005463484538304062,
      "loss": 3.8627,
      "step": 8400
    },
    {
      "epoch": 0.9094822946937897,
      "grad_norm": 0.5455643534660339,
      "learning_rate": 0.0005460252128003447,
      "loss": 3.8711,
      "step": 8450
    },
    {
      "epoch": 0.9148638467333979,
      "grad_norm": 0.5501160621643066,
      "learning_rate": 0.0005457019717702833,
      "loss": 3.8719,
      "step": 8500
    },
    {
      "epoch": 0.9202453987730062,
      "grad_norm": 0.5473130941390991,
      "learning_rate": 0.0005453787307402219,
      "loss": 3.8563,
      "step": 8550
    },
    {
      "epoch": 0.9256269508126144,
      "grad_norm": 0.5105751156806946,
      "learning_rate": 0.0005450554897101605,
      "loss": 3.8604,
      "step": 8600
    },
    {
      "epoch": 0.9310085028522226,
      "grad_norm": 0.6234155893325806,
      "learning_rate": 0.0005447322486800991,
      "loss": 3.8682,
      "step": 8650
    },
    {
      "epoch": 0.9363900548918308,
      "grad_norm": 0.6187501549720764,
      "learning_rate": 0.0005444090076500377,
      "loss": 3.8721,
      "step": 8700
    },
    {
      "epoch": 0.941771606931439,
      "grad_norm": 0.563278317451477,
      "learning_rate": 0.0005440857666199763,
      "loss": 3.8535,
      "step": 8750
    },
    {
      "epoch": 0.9471531589710472,
      "grad_norm": 0.5937185883522034,
      "learning_rate": 0.0005437625255899148,
      "loss": 3.868,
      "step": 8800
    },
    {
      "epoch": 0.9525347110106555,
      "grad_norm": 0.592932403087616,
      "learning_rate": 0.0005434392845598534,
      "loss": 3.8573,
      "step": 8850
    },
    {
      "epoch": 0.9579162630502637,
      "grad_norm": 0.6201167106628418,
      "learning_rate": 0.000543116043529792,
      "loss": 3.8481,
      "step": 8900
    },
    {
      "epoch": 0.9632978150898719,
      "grad_norm": 0.554725706577301,
      "learning_rate": 0.0005427928024997306,
      "loss": 3.8541,
      "step": 8950
    },
    {
      "epoch": 0.9686793671294801,
      "grad_norm": 0.5520325899124146,
      "learning_rate": 0.0005424695614696692,
      "loss": 3.85,
      "step": 9000
    },
    {
      "epoch": 0.9686793671294801,
      "eval_accuracy": 0.3408027123237691,
      "eval_loss": 3.7828752994537354,
      "eval_runtime": 184.162,
      "eval_samples_per_second": 97.8,
      "eval_steps_per_second": 6.114,
      "step": 9000
    },
    {
      "epoch": 0.9740609191690883,
      "grad_norm": 0.5292057394981384,
      "learning_rate": 0.0005421463204396078,
      "loss": 3.8381,
      "step": 9050
    },
    {
      "epoch": 0.9794424712086965,
      "grad_norm": 0.5707278847694397,
      "learning_rate": 0.0005418230794095463,
      "loss": 3.8459,
      "step": 9100
    },
    {
      "epoch": 0.9848240232483048,
      "grad_norm": 0.5048800110816956,
      "learning_rate": 0.000541499838379485,
      "loss": 3.8483,
      "step": 9150
    },
    {
      "epoch": 0.9902055752879131,
      "grad_norm": 0.6456687450408936,
      "learning_rate": 0.0005411765973494235,
      "loss": 3.8333,
      "step": 9200
    },
    {
      "epoch": 0.9955871273275213,
      "grad_norm": 0.6426123976707458,
      "learning_rate": 0.0005408533563193621,
      "loss": 3.8645,
      "step": 9250
    },
    {
      "epoch": 1.0009686793671295,
      "grad_norm": 0.6812326908111572,
      "learning_rate": 0.0005405301152893007,
      "loss": 3.8295,
      "step": 9300
    },
    {
      "epoch": 1.0063502314067376,
      "grad_norm": 0.5232545733451843,
      "learning_rate": 0.0005402068742592392,
      "loss": 3.7647,
      "step": 9350
    },
    {
      "epoch": 1.011731783446346,
      "grad_norm": 0.5876726508140564,
      "learning_rate": 0.0005398836332291779,
      "loss": 3.7793,
      "step": 9400
    },
    {
      "epoch": 1.017113335485954,
      "grad_norm": 0.5287936925888062,
      "learning_rate": 0.0005395603921991164,
      "loss": 3.7686,
      "step": 9450
    },
    {
      "epoch": 1.0224948875255624,
      "grad_norm": 0.5465691089630127,
      "learning_rate": 0.0005392436159896562,
      "loss": 3.7749,
      "step": 9500
    },
    {
      "epoch": 1.0278764395651705,
      "grad_norm": 0.5962051749229431,
      "learning_rate": 0.0005389203749595948,
      "loss": 3.7751,
      "step": 9550
    },
    {
      "epoch": 1.0332579916047788,
      "grad_norm": 0.543983519077301,
      "learning_rate": 0.0005385971339295334,
      "loss": 3.7671,
      "step": 9600
    },
    {
      "epoch": 1.0386395436443872,
      "grad_norm": 0.5294556021690369,
      "learning_rate": 0.000538273892899472,
      "loss": 3.7737,
      "step": 9650
    },
    {
      "epoch": 1.0440210956839953,
      "grad_norm": 0.6221946477890015,
      "learning_rate": 0.0005379506518694106,
      "loss": 3.7698,
      "step": 9700
    },
    {
      "epoch": 1.0494026477236036,
      "grad_norm": 0.5739998817443848,
      "learning_rate": 0.0005376274108393491,
      "loss": 3.7579,
      "step": 9750
    },
    {
      "epoch": 1.0547841997632117,
      "grad_norm": 0.6448442339897156,
      "learning_rate": 0.0005373041698092877,
      "loss": 3.779,
      "step": 9800
    },
    {
      "epoch": 1.06016575180282,
      "grad_norm": 0.5370635986328125,
      "learning_rate": 0.0005369809287792263,
      "loss": 3.7778,
      "step": 9850
    },
    {
      "epoch": 1.0655473038424281,
      "grad_norm": 0.5538362860679626,
      "learning_rate": 0.000536657687749165,
      "loss": 3.7705,
      "step": 9900
    },
    {
      "epoch": 1.0709288558820365,
      "grad_norm": 0.5770606994628906,
      "learning_rate": 0.0005363344467191035,
      "loss": 3.7701,
      "step": 9950
    },
    {
      "epoch": 1.0763104079216446,
      "grad_norm": 0.5529940128326416,
      "learning_rate": 0.000536011205689042,
      "loss": 3.768,
      "step": 10000
    },
    {
      "epoch": 1.0763104079216446,
      "eval_accuracy": 0.3442798220396081,
      "eval_loss": 3.7558341026306152,
      "eval_runtime": 184.4409,
      "eval_samples_per_second": 97.652,
      "eval_steps_per_second": 6.105,
      "step": 10000
    },
    {
      "epoch": 1.081691959961253,
      "grad_norm": 0.500934362411499,
      "learning_rate": 0.0005356879646589807,
      "loss": 3.7708,
      "step": 10050
    },
    {
      "epoch": 1.087073512000861,
      "grad_norm": 0.5757973790168762,
      "learning_rate": 0.0005353647236289192,
      "loss": 3.7773,
      "step": 10100
    },
    {
      "epoch": 1.0924550640404693,
      "grad_norm": 0.580998420715332,
      "learning_rate": 0.0005350414825988579,
      "loss": 3.7801,
      "step": 10150
    },
    {
      "epoch": 1.0978366160800774,
      "grad_norm": 0.5978397727012634,
      "learning_rate": 0.0005347182415687964,
      "loss": 3.7486,
      "step": 10200
    },
    {
      "epoch": 1.1032181681196858,
      "grad_norm": 0.5137054324150085,
      "learning_rate": 0.000534395000538735,
      "loss": 3.759,
      "step": 10250
    },
    {
      "epoch": 1.1085997201592939,
      "grad_norm": 0.5786076188087463,
      "learning_rate": 0.0005340717595086736,
      "loss": 3.7711,
      "step": 10300
    },
    {
      "epoch": 1.1139812721989022,
      "grad_norm": 0.6630505919456482,
      "learning_rate": 0.0005337485184786122,
      "loss": 3.7715,
      "step": 10350
    },
    {
      "epoch": 1.1193628242385103,
      "grad_norm": 0.5464471578598022,
      "learning_rate": 0.0005334252774485507,
      "loss": 3.7483,
      "step": 10400
    },
    {
      "epoch": 1.1247443762781186,
      "grad_norm": 0.5084022283554077,
      "learning_rate": 0.0005331020364184894,
      "loss": 3.7667,
      "step": 10450
    },
    {
      "epoch": 1.1301259283177267,
      "grad_norm": 0.5448545217514038,
      "learning_rate": 0.0005327787953884279,
      "loss": 3.7683,
      "step": 10500
    },
    {
      "epoch": 1.135507480357335,
      "grad_norm": 0.5716451406478882,
      "learning_rate": 0.0005324555543583665,
      "loss": 3.7649,
      "step": 10550
    },
    {
      "epoch": 1.1408890323969434,
      "grad_norm": 0.6230728626251221,
      "learning_rate": 0.0005321323133283051,
      "loss": 3.7639,
      "step": 10600
    },
    {
      "epoch": 1.1462705844365515,
      "grad_norm": 0.5730423927307129,
      "learning_rate": 0.0005318090722982436,
      "loss": 3.7511,
      "step": 10650
    },
    {
      "epoch": 1.1516521364761596,
      "grad_norm": 0.5485336780548096,
      "learning_rate": 0.0005314858312681823,
      "loss": 3.7601,
      "step": 10700
    },
    {
      "epoch": 1.157033688515768,
      "grad_norm": 0.6366468071937561,
      "learning_rate": 0.0005311625902381209,
      "loss": 3.7495,
      "step": 10750
    },
    {
      "epoch": 1.1624152405553763,
      "grad_norm": 0.5823078751564026,
      "learning_rate": 0.0005308393492080595,
      "loss": 3.7517,
      "step": 10800
    },
    {
      "epoch": 1.1677967925949844,
      "grad_norm": 0.5976313352584839,
      "learning_rate": 0.000530516108177998,
      "loss": 3.7573,
      "step": 10850
    },
    {
      "epoch": 1.1731783446345927,
      "grad_norm": 0.5455477237701416,
      "learning_rate": 0.0005301928671479365,
      "loss": 3.7558,
      "step": 10900
    },
    {
      "epoch": 1.1785598966742008,
      "grad_norm": 0.5552336573600769,
      "learning_rate": 0.0005298696261178752,
      "loss": 3.7501,
      "step": 10950
    },
    {
      "epoch": 1.1839414487138091,
      "grad_norm": 0.5553588271141052,
      "learning_rate": 0.0005295463850878138,
      "loss": 3.7642,
      "step": 11000
    },
    {
      "epoch": 1.1839414487138091,
      "eval_accuracy": 0.34638942650941407,
      "eval_loss": 3.727562427520752,
      "eval_runtime": 184.3893,
      "eval_samples_per_second": 97.679,
      "eval_steps_per_second": 6.107,
      "step": 11000
    },
    {
      "epoch": 1.1893230007534172,
      "grad_norm": 0.5498607754707336,
      "learning_rate": 0.0005292231440577524,
      "loss": 3.7585,
      "step": 11050
    },
    {
      "epoch": 1.1947045527930256,
      "grad_norm": 0.5491752624511719,
      "learning_rate": 0.0005288999030276909,
      "loss": 3.7383,
      "step": 11100
    },
    {
      "epoch": 1.2000861048326337,
      "grad_norm": 0.6076170206069946,
      "learning_rate": 0.0005285766619976295,
      "loss": 3.7506,
      "step": 11150
    },
    {
      "epoch": 1.205467656872242,
      "grad_norm": 0.5911027193069458,
      "learning_rate": 0.0005282534209675681,
      "loss": 3.7489,
      "step": 11200
    },
    {
      "epoch": 1.21084920891185,
      "grad_norm": 0.5287796854972839,
      "learning_rate": 0.0005279301799375066,
      "loss": 3.7516,
      "step": 11250
    },
    {
      "epoch": 1.2162307609514584,
      "grad_norm": 0.5889896750450134,
      "learning_rate": 0.0005276069389074453,
      "loss": 3.7365,
      "step": 11300
    },
    {
      "epoch": 1.2216123129910665,
      "grad_norm": 0.5834380388259888,
      "learning_rate": 0.0005272836978773838,
      "loss": 3.7473,
      "step": 11350
    },
    {
      "epoch": 1.2269938650306749,
      "grad_norm": 0.564867377281189,
      "learning_rate": 0.0005269604568473225,
      "loss": 3.7242,
      "step": 11400
    },
    {
      "epoch": 1.232375417070283,
      "grad_norm": 0.5929306745529175,
      "learning_rate": 0.000526637215817261,
      "loss": 3.7374,
      "step": 11450
    },
    {
      "epoch": 1.2377569691098913,
      "grad_norm": 0.5987567901611328,
      "learning_rate": 0.0005263139747871996,
      "loss": 3.748,
      "step": 11500
    },
    {
      "epoch": 1.2431385211494996,
      "grad_norm": 0.5246814489364624,
      "learning_rate": 0.0005259907337571381,
      "loss": 3.7468,
      "step": 11550
    },
    {
      "epoch": 1.2485200731891077,
      "grad_norm": 0.6484177112579346,
      "learning_rate": 0.0005256674927270768,
      "loss": 3.7486,
      "step": 11600
    },
    {
      "epoch": 1.2539016252287158,
      "grad_norm": 0.558538019657135,
      "learning_rate": 0.0005253442516970154,
      "loss": 3.7339,
      "step": 11650
    },
    {
      "epoch": 1.2592831772683242,
      "grad_norm": 0.5690079927444458,
      "learning_rate": 0.000525021010666954,
      "loss": 3.7292,
      "step": 11700
    },
    {
      "epoch": 1.2646647293079325,
      "grad_norm": 0.6083813309669495,
      "learning_rate": 0.0005246977696368925,
      "loss": 3.7407,
      "step": 11750
    },
    {
      "epoch": 1.2700462813475406,
      "grad_norm": 0.5509043335914612,
      "learning_rate": 0.0005243745286068311,
      "loss": 3.7299,
      "step": 11800
    },
    {
      "epoch": 1.275427833387149,
      "grad_norm": 0.5929333567619324,
      "learning_rate": 0.0005240512875767697,
      "loss": 3.7367,
      "step": 11850
    },
    {
      "epoch": 1.280809385426757,
      "grad_norm": 0.5754917860031128,
      "learning_rate": 0.0005237280465467083,
      "loss": 3.75,
      "step": 11900
    },
    {
      "epoch": 1.2861909374663654,
      "grad_norm": 0.6693816781044006,
      "learning_rate": 0.0005234048055166469,
      "loss": 3.7461,
      "step": 11950
    },
    {
      "epoch": 1.2915724895059735,
      "grad_norm": 0.5702859163284302,
      "learning_rate": 0.0005230815644865854,
      "loss": 3.7662,
      "step": 12000
    },
    {
      "epoch": 1.2915724895059735,
      "eval_accuracy": 0.34887518727683375,
      "eval_loss": 3.7015981674194336,
      "eval_runtime": 184.6259,
      "eval_samples_per_second": 97.554,
      "eval_steps_per_second": 6.099,
      "step": 12000
    },
    {
      "epoch": 1.2969540415455818,
      "grad_norm": 0.5472993850708008,
      "learning_rate": 0.0005227583234565241,
      "loss": 3.7416,
      "step": 12050
    },
    {
      "epoch": 1.30233559358519,
      "grad_norm": 0.5625805258750916,
      "learning_rate": 0.0005224350824264626,
      "loss": 3.7177,
      "step": 12100
    },
    {
      "epoch": 1.3077171456247982,
      "grad_norm": 0.6123823523521423,
      "learning_rate": 0.0005221118413964012,
      "loss": 3.7213,
      "step": 12150
    },
    {
      "epoch": 1.3130986976644063,
      "grad_norm": 0.5814079642295837,
      "learning_rate": 0.0005217886003663398,
      "loss": 3.7236,
      "step": 12200
    },
    {
      "epoch": 1.3184802497040147,
      "grad_norm": 0.5321001410484314,
      "learning_rate": 0.0005214653593362783,
      "loss": 3.743,
      "step": 12250
    },
    {
      "epoch": 1.3238618017436228,
      "grad_norm": 0.5726140737533569,
      "learning_rate": 0.000521142118306217,
      "loss": 3.7532,
      "step": 12300
    },
    {
      "epoch": 1.329243353783231,
      "grad_norm": 0.6130521893501282,
      "learning_rate": 0.0005208188772761555,
      "loss": 3.7071,
      "step": 12350
    },
    {
      "epoch": 1.3346249058228392,
      "grad_norm": 0.5689705014228821,
      "learning_rate": 0.0005204956362460941,
      "loss": 3.7282,
      "step": 12400
    },
    {
      "epoch": 1.3400064578624475,
      "grad_norm": 0.5570671558380127,
      "learning_rate": 0.0005201723952160327,
      "loss": 3.7374,
      "step": 12450
    },
    {
      "epoch": 1.3453880099020559,
      "grad_norm": 0.4863152801990509,
      "learning_rate": 0.0005198491541859714,
      "loss": 3.7376,
      "step": 12500
    },
    {
      "epoch": 1.350769561941664,
      "grad_norm": 0.5546500086784363,
      "learning_rate": 0.0005195259131559099,
      "loss": 3.737,
      "step": 12550
    },
    {
      "epoch": 1.356151113981272,
      "grad_norm": 0.6091620922088623,
      "learning_rate": 0.0005192026721258484,
      "loss": 3.7378,
      "step": 12600
    },
    {
      "epoch": 1.3615326660208804,
      "grad_norm": 0.5455090999603271,
      "learning_rate": 0.000518879431095787,
      "loss": 3.7146,
      "step": 12650
    },
    {
      "epoch": 1.3669142180604887,
      "grad_norm": 0.5660754442214966,
      "learning_rate": 0.0005185561900657256,
      "loss": 3.7174,
      "step": 12700
    },
    {
      "epoch": 1.3722957701000968,
      "grad_norm": 0.5691196322441101,
      "learning_rate": 0.0005182329490356643,
      "loss": 3.7193,
      "step": 12750
    },
    {
      "epoch": 1.3776773221397052,
      "grad_norm": 0.5520552396774292,
      "learning_rate": 0.0005179097080056028,
      "loss": 3.7088,
      "step": 12800
    },
    {
      "epoch": 1.3830588741793133,
      "grad_norm": 0.5134276747703552,
      "learning_rate": 0.0005175864669755414,
      "loss": 3.733,
      "step": 12850
    },
    {
      "epoch": 1.3884404262189216,
      "grad_norm": 0.5690860748291016,
      "learning_rate": 0.00051726322594548,
      "loss": 3.7421,
      "step": 12900
    },
    {
      "epoch": 1.3938219782585297,
      "grad_norm": 0.5222173929214478,
      "learning_rate": 0.0005169399849154185,
      "loss": 3.7287,
      "step": 12950
    },
    {
      "epoch": 1.399203530298138,
      "grad_norm": 0.6103889346122742,
      "learning_rate": 0.0005166167438853572,
      "loss": 3.7379,
      "step": 13000
    },
    {
      "epoch": 1.399203530298138,
      "eval_accuracy": 0.3510507440495431,
      "eval_loss": 3.6815786361694336,
      "eval_runtime": 184.6188,
      "eval_samples_per_second": 97.558,
      "eval_steps_per_second": 6.099,
      "step": 13000
    },
    {
      "epoch": 1.4045850823377461,
      "grad_norm": 0.6409273147583008,
      "learning_rate": 0.0005162999676758969,
      "loss": 3.7096,
      "step": 13050
    },
    {
      "epoch": 1.4099666343773545,
      "grad_norm": 0.5928708910942078,
      "learning_rate": 0.0005159767266458355,
      "loss": 3.7176,
      "step": 13100
    },
    {
      "epoch": 1.4153481864169626,
      "grad_norm": 0.5624579787254333,
      "learning_rate": 0.0005156534856157741,
      "loss": 3.7212,
      "step": 13150
    },
    {
      "epoch": 1.420729738456571,
      "grad_norm": 0.5392971634864807,
      "learning_rate": 0.0005153302445857127,
      "loss": 3.7227,
      "step": 13200
    },
    {
      "epoch": 1.426111290496179,
      "grad_norm": 0.5955464839935303,
      "learning_rate": 0.0005150070035556513,
      "loss": 3.7138,
      "step": 13250
    },
    {
      "epoch": 1.4314928425357873,
      "grad_norm": 0.5630555748939514,
      "learning_rate": 0.0005146837625255898,
      "loss": 3.7096,
      "step": 13300
    },
    {
      "epoch": 1.4368743945753955,
      "grad_norm": 0.5762878656387329,
      "learning_rate": 0.0005143605214955285,
      "loss": 3.7116,
      "step": 13350
    },
    {
      "epoch": 1.4422559466150038,
      "grad_norm": 0.5761459469795227,
      "learning_rate": 0.0005140372804654671,
      "loss": 3.7244,
      "step": 13400
    },
    {
      "epoch": 1.447637498654612,
      "grad_norm": 0.5489415526390076,
      "learning_rate": 0.0005137140394354056,
      "loss": 3.6976,
      "step": 13450
    },
    {
      "epoch": 1.4530190506942202,
      "grad_norm": 0.5890822410583496,
      "learning_rate": 0.0005133907984053442,
      "loss": 3.72,
      "step": 13500
    },
    {
      "epoch": 1.4584006027338283,
      "grad_norm": 0.5874429941177368,
      "learning_rate": 0.0005130675573752827,
      "loss": 3.7224,
      "step": 13550
    },
    {
      "epoch": 1.4637821547734367,
      "grad_norm": 0.5852620601654053,
      "learning_rate": 0.0005127443163452214,
      "loss": 3.6857,
      "step": 13600
    },
    {
      "epoch": 1.469163706813045,
      "grad_norm": 0.5440662503242493,
      "learning_rate": 0.00051242107531516,
      "loss": 3.7188,
      "step": 13650
    },
    {
      "epoch": 1.474545258852653,
      "grad_norm": 0.5821568369865417,
      "learning_rate": 0.0005120978342850986,
      "loss": 3.6959,
      "step": 13700
    },
    {
      "epoch": 1.4799268108922612,
      "grad_norm": 0.5682273507118225,
      "learning_rate": 0.0005117745932550371,
      "loss": 3.7115,
      "step": 13750
    },
    {
      "epoch": 1.4853083629318695,
      "grad_norm": 0.5574936270713806,
      "learning_rate": 0.0005114513522249758,
      "loss": 3.6976,
      "step": 13800
    },
    {
      "epoch": 1.4906899149714778,
      "grad_norm": 0.5789446830749512,
      "learning_rate": 0.0005111281111949143,
      "loss": 3.7158,
      "step": 13850
    },
    {
      "epoch": 1.496071467011086,
      "grad_norm": 0.5653371214866638,
      "learning_rate": 0.0005108048701648528,
      "loss": 3.7058,
      "step": 13900
    },
    {
      "epoch": 1.501453019050694,
      "grad_norm": 0.6067124605178833,
      "learning_rate": 0.0005104816291347915,
      "loss": 3.7109,
      "step": 13950
    },
    {
      "epoch": 1.5068345710903024,
      "grad_norm": 0.5582537651062012,
      "learning_rate": 0.00051015838810473,
      "loss": 3.704,
      "step": 14000
    },
    {
      "epoch": 1.5068345710903024,
      "eval_accuracy": 0.3532685667956123,
      "eval_loss": 3.662456750869751,
      "eval_runtime": 184.4691,
      "eval_samples_per_second": 97.637,
      "eval_steps_per_second": 6.104,
      "step": 14000
    },
    {
      "epoch": 1.5122161231299107,
      "grad_norm": 0.5308018922805786,
      "learning_rate": 0.0005098351470746687,
      "loss": 3.7025,
      "step": 14050
    },
    {
      "epoch": 1.5175976751695188,
      "grad_norm": 0.5645959377288818,
      "learning_rate": 0.0005095119060446072,
      "loss": 3.6898,
      "step": 14100
    },
    {
      "epoch": 1.5229792272091272,
      "grad_norm": 0.5427910685539246,
      "learning_rate": 0.0005091886650145458,
      "loss": 3.7139,
      "step": 14150
    },
    {
      "epoch": 1.5283607792487355,
      "grad_norm": 0.5312597155570984,
      "learning_rate": 0.0005088654239844844,
      "loss": 3.6909,
      "step": 14200
    },
    {
      "epoch": 1.5337423312883436,
      "grad_norm": 0.5827968716621399,
      "learning_rate": 0.0005085421829544229,
      "loss": 3.684,
      "step": 14250
    },
    {
      "epoch": 1.5391238833279517,
      "grad_norm": 0.5666502714157104,
      "learning_rate": 0.0005082189419243616,
      "loss": 3.6984,
      "step": 14300
    },
    {
      "epoch": 1.54450543536756,
      "grad_norm": 0.4743129014968872,
      "learning_rate": 0.0005078957008943001,
      "loss": 3.7008,
      "step": 14350
    },
    {
      "epoch": 1.5498869874071683,
      "grad_norm": 0.5577555894851685,
      "learning_rate": 0.0005075724598642387,
      "loss": 3.6984,
      "step": 14400
    },
    {
      "epoch": 1.5552685394467765,
      "grad_norm": 0.5439442992210388,
      "learning_rate": 0.0005072492188341773,
      "loss": 3.6929,
      "step": 14450
    },
    {
      "epoch": 1.5606500914863846,
      "grad_norm": 0.6000552773475647,
      "learning_rate": 0.000506925977804116,
      "loss": 3.7026,
      "step": 14500
    },
    {
      "epoch": 1.566031643525993,
      "grad_norm": 0.5914868712425232,
      "learning_rate": 0.0005066027367740545,
      "loss": 3.7021,
      "step": 14550
    },
    {
      "epoch": 1.5714131955656012,
      "grad_norm": 0.6045658588409424,
      "learning_rate": 0.000506279495743993,
      "loss": 3.6795,
      "step": 14600
    },
    {
      "epoch": 1.5767947476052093,
      "grad_norm": 0.5576021671295166,
      "learning_rate": 0.0005059562547139316,
      "loss": 3.6943,
      "step": 14650
    },
    {
      "epoch": 1.5821762996448174,
      "grad_norm": 0.544590175151825,
      "learning_rate": 0.0005056330136838702,
      "loss": 3.6921,
      "step": 14700
    },
    {
      "epoch": 1.5875578516844258,
      "grad_norm": 0.5793708562850952,
      "learning_rate": 0.0005053097726538088,
      "loss": 3.6954,
      "step": 14750
    },
    {
      "epoch": 1.592939403724034,
      "grad_norm": 0.5727635622024536,
      "learning_rate": 0.0005049865316237474,
      "loss": 3.7,
      "step": 14800
    },
    {
      "epoch": 1.5983209557636422,
      "grad_norm": 0.5338447093963623,
      "learning_rate": 0.000504663290593686,
      "loss": 3.7025,
      "step": 14850
    },
    {
      "epoch": 1.6037025078032503,
      "grad_norm": 0.5906487107276917,
      "learning_rate": 0.0005043400495636246,
      "loss": 3.6778,
      "step": 14900
    },
    {
      "epoch": 1.6090840598428586,
      "grad_norm": 0.5547481179237366,
      "learning_rate": 0.0005040168085335632,
      "loss": 3.6889,
      "step": 14950
    },
    {
      "epoch": 1.614465611882467,
      "grad_norm": 0.5554806590080261,
      "learning_rate": 0.0005036935675035017,
      "loss": 3.6819,
      "step": 15000
    },
    {
      "epoch": 1.614465611882467,
      "eval_accuracy": 0.3546605189414036,
      "eval_loss": 3.6441338062286377,
      "eval_runtime": 184.144,
      "eval_samples_per_second": 97.809,
      "eval_steps_per_second": 6.115,
      "step": 15000
    },
    {
      "epoch": 1.619847163922075,
      "grad_norm": 0.5668731927871704,
      "learning_rate": 0.0005033703264734402,
      "loss": 3.6605,
      "step": 15050
    },
    {
      "epoch": 1.6252287159616834,
      "grad_norm": 0.5433127284049988,
      "learning_rate": 0.0005030470854433789,
      "loss": 3.6804,
      "step": 15100
    },
    {
      "epoch": 1.6306102680012917,
      "grad_norm": 0.5904312133789062,
      "learning_rate": 0.0005027238444133175,
      "loss": 3.6927,
      "step": 15150
    },
    {
      "epoch": 1.6359918200408998,
      "grad_norm": 0.57832932472229,
      "learning_rate": 0.0005024070682038573,
      "loss": 3.671,
      "step": 15200
    },
    {
      "epoch": 1.641373372080508,
      "grad_norm": 0.6239519119262695,
      "learning_rate": 0.0005020838271737959,
      "loss": 3.6763,
      "step": 15250
    },
    {
      "epoch": 1.6467549241201163,
      "grad_norm": 0.5921240448951721,
      "learning_rate": 0.0005017605861437344,
      "loss": 3.6776,
      "step": 15300
    },
    {
      "epoch": 1.6521364761597246,
      "grad_norm": 0.5674989223480225,
      "learning_rate": 0.0005014373451136731,
      "loss": 3.6907,
      "step": 15350
    },
    {
      "epoch": 1.6575180281993327,
      "grad_norm": 0.5791436433792114,
      "learning_rate": 0.0005011141040836116,
      "loss": 3.6828,
      "step": 15400
    },
    {
      "epoch": 1.6628995802389408,
      "grad_norm": 0.5726423859596252,
      "learning_rate": 0.0005007908630535503,
      "loss": 3.6801,
      "step": 15450
    },
    {
      "epoch": 1.6682811322785491,
      "grad_norm": 0.5802739858627319,
      "learning_rate": 0.0005004676220234888,
      "loss": 3.6648,
      "step": 15500
    },
    {
      "epoch": 1.6736626843181575,
      "grad_norm": 0.6185660362243652,
      "learning_rate": 0.0005001443809934273,
      "loss": 3.6886,
      "step": 15550
    },
    {
      "epoch": 1.6790442363577656,
      "grad_norm": 0.5736721158027649,
      "learning_rate": 0.000499821139963366,
      "loss": 3.6699,
      "step": 15600
    },
    {
      "epoch": 1.6844257883973737,
      "grad_norm": 0.5160577297210693,
      "learning_rate": 0.0004994978989333045,
      "loss": 3.6666,
      "step": 15650
    },
    {
      "epoch": 1.689807340436982,
      "grad_norm": 0.6582714915275574,
      "learning_rate": 0.0004991746579032431,
      "loss": 3.6816,
      "step": 15700
    },
    {
      "epoch": 1.6951888924765903,
      "grad_norm": 0.5677745938301086,
      "learning_rate": 0.0004988514168731817,
      "loss": 3.672,
      "step": 15750
    },
    {
      "epoch": 1.7005704445161984,
      "grad_norm": 0.5664964318275452,
      "learning_rate": 0.0004985281758431204,
      "loss": 3.672,
      "step": 15800
    },
    {
      "epoch": 1.7059519965558065,
      "grad_norm": 0.5303553938865662,
      "learning_rate": 0.0004982049348130589,
      "loss": 3.6698,
      "step": 15850
    },
    {
      "epoch": 1.7113335485954149,
      "grad_norm": 0.5693548321723938,
      "learning_rate": 0.0004978816937829975,
      "loss": 3.6643,
      "step": 15900
    },
    {
      "epoch": 1.7167151006350232,
      "grad_norm": 0.5552079081535339,
      "learning_rate": 0.000497558452752936,
      "loss": 3.6741,
      "step": 15950
    },
    {
      "epoch": 1.7220966526746313,
      "grad_norm": 0.544607937335968,
      "learning_rate": 0.0004972352117228746,
      "loss": 3.6824,
      "step": 16000
    },
    {
      "epoch": 1.7220966526746313,
      "eval_accuracy": 0.3563871219816635,
      "eval_loss": 3.626404285430908,
      "eval_runtime": 184.7602,
      "eval_samples_per_second": 97.483,
      "eval_steps_per_second": 6.094,
      "step": 16000
    },
    {
      "epoch": 1.7274782047142396,
      "grad_norm": 0.5834277272224426,
      "learning_rate": 0.0004969119706928133,
      "loss": 3.6684,
      "step": 16050
    },
    {
      "epoch": 1.732859756753848,
      "grad_norm": 0.5835748314857483,
      "learning_rate": 0.0004965887296627518,
      "loss": 3.6569,
      "step": 16100
    },
    {
      "epoch": 1.738241308793456,
      "grad_norm": 0.7389477491378784,
      "learning_rate": 0.0004962654886326904,
      "loss": 3.6593,
      "step": 16150
    },
    {
      "epoch": 1.7436228608330642,
      "grad_norm": 0.5737412571907043,
      "learning_rate": 0.000495942247602629,
      "loss": 3.6809,
      "step": 16200
    },
    {
      "epoch": 1.7490044128726725,
      "grad_norm": 0.5486642718315125,
      "learning_rate": 0.0004956190065725676,
      "loss": 3.6831,
      "step": 16250
    },
    {
      "epoch": 1.7543859649122808,
      "grad_norm": 0.5566287040710449,
      "learning_rate": 0.0004952957655425062,
      "loss": 3.653,
      "step": 16300
    },
    {
      "epoch": 1.759767516951889,
      "grad_norm": 0.5503498911857605,
      "learning_rate": 0.0004949725245124448,
      "loss": 3.6703,
      "step": 16350
    },
    {
      "epoch": 1.765149068991497,
      "grad_norm": 0.5988386273384094,
      "learning_rate": 0.0004946492834823833,
      "loss": 3.6629,
      "step": 16400
    },
    {
      "epoch": 1.7705306210311054,
      "grad_norm": 0.6045162677764893,
      "learning_rate": 0.0004943260424523219,
      "loss": 3.6857,
      "step": 16450
    },
    {
      "epoch": 1.7759121730707137,
      "grad_norm": 0.554808497428894,
      "learning_rate": 0.0004940028014222605,
      "loss": 3.658,
      "step": 16500
    },
    {
      "epoch": 1.7812937251103218,
      "grad_norm": 0.542362630367279,
      "learning_rate": 0.000493679560392199,
      "loss": 3.6895,
      "step": 16550
    },
    {
      "epoch": 1.78667527714993,
      "grad_norm": 0.5683366060256958,
      "learning_rate": 0.0004933563193621377,
      "loss": 3.6761,
      "step": 16600
    },
    {
      "epoch": 1.7920568291895382,
      "grad_norm": 0.5613236427307129,
      "learning_rate": 0.0004930330783320762,
      "loss": 3.6734,
      "step": 16650
    },
    {
      "epoch": 1.7974383812291466,
      "grad_norm": 0.5919580459594727,
      "learning_rate": 0.0004927098373020149,
      "loss": 3.6589,
      "step": 16700
    },
    {
      "epoch": 1.8028199332687547,
      "grad_norm": 0.5223198533058167,
      "learning_rate": 0.0004923865962719534,
      "loss": 3.6613,
      "step": 16750
    },
    {
      "epoch": 1.8082014853083628,
      "grad_norm": 0.5610350966453552,
      "learning_rate": 0.0004920633552418919,
      "loss": 3.6667,
      "step": 16800
    },
    {
      "epoch": 1.813583037347971,
      "grad_norm": 0.5463954210281372,
      "learning_rate": 0.0004917401142118306,
      "loss": 3.6762,
      "step": 16850
    },
    {
      "epoch": 1.8189645893875794,
      "grad_norm": 0.5870181322097778,
      "learning_rate": 0.0004914168731817692,
      "loss": 3.6646,
      "step": 16900
    },
    {
      "epoch": 1.8243461414271875,
      "grad_norm": 0.5730648040771484,
      "learning_rate": 0.0004910936321517078,
      "loss": 3.6575,
      "step": 16950
    },
    {
      "epoch": 1.8297276934667959,
      "grad_norm": 0.5852705836296082,
      "learning_rate": 0.0004907703911216463,
      "loss": 3.6593,
      "step": 17000
    },
    {
      "epoch": 1.8297276934667959,
      "eval_accuracy": 0.3582260721079236,
      "eval_loss": 3.6089930534362793,
      "eval_runtime": 184.3086,
      "eval_samples_per_second": 97.722,
      "eval_steps_per_second": 6.109,
      "step": 17000
    },
    {
      "epoch": 1.8351092455064042,
      "grad_norm": 0.5546532273292542,
      "learning_rate": 0.0004904471500915849,
      "loss": 3.6648,
      "step": 17050
    },
    {
      "epoch": 1.8404907975460123,
      "grad_norm": 0.5215673446655273,
      "learning_rate": 0.0004901239090615235,
      "loss": 3.6543,
      "step": 17100
    },
    {
      "epoch": 1.8458723495856204,
      "grad_norm": 0.5807424783706665,
      "learning_rate": 0.000489800668031462,
      "loss": 3.6671,
      "step": 17150
    },
    {
      "epoch": 1.8512539016252287,
      "grad_norm": 0.5335880517959595,
      "learning_rate": 0.0004894774270014007,
      "loss": 3.6385,
      "step": 17200
    },
    {
      "epoch": 1.856635453664837,
      "grad_norm": 0.6046630144119263,
      "learning_rate": 0.0004891606507919405,
      "loss": 3.6429,
      "step": 17250
    },
    {
      "epoch": 1.8620170057044452,
      "grad_norm": 0.5652769804000854,
      "learning_rate": 0.000488837409761879,
      "loss": 3.6444,
      "step": 17300
    },
    {
      "epoch": 1.8673985577440533,
      "grad_norm": 0.5628509521484375,
      "learning_rate": 0.0004885141687318177,
      "loss": 3.6182,
      "step": 17350
    },
    {
      "epoch": 1.8727801097836616,
      "grad_norm": 0.5531663298606873,
      "learning_rate": 0.00048819092770175623,
      "loss": 3.658,
      "step": 17400
    },
    {
      "epoch": 1.87816166182327,
      "grad_norm": 0.6234077215194702,
      "learning_rate": 0.0004878676866716948,
      "loss": 3.6671,
      "step": 17450
    },
    {
      "epoch": 1.883543213862878,
      "grad_norm": 0.6239759922027588,
      "learning_rate": 0.00048754444564163337,
      "loss": 3.6746,
      "step": 17500
    },
    {
      "epoch": 1.8889247659024861,
      "grad_norm": 0.5652669072151184,
      "learning_rate": 0.000487221204611572,
      "loss": 3.6429,
      "step": 17550
    },
    {
      "epoch": 1.8943063179420945,
      "grad_norm": 0.566386878490448,
      "learning_rate": 0.00048689796358151056,
      "loss": 3.6528,
      "step": 17600
    },
    {
      "epoch": 1.8996878699817028,
      "grad_norm": 0.5329195857048035,
      "learning_rate": 0.00048657472255144915,
      "loss": 3.6562,
      "step": 17650
    },
    {
      "epoch": 1.905069422021311,
      "grad_norm": 0.5578447580337524,
      "learning_rate": 0.00048625148152138775,
      "loss": 3.6657,
      "step": 17700
    },
    {
      "epoch": 1.910450974060919,
      "grad_norm": 0.5059750080108643,
      "learning_rate": 0.0004859282404913263,
      "loss": 3.6352,
      "step": 17750
    },
    {
      "epoch": 1.9158325261005273,
      "grad_norm": 0.6057028770446777,
      "learning_rate": 0.0004856049994612649,
      "loss": 3.6454,
      "step": 17800
    },
    {
      "epoch": 1.9212140781401357,
      "grad_norm": 0.5666162371635437,
      "learning_rate": 0.00048528175843120353,
      "loss": 3.652,
      "step": 17850
    },
    {
      "epoch": 1.9265956301797438,
      "grad_norm": 0.6209063529968262,
      "learning_rate": 0.0004849585174011421,
      "loss": 3.6435,
      "step": 17900
    },
    {
      "epoch": 1.931977182219352,
      "grad_norm": 0.5998473167419434,
      "learning_rate": 0.00048463527637108067,
      "loss": 3.6587,
      "step": 17950
    },
    {
      "epoch": 1.9373587342589604,
      "grad_norm": 0.5225685238838196,
      "learning_rate": 0.0004843120353410192,
      "loss": 3.6463,
      "step": 18000
    },
    {
      "epoch": 1.9373587342589604,
      "eval_accuracy": 0.359626173220301,
      "eval_loss": 3.5961034297943115,
      "eval_runtime": 184.6918,
      "eval_samples_per_second": 97.519,
      "eval_steps_per_second": 6.097,
      "step": 18000
    },
    {
      "epoch": 1.9427402862985685,
      "grad_norm": 0.626304030418396,
      "learning_rate": 0.0004839887943109578,
      "loss": 3.6488,
      "step": 18050
    },
    {
      "epoch": 1.9481218383381766,
      "grad_norm": 0.5639821887016296,
      "learning_rate": 0.00048366555328089645,
      "loss": 3.6348,
      "step": 18100
    },
    {
      "epoch": 1.953503390377785,
      "grad_norm": 0.6421195268630981,
      "learning_rate": 0.000483342312250835,
      "loss": 3.648,
      "step": 18150
    },
    {
      "epoch": 1.9588849424173933,
      "grad_norm": 0.5397465825080872,
      "learning_rate": 0.0004830190712207736,
      "loss": 3.6392,
      "step": 18200
    },
    {
      "epoch": 1.9642664944570014,
      "grad_norm": 0.5502766966819763,
      "learning_rate": 0.0004826958301907122,
      "loss": 3.6678,
      "step": 18250
    },
    {
      "epoch": 1.9696480464966095,
      "grad_norm": 0.5584690570831299,
      "learning_rate": 0.0004823725891606507,
      "loss": 3.6432,
      "step": 18300
    },
    {
      "epoch": 1.9750295985362178,
      "grad_norm": 0.5158748030662537,
      "learning_rate": 0.0004820493481305893,
      "loss": 3.635,
      "step": 18350
    },
    {
      "epoch": 1.9804111505758262,
      "grad_norm": 0.5889758467674255,
      "learning_rate": 0.00048172610710052797,
      "loss": 3.6413,
      "step": 18400
    },
    {
      "epoch": 1.9857927026154343,
      "grad_norm": 0.5550277829170227,
      "learning_rate": 0.0004814028660704665,
      "loss": 3.6224,
      "step": 18450
    },
    {
      "epoch": 1.9911742546550424,
      "grad_norm": 0.6472018361091614,
      "learning_rate": 0.0004810796250404051,
      "loss": 3.6482,
      "step": 18500
    },
    {
      "epoch": 1.9965558066946507,
      "grad_norm": 0.6527708172798157,
      "learning_rate": 0.00048075638401034364,
      "loss": 3.6119,
      "step": 18550
    },
    {
      "epoch": 2.001937358734259,
      "grad_norm": 0.5866058468818665,
      "learning_rate": 0.00048043314298028224,
      "loss": 3.5859,
      "step": 18600
    },
    {
      "epoch": 2.007318910773867,
      "grad_norm": 0.572261393070221,
      "learning_rate": 0.00048010990195022083,
      "loss": 3.5637,
      "step": 18650
    },
    {
      "epoch": 2.0127004628134753,
      "grad_norm": 0.5646705031394958,
      "learning_rate": 0.0004797866609201594,
      "loss": 3.5563,
      "step": 18700
    },
    {
      "epoch": 2.018082014853084,
      "grad_norm": 0.5501309633255005,
      "learning_rate": 0.000479463419890098,
      "loss": 3.5623,
      "step": 18750
    },
    {
      "epoch": 2.023463566892692,
      "grad_norm": 0.5593372583389282,
      "learning_rate": 0.0004791401788600366,
      "loss": 3.5727,
      "step": 18800
    },
    {
      "epoch": 2.0288451189323,
      "grad_norm": 0.6009753346443176,
      "learning_rate": 0.00047881693782997515,
      "loss": 3.5553,
      "step": 18850
    },
    {
      "epoch": 2.034226670971908,
      "grad_norm": 0.7166525721549988,
      "learning_rate": 0.00047849369679991375,
      "loss": 3.5668,
      "step": 18900
    },
    {
      "epoch": 2.0396082230115167,
      "grad_norm": 0.5624678134918213,
      "learning_rate": 0.0004781704557698523,
      "loss": 3.5726,
      "step": 18950
    },
    {
      "epoch": 2.044989775051125,
      "grad_norm": 0.5785338282585144,
      "learning_rate": 0.00047784721473979094,
      "loss": 3.5643,
      "step": 19000
    },
    {
      "epoch": 2.044989775051125,
      "eval_accuracy": 0.36080744741662,
      "eval_loss": 3.5861058235168457,
      "eval_runtime": 184.2445,
      "eval_samples_per_second": 97.756,
      "eval_steps_per_second": 6.111,
      "step": 19000
    },
    {
      "epoch": 2.050371327090733,
      "grad_norm": 0.5863303542137146,
      "learning_rate": 0.00047752397370972953,
      "loss": 3.5615,
      "step": 19050
    },
    {
      "epoch": 2.055752879130341,
      "grad_norm": 0.516839861869812,
      "learning_rate": 0.0004772007326796681,
      "loss": 3.5527,
      "step": 19100
    },
    {
      "epoch": 2.0611344311699495,
      "grad_norm": 0.5679395198822021,
      "learning_rate": 0.00047687749164960667,
      "loss": 3.5529,
      "step": 19150
    },
    {
      "epoch": 2.0665159832095576,
      "grad_norm": 0.633868932723999,
      "learning_rate": 0.00047655425061954526,
      "loss": 3.5628,
      "step": 19200
    },
    {
      "epoch": 2.0718975352491658,
      "grad_norm": 0.6057083606719971,
      "learning_rate": 0.00047623100958948386,
      "loss": 3.5349,
      "step": 19250
    },
    {
      "epoch": 2.0772790872887743,
      "grad_norm": 0.5921409726142883,
      "learning_rate": 0.00047590776855942245,
      "loss": 3.5432,
      "step": 19300
    },
    {
      "epoch": 2.0826606393283824,
      "grad_norm": 0.594826340675354,
      "learning_rate": 0.00047558452752936105,
      "loss": 3.5675,
      "step": 19350
    },
    {
      "epoch": 2.0880421913679905,
      "grad_norm": 0.563567042350769,
      "learning_rate": 0.0004752612864992996,
      "loss": 3.5716,
      "step": 19400
    },
    {
      "epoch": 2.0934237434075986,
      "grad_norm": 0.6127414107322693,
      "learning_rate": 0.0004749380454692382,
      "loss": 3.5576,
      "step": 19450
    },
    {
      "epoch": 2.098805295447207,
      "grad_norm": 0.5405972599983215,
      "learning_rate": 0.0004746148044391767,
      "loss": 3.5377,
      "step": 19500
    },
    {
      "epoch": 2.1041868474868153,
      "grad_norm": 0.5919598937034607,
      "learning_rate": 0.00047429156340911537,
      "loss": 3.559,
      "step": 19550
    },
    {
      "epoch": 2.1095683995264234,
      "grad_norm": 0.6029935479164124,
      "learning_rate": 0.0004739747871996551,
      "loss": 3.5603,
      "step": 19600
    },
    {
      "epoch": 2.1149499515660315,
      "grad_norm": 0.5658677220344543,
      "learning_rate": 0.00047365154616959377,
      "loss": 3.5638,
      "step": 19650
    },
    {
      "epoch": 2.12033150360564,
      "grad_norm": 0.5634831786155701,
      "learning_rate": 0.00047332830513953237,
      "loss": 3.5569,
      "step": 19700
    },
    {
      "epoch": 2.125713055645248,
      "grad_norm": 0.5616461634635925,
      "learning_rate": 0.0004730050641094709,
      "loss": 3.5731,
      "step": 19750
    },
    {
      "epoch": 2.1310946076848563,
      "grad_norm": 0.6075220108032227,
      "learning_rate": 0.0004726818230794095,
      "loss": 3.56,
      "step": 19800
    },
    {
      "epoch": 2.1364761597244644,
      "grad_norm": 0.6314546465873718,
      "learning_rate": 0.00047235858204934804,
      "loss": 3.5832,
      "step": 19850
    },
    {
      "epoch": 2.141857711764073,
      "grad_norm": 0.5252575278282166,
      "learning_rate": 0.0004720353410192867,
      "loss": 3.5728,
      "step": 19900
    },
    {
      "epoch": 2.147239263803681,
      "grad_norm": 0.5813453197479248,
      "learning_rate": 0.0004717120999892253,
      "loss": 3.5722,
      "step": 19950
    },
    {
      "epoch": 2.152620815843289,
      "grad_norm": 0.5590271949768066,
      "learning_rate": 0.0004713888589591638,
      "loss": 3.5555,
      "step": 20000
    },
    {
      "epoch": 2.152620815843289,
      "eval_accuracy": 0.3620278366525523,
      "eval_loss": 3.5756726264953613,
      "eval_runtime": 184.4292,
      "eval_samples_per_second": 97.658,
      "eval_steps_per_second": 6.105,
      "step": 20000
    },
    {
      "epoch": 2.1580023678828972,
      "grad_norm": 0.569237470626831,
      "learning_rate": 0.0004710656179291024,
      "loss": 3.5626,
      "step": 20050
    },
    {
      "epoch": 2.163383919922506,
      "grad_norm": 0.6128169894218445,
      "learning_rate": 0.000470742376899041,
      "loss": 3.5588,
      "step": 20100
    },
    {
      "epoch": 2.168765471962114,
      "grad_norm": 0.5507939457893372,
      "learning_rate": 0.00047041913586897956,
      "loss": 3.5712,
      "step": 20150
    },
    {
      "epoch": 2.174147024001722,
      "grad_norm": 0.5997487306594849,
      "learning_rate": 0.0004700958948389182,
      "loss": 3.5793,
      "step": 20200
    },
    {
      "epoch": 2.1795285760413305,
      "grad_norm": 0.6164665222167969,
      "learning_rate": 0.0004697726538088568,
      "loss": 3.5544,
      "step": 20250
    },
    {
      "epoch": 2.1849101280809387,
      "grad_norm": 0.6177306771278381,
      "learning_rate": 0.00046944941277879534,
      "loss": 3.5816,
      "step": 20300
    },
    {
      "epoch": 2.1902916801205468,
      "grad_norm": 0.5694639086723328,
      "learning_rate": 0.00046912617174873394,
      "loss": 3.5634,
      "step": 20350
    },
    {
      "epoch": 2.195673232160155,
      "grad_norm": 0.5714508295059204,
      "learning_rate": 0.0004688029307186725,
      "loss": 3.5552,
      "step": 20400
    },
    {
      "epoch": 2.2010547841997634,
      "grad_norm": 0.5863978266716003,
      "learning_rate": 0.00046847968968861107,
      "loss": 3.5663,
      "step": 20450
    },
    {
      "epoch": 2.2064363362393715,
      "grad_norm": 0.567820131778717,
      "learning_rate": 0.0004681564486585497,
      "loss": 3.5664,
      "step": 20500
    },
    {
      "epoch": 2.2118178882789796,
      "grad_norm": 0.599705159664154,
      "learning_rate": 0.00046783320762848826,
      "loss": 3.5713,
      "step": 20550
    },
    {
      "epoch": 2.2171994403185877,
      "grad_norm": 0.5928988456726074,
      "learning_rate": 0.00046750996659842685,
      "loss": 3.5539,
      "step": 20600
    },
    {
      "epoch": 2.2225809923581963,
      "grad_norm": 0.6215653419494629,
      "learning_rate": 0.00046718672556836545,
      "loss": 3.5643,
      "step": 20650
    },
    {
      "epoch": 2.2279625443978044,
      "grad_norm": 1.1149789094924927,
      "learning_rate": 0.000466863484538304,
      "loss": 3.5393,
      "step": 20700
    },
    {
      "epoch": 2.2333440964374125,
      "grad_norm": 0.5954315066337585,
      "learning_rate": 0.0004665402435082426,
      "loss": 3.5733,
      "step": 20750
    },
    {
      "epoch": 2.2387256484770206,
      "grad_norm": 0.5726349949836731,
      "learning_rate": 0.00046621700247818123,
      "loss": 3.5703,
      "step": 20800
    },
    {
      "epoch": 2.244107200516629,
      "grad_norm": 0.5380665063858032,
      "learning_rate": 0.0004658937614481198,
      "loss": 3.5658,
      "step": 20850
    },
    {
      "epoch": 2.2494887525562373,
      "grad_norm": 0.5875298976898193,
      "learning_rate": 0.00046557052041805837,
      "loss": 3.5581,
      "step": 20900
    },
    {
      "epoch": 2.2548703045958454,
      "grad_norm": 0.6029031872749329,
      "learning_rate": 0.0004652472793879969,
      "loss": 3.5589,
      "step": 20950
    },
    {
      "epoch": 2.2602518566354535,
      "grad_norm": 0.5990302562713623,
      "learning_rate": 0.0004649240383579355,
      "loss": 3.5354,
      "step": 21000
    },
    {
      "epoch": 2.2602518566354535,
      "eval_accuracy": 0.3630543891366235,
      "eval_loss": 3.5674469470977783,
      "eval_runtime": 184.3873,
      "eval_samples_per_second": 97.68,
      "eval_steps_per_second": 6.107,
      "step": 21000
    },
    {
      "epoch": 2.265633408675062,
      "grad_norm": 0.5953611731529236,
      "learning_rate": 0.00046460079732787415,
      "loss": 3.5803,
      "step": 21050
    },
    {
      "epoch": 2.27101496071467,
      "grad_norm": 0.6006433963775635,
      "learning_rate": 0.0004642775562978127,
      "loss": 3.5661,
      "step": 21100
    },
    {
      "epoch": 2.2763965127542782,
      "grad_norm": 0.5515443086624146,
      "learning_rate": 0.0004639543152677513,
      "loss": 3.5575,
      "step": 21150
    },
    {
      "epoch": 2.281778064793887,
      "grad_norm": 0.6373427510261536,
      "learning_rate": 0.0004636310742376899,
      "loss": 3.5672,
      "step": 21200
    },
    {
      "epoch": 2.287159616833495,
      "grad_norm": 0.6004770994186401,
      "learning_rate": 0.0004633078332076284,
      "loss": 3.5786,
      "step": 21250
    },
    {
      "epoch": 2.292541168873103,
      "grad_norm": 0.5543513894081116,
      "learning_rate": 0.000462984592177567,
      "loss": 3.5707,
      "step": 21300
    },
    {
      "epoch": 2.297922720912711,
      "grad_norm": 0.5607895255088806,
      "learning_rate": 0.00046266135114750567,
      "loss": 3.5585,
      "step": 21350
    },
    {
      "epoch": 2.303304272952319,
      "grad_norm": 0.5842389464378357,
      "learning_rate": 0.0004623381101174442,
      "loss": 3.568,
      "step": 21400
    },
    {
      "epoch": 2.3086858249919278,
      "grad_norm": 0.5327559113502502,
      "learning_rate": 0.0004620148690873828,
      "loss": 3.5628,
      "step": 21450
    },
    {
      "epoch": 2.314067377031536,
      "grad_norm": 0.6334641575813293,
      "learning_rate": 0.00046169162805732134,
      "loss": 3.5579,
      "step": 21500
    },
    {
      "epoch": 2.319448929071144,
      "grad_norm": 0.6180813908576965,
      "learning_rate": 0.00046136838702725994,
      "loss": 3.5685,
      "step": 21550
    },
    {
      "epoch": 2.3248304811107525,
      "grad_norm": 0.5589969754219055,
      "learning_rate": 0.00046105161081779974,
      "loss": 3.5638,
      "step": 21600
    },
    {
      "epoch": 2.3302120331503606,
      "grad_norm": 0.6755582094192505,
      "learning_rate": 0.00046072836978773834,
      "loss": 3.5614,
      "step": 21650
    },
    {
      "epoch": 2.3355935851899687,
      "grad_norm": 0.5525728464126587,
      "learning_rate": 0.000460405128757677,
      "loss": 3.551,
      "step": 21700
    },
    {
      "epoch": 2.340975137229577,
      "grad_norm": 0.5571237206459045,
      "learning_rate": 0.0004600818877276155,
      "loss": 3.5775,
      "step": 21750
    },
    {
      "epoch": 2.3463566892691854,
      "grad_norm": 0.5496062636375427,
      "learning_rate": 0.0004597651115181554,
      "loss": 3.5669,
      "step": 21800
    },
    {
      "epoch": 2.3517382413087935,
      "grad_norm": 0.5897806286811829,
      "learning_rate": 0.00045944187048809393,
      "loss": 3.5748,
      "step": 21850
    },
    {
      "epoch": 2.3571197933484016,
      "grad_norm": 0.576259970664978,
      "learning_rate": 0.0004591186294580325,
      "loss": 3.5555,
      "step": 21900
    },
    {
      "epoch": 2.3625013453880097,
      "grad_norm": 0.6202755570411682,
      "learning_rate": 0.00045879538842797106,
      "loss": 3.5582,
      "step": 21950
    },
    {
      "epoch": 2.3678828974276183,
      "grad_norm": 0.575225293636322,
      "learning_rate": 0.00045847214739790966,
      "loss": 3.5567,
      "step": 22000
    },
    {
      "epoch": 2.3678828974276183,
      "eval_accuracy": 0.3642527218363294,
      "eval_loss": 3.5551345348358154,
      "eval_runtime": 184.3868,
      "eval_samples_per_second": 97.681,
      "eval_steps_per_second": 6.107,
      "step": 22000
    },
    {
      "epoch": 2.3732644494672264,
      "grad_norm": 0.6205533146858215,
      "learning_rate": 0.0004581489063678482,
      "loss": 3.5536,
      "step": 22050
    },
    {
      "epoch": 2.3786460015068345,
      "grad_norm": 0.6232433319091797,
      "learning_rate": 0.00045782566533778685,
      "loss": 3.5529,
      "step": 22100
    },
    {
      "epoch": 2.384027553546443,
      "grad_norm": 0.5750186443328857,
      "learning_rate": 0.00045750242430772544,
      "loss": 3.5585,
      "step": 22150
    },
    {
      "epoch": 2.389409105586051,
      "grad_norm": 0.6692588329315186,
      "learning_rate": 0.000457179183277664,
      "loss": 3.5555,
      "step": 22200
    },
    {
      "epoch": 2.3947906576256592,
      "grad_norm": 0.6092421412467957,
      "learning_rate": 0.0004568559422476026,
      "loss": 3.5597,
      "step": 22250
    },
    {
      "epoch": 2.4001722096652673,
      "grad_norm": 0.5946943759918213,
      "learning_rate": 0.00045653270121754117,
      "loss": 3.5556,
      "step": 22300
    },
    {
      "epoch": 2.4055537617048754,
      "grad_norm": 0.6357599496841431,
      "learning_rate": 0.0004562094601874797,
      "loss": 3.5606,
      "step": 22350
    },
    {
      "epoch": 2.410935313744484,
      "grad_norm": 0.6717286705970764,
      "learning_rate": 0.00045588621915741836,
      "loss": 3.5604,
      "step": 22400
    },
    {
      "epoch": 2.416316865784092,
      "grad_norm": 0.6248000264167786,
      "learning_rate": 0.00045556297812735696,
      "loss": 3.5832,
      "step": 22450
    },
    {
      "epoch": 2.4216984178237,
      "grad_norm": 0.6078273057937622,
      "learning_rate": 0.0004552397370972955,
      "loss": 3.5498,
      "step": 22500
    },
    {
      "epoch": 2.4270799698633088,
      "grad_norm": 0.660775363445282,
      "learning_rate": 0.0004549164960672341,
      "loss": 3.554,
      "step": 22550
    },
    {
      "epoch": 2.432461521902917,
      "grad_norm": 0.6106600165367126,
      "learning_rate": 0.00045459325503717263,
      "loss": 3.5698,
      "step": 22600
    },
    {
      "epoch": 2.437843073942525,
      "grad_norm": 0.6340673565864563,
      "learning_rate": 0.0004542700140071113,
      "loss": 3.55,
      "step": 22650
    },
    {
      "epoch": 2.443224625982133,
      "grad_norm": 0.5620933175086975,
      "learning_rate": 0.0004539467729770499,
      "loss": 3.5538,
      "step": 22700
    },
    {
      "epoch": 2.4486061780217416,
      "grad_norm": 0.6145320534706116,
      "learning_rate": 0.0004536235319469884,
      "loss": 3.5755,
      "step": 22750
    },
    {
      "epoch": 2.4539877300613497,
      "grad_norm": 0.5466718077659607,
      "learning_rate": 0.000453300290916927,
      "loss": 3.5512,
      "step": 22800
    },
    {
      "epoch": 2.459369282100958,
      "grad_norm": 0.5778665542602539,
      "learning_rate": 0.0004529770498868656,
      "loss": 3.574,
      "step": 22850
    },
    {
      "epoch": 2.464750834140566,
      "grad_norm": 0.5278246998786926,
      "learning_rate": 0.00045265380885680414,
      "loss": 3.5621,
      "step": 22900
    },
    {
      "epoch": 2.4701323861801745,
      "grad_norm": 0.6504663228988647,
      "learning_rate": 0.0004523305678267428,
      "loss": 3.545,
      "step": 22950
    },
    {
      "epoch": 2.4755139382197826,
      "grad_norm": 0.6087246537208557,
      "learning_rate": 0.0004520073267966814,
      "loss": 3.5591,
      "step": 23000
    },
    {
      "epoch": 2.4755139382197826,
      "eval_accuracy": 0.365130854475646,
      "eval_loss": 3.546229600906372,
      "eval_runtime": 184.3114,
      "eval_samples_per_second": 97.72,
      "eval_steps_per_second": 6.109,
      "step": 23000
    },
    {
      "epoch": 2.4808954902593907,
      "grad_norm": 0.5887711644172668,
      "learning_rate": 0.00045168408576661993,
      "loss": 3.5464,
      "step": 23050
    },
    {
      "epoch": 2.4862770422989993,
      "grad_norm": 0.6163349747657776,
      "learning_rate": 0.0004513608447365585,
      "loss": 3.5743,
      "step": 23100
    },
    {
      "epoch": 2.4916585943386074,
      "grad_norm": 0.6012129187583923,
      "learning_rate": 0.00045103760370649706,
      "loss": 3.5533,
      "step": 23150
    },
    {
      "epoch": 2.4970401463782155,
      "grad_norm": 0.5883326530456543,
      "learning_rate": 0.00045071436267643566,
      "loss": 3.5591,
      "step": 23200
    },
    {
      "epoch": 2.5024216984178236,
      "grad_norm": 0.5886674523353577,
      "learning_rate": 0.0004503911216463743,
      "loss": 3.5569,
      "step": 23250
    },
    {
      "epoch": 2.5078032504574317,
      "grad_norm": 0.5826066136360168,
      "learning_rate": 0.00045006788061631285,
      "loss": 3.5646,
      "step": 23300
    },
    {
      "epoch": 2.5131848024970402,
      "grad_norm": 0.5926262140274048,
      "learning_rate": 0.00044974463958625144,
      "loss": 3.5606,
      "step": 23350
    },
    {
      "epoch": 2.5185663545366483,
      "grad_norm": 0.580472469329834,
      "learning_rate": 0.00044942139855619004,
      "loss": 3.577,
      "step": 23400
    },
    {
      "epoch": 2.5239479065762565,
      "grad_norm": 0.6202252507209778,
      "learning_rate": 0.0004490981575261286,
      "loss": 3.5596,
      "step": 23450
    },
    {
      "epoch": 2.529329458615865,
      "grad_norm": 0.5866471529006958,
      "learning_rate": 0.0004487749164960672,
      "loss": 3.5524,
      "step": 23500
    },
    {
      "epoch": 2.534711010655473,
      "grad_norm": 0.6061303019523621,
      "learning_rate": 0.0004484516754660058,
      "loss": 3.5537,
      "step": 23550
    },
    {
      "epoch": 2.540092562695081,
      "grad_norm": 0.5850672125816345,
      "learning_rate": 0.00044812843443594436,
      "loss": 3.5553,
      "step": 23600
    },
    {
      "epoch": 2.5454741147346893,
      "grad_norm": 0.6245715618133545,
      "learning_rate": 0.00044780519340588296,
      "loss": 3.5746,
      "step": 23650
    },
    {
      "epoch": 2.550855666774298,
      "grad_norm": 0.5901142358779907,
      "learning_rate": 0.0004474819523758215,
      "loss": 3.5462,
      "step": 23700
    },
    {
      "epoch": 2.556237218813906,
      "grad_norm": 0.6326914429664612,
      "learning_rate": 0.0004471587113457601,
      "loss": 3.5691,
      "step": 23750
    },
    {
      "epoch": 2.561618770853514,
      "grad_norm": 0.5844827890396118,
      "learning_rate": 0.00044683547031569874,
      "loss": 3.555,
      "step": 23800
    },
    {
      "epoch": 2.567000322893122,
      "grad_norm": 0.588846743106842,
      "learning_rate": 0.0004465122292856373,
      "loss": 3.5626,
      "step": 23850
    },
    {
      "epoch": 2.5723818749327307,
      "grad_norm": 0.5835680961608887,
      "learning_rate": 0.0004461889882555759,
      "loss": 3.5559,
      "step": 23900
    },
    {
      "epoch": 2.577763426972339,
      "grad_norm": 0.6164372563362122,
      "learning_rate": 0.00044586574722551447,
      "loss": 3.5453,
      "step": 23950
    },
    {
      "epoch": 2.583144979011947,
      "grad_norm": 0.5695710182189941,
      "learning_rate": 0.000445542506195453,
      "loss": 3.5687,
      "step": 24000
    },
    {
      "epoch": 2.583144979011947,
      "eval_accuracy": 0.36589109873168396,
      "eval_loss": 3.536973237991333,
      "eval_runtime": 184.4961,
      "eval_samples_per_second": 97.623,
      "eval_steps_per_second": 6.103,
      "step": 24000
    },
    {
      "epoch": 2.5885265310515555,
      "grad_norm": 0.6265692114830017,
      "learning_rate": 0.0004452192651653916,
      "loss": 3.5648,
      "step": 24050
    },
    {
      "epoch": 2.5939080830911636,
      "grad_norm": 0.5727704167366028,
      "learning_rate": 0.00044489602413533025,
      "loss": 3.5513,
      "step": 24100
    },
    {
      "epoch": 2.5992896351307717,
      "grad_norm": 0.5817157030105591,
      "learning_rate": 0.0004445727831052688,
      "loss": 3.5297,
      "step": 24150
    },
    {
      "epoch": 2.60467118717038,
      "grad_norm": 0.6151541471481323,
      "learning_rate": 0.0004442495420752074,
      "loss": 3.5545,
      "step": 24200
    },
    {
      "epoch": 2.610052739209988,
      "grad_norm": 0.6546781063079834,
      "learning_rate": 0.00044392630104514593,
      "loss": 3.546,
      "step": 24250
    },
    {
      "epoch": 2.6154342912495965,
      "grad_norm": 0.5718879699707031,
      "learning_rate": 0.0004436030600150845,
      "loss": 3.5443,
      "step": 24300
    },
    {
      "epoch": 2.6208158432892046,
      "grad_norm": 0.5507016181945801,
      "learning_rate": 0.0004432798189850231,
      "loss": 3.5538,
      "step": 24350
    },
    {
      "epoch": 2.6261973953288127,
      "grad_norm": 0.5672619342803955,
      "learning_rate": 0.0004429565779549617,
      "loss": 3.5482,
      "step": 24400
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 0.5896308422088623,
      "learning_rate": 0.0004426333369249003,
      "loss": 3.5468,
      "step": 24450
    },
    {
      "epoch": 2.6369604994080293,
      "grad_norm": 0.6065009832382202,
      "learning_rate": 0.0004423100958948389,
      "loss": 3.5559,
      "step": 24500
    },
    {
      "epoch": 2.6423420514476375,
      "grad_norm": 0.580746591091156,
      "learning_rate": 0.00044198685486477744,
      "loss": 3.565,
      "step": 24550
    },
    {
      "epoch": 2.6477236034872456,
      "grad_norm": 0.5891657471656799,
      "learning_rate": 0.00044166361383471604,
      "loss": 3.5625,
      "step": 24600
    },
    {
      "epoch": 2.653105155526854,
      "grad_norm": 0.5834586024284363,
      "learning_rate": 0.0004413403728046547,
      "loss": 3.567,
      "step": 24650
    },
    {
      "epoch": 2.658486707566462,
      "grad_norm": 0.6075326800346375,
      "learning_rate": 0.0004410171317745932,
      "loss": 3.5495,
      "step": 24700
    },
    {
      "epoch": 2.6638682596060703,
      "grad_norm": 0.5860100984573364,
      "learning_rate": 0.0004406938907445318,
      "loss": 3.5476,
      "step": 24750
    },
    {
      "epoch": 2.6692498116456784,
      "grad_norm": 0.5915664434432983,
      "learning_rate": 0.00044037064971447036,
      "loss": 3.5393,
      "step": 24800
    },
    {
      "epoch": 2.674631363685287,
      "grad_norm": 0.6533091068267822,
      "learning_rate": 0.00044004740868440896,
      "loss": 3.5503,
      "step": 24850
    },
    {
      "epoch": 2.680012915724895,
      "grad_norm": 0.5966440439224243,
      "learning_rate": 0.00043972416765434755,
      "loss": 3.5489,
      "step": 24900
    },
    {
      "epoch": 2.685394467764503,
      "grad_norm": 0.543996274471283,
      "learning_rate": 0.00043940092662428615,
      "loss": 3.5434,
      "step": 24950
    },
    {
      "epoch": 2.6907760198041117,
      "grad_norm": 0.5796987414360046,
      "learning_rate": 0.00043907768559422474,
      "loss": 3.5648,
      "step": 25000
    },
    {
      "epoch": 2.6907760198041117,
      "eval_accuracy": 0.3670347790288192,
      "eval_loss": 3.5273029804229736,
      "eval_runtime": 184.4572,
      "eval_samples_per_second": 97.643,
      "eval_steps_per_second": 6.104,
      "step": 25000
    },
    {
      "epoch": 2.69615757184372,
      "grad_norm": 0.598313570022583,
      "learning_rate": 0.00043875444456416334,
      "loss": 3.5445,
      "step": 25050
    },
    {
      "epoch": 2.701539123883328,
      "grad_norm": 0.5599380135536194,
      "learning_rate": 0.0004384312035341019,
      "loss": 3.5487,
      "step": 25100
    },
    {
      "epoch": 2.706920675922936,
      "grad_norm": 0.6336055397987366,
      "learning_rate": 0.0004381144273246417,
      "loss": 3.5344,
      "step": 25150
    },
    {
      "epoch": 2.712302227962544,
      "grad_norm": 0.6747413277626038,
      "learning_rate": 0.0004377911862945803,
      "loss": 3.5519,
      "step": 25200
    },
    {
      "epoch": 2.7176837800021527,
      "grad_norm": 0.5955970287322998,
      "learning_rate": 0.00043746794526451887,
      "loss": 3.5387,
      "step": 25250
    },
    {
      "epoch": 2.723065332041761,
      "grad_norm": 0.6370869874954224,
      "learning_rate": 0.00043714470423445747,
      "loss": 3.5252,
      "step": 25300
    },
    {
      "epoch": 2.728446884081369,
      "grad_norm": 0.5500616431236267,
      "learning_rate": 0.00043682146320439606,
      "loss": 3.541,
      "step": 25350
    },
    {
      "epoch": 2.7338284361209775,
      "grad_norm": 0.5476239919662476,
      "learning_rate": 0.00043649822217433466,
      "loss": 3.5346,
      "step": 25400
    },
    {
      "epoch": 2.7392099881605856,
      "grad_norm": 0.6034436225891113,
      "learning_rate": 0.0004361749811442732,
      "loss": 3.5598,
      "step": 25450
    },
    {
      "epoch": 2.7445915402001937,
      "grad_norm": 0.539374589920044,
      "learning_rate": 0.0004358517401142118,
      "loss": 3.5383,
      "step": 25500
    },
    {
      "epoch": 2.749973092239802,
      "grad_norm": 0.5982243418693542,
      "learning_rate": 0.00043552849908415033,
      "loss": 3.5526,
      "step": 25550
    },
    {
      "epoch": 2.7553546442794103,
      "grad_norm": 0.6408721804618835,
      "learning_rate": 0.000435205258054089,
      "loss": 3.5413,
      "step": 25600
    },
    {
      "epoch": 2.7607361963190185,
      "grad_norm": 0.5847112536430359,
      "learning_rate": 0.0004348820170240276,
      "loss": 3.524,
      "step": 25650
    },
    {
      "epoch": 2.7661177483586266,
      "grad_norm": 0.6375226378440857,
      "learning_rate": 0.0004345587759939661,
      "loss": 3.5437,
      "step": 25700
    },
    {
      "epoch": 2.7714993003982347,
      "grad_norm": 0.5986684560775757,
      "learning_rate": 0.0004342355349639047,
      "loss": 3.5426,
      "step": 25750
    },
    {
      "epoch": 2.776880852437843,
      "grad_norm": 0.6059084534645081,
      "learning_rate": 0.0004339122939338433,
      "loss": 3.567,
      "step": 25800
    },
    {
      "epoch": 2.7822624044774513,
      "grad_norm": 0.5697919130325317,
      "learning_rate": 0.00043358905290378184,
      "loss": 3.5538,
      "step": 25850
    },
    {
      "epoch": 2.7876439565170594,
      "grad_norm": 0.6090335249900818,
      "learning_rate": 0.0004332658118737205,
      "loss": 3.5257,
      "step": 25900
    },
    {
      "epoch": 2.793025508556668,
      "grad_norm": 0.5608583688735962,
      "learning_rate": 0.0004329425708436591,
      "loss": 3.5249,
      "step": 25950
    },
    {
      "epoch": 2.798407060596276,
      "grad_norm": 0.6095306277275085,
      "learning_rate": 0.00043261932981359763,
      "loss": 3.5415,
      "step": 26000
    },
    {
      "epoch": 2.798407060596276,
      "eval_accuracy": 0.36779230696266174,
      "eval_loss": 3.518670082092285,
      "eval_runtime": 184.5683,
      "eval_samples_per_second": 97.584,
      "eval_steps_per_second": 6.101,
      "step": 26000
    },
    {
      "epoch": 2.803788612635884,
      "grad_norm": 0.6020979285240173,
      "learning_rate": 0.0004322960887835362,
      "loss": 3.5443,
      "step": 26050
    },
    {
      "epoch": 2.8091701646754923,
      "grad_norm": 0.6217984557151794,
      "learning_rate": 0.00043197284775347476,
      "loss": 3.5535,
      "step": 26100
    },
    {
      "epoch": 2.8145517167151004,
      "grad_norm": 0.5924156904220581,
      "learning_rate": 0.00043164960672341336,
      "loss": 3.5478,
      "step": 26150
    },
    {
      "epoch": 2.819933268754709,
      "grad_norm": 0.596008837223053,
      "learning_rate": 0.000431326365693352,
      "loss": 3.5479,
      "step": 26200
    },
    {
      "epoch": 2.825314820794317,
      "grad_norm": 0.602006196975708,
      "learning_rate": 0.00043100312466329055,
      "loss": 3.5386,
      "step": 26250
    },
    {
      "epoch": 2.830696372833925,
      "grad_norm": 0.6307808756828308,
      "learning_rate": 0.00043067988363322914,
      "loss": 3.5312,
      "step": 26300
    },
    {
      "epoch": 2.8360779248735337,
      "grad_norm": 0.5755499601364136,
      "learning_rate": 0.00043035664260316774,
      "loss": 3.551,
      "step": 26350
    },
    {
      "epoch": 2.841459476913142,
      "grad_norm": 0.6352705359458923,
      "learning_rate": 0.0004300334015731063,
      "loss": 3.5311,
      "step": 26400
    },
    {
      "epoch": 2.84684102895275,
      "grad_norm": 0.5776434540748596,
      "learning_rate": 0.0004297101605430449,
      "loss": 3.5378,
      "step": 26450
    },
    {
      "epoch": 2.852222580992358,
      "grad_norm": 0.5926163792610168,
      "learning_rate": 0.0004293869195129835,
      "loss": 3.5311,
      "step": 26500
    },
    {
      "epoch": 2.857604133031966,
      "grad_norm": 0.6265907287597656,
      "learning_rate": 0.00042906367848292206,
      "loss": 3.5451,
      "step": 26550
    },
    {
      "epoch": 2.8629856850715747,
      "grad_norm": 0.6043955087661743,
      "learning_rate": 0.00042874043745286066,
      "loss": 3.5335,
      "step": 26600
    },
    {
      "epoch": 2.868367237111183,
      "grad_norm": 0.5979616641998291,
      "learning_rate": 0.0004284171964227992,
      "loss": 3.5438,
      "step": 26650
    },
    {
      "epoch": 2.873748789150791,
      "grad_norm": 0.6957486867904663,
      "learning_rate": 0.0004280939553927378,
      "loss": 3.5178,
      "step": 26700
    },
    {
      "epoch": 2.8791303411903995,
      "grad_norm": 0.5793423056602478,
      "learning_rate": 0.00042777071436267644,
      "loss": 3.5425,
      "step": 26750
    },
    {
      "epoch": 2.8845118932300076,
      "grad_norm": 0.5961024761199951,
      "learning_rate": 0.000427447473332615,
      "loss": 3.5399,
      "step": 26800
    },
    {
      "epoch": 2.8898934452696157,
      "grad_norm": 0.5859310626983643,
      "learning_rate": 0.0004271242323025536,
      "loss": 3.5324,
      "step": 26850
    },
    {
      "epoch": 2.895274997309224,
      "grad_norm": 0.6074236035346985,
      "learning_rate": 0.00042680099127249217,
      "loss": 3.541,
      "step": 26900
    },
    {
      "epoch": 2.9006565493488323,
      "grad_norm": 0.5824597477912903,
      "learning_rate": 0.0004264777502424307,
      "loss": 3.5303,
      "step": 26950
    },
    {
      "epoch": 2.9060381013884404,
      "grad_norm": 0.6311275362968445,
      "learning_rate": 0.0004261545092123693,
      "loss": 3.5276,
      "step": 27000
    },
    {
      "epoch": 2.9060381013884404,
      "eval_accuracy": 0.36884406691670596,
      "eval_loss": 3.5137550830841064,
      "eval_runtime": 184.352,
      "eval_samples_per_second": 97.699,
      "eval_steps_per_second": 6.108,
      "step": 27000
    },
    {
      "epoch": 2.9114196534280485,
      "grad_norm": 0.5767911672592163,
      "learning_rate": 0.00042583126818230795,
      "loss": 3.5308,
      "step": 27050
    },
    {
      "epoch": 2.9168012054676566,
      "grad_norm": 0.7029990553855896,
      "learning_rate": 0.0004255080271522465,
      "loss": 3.5432,
      "step": 27100
    },
    {
      "epoch": 2.922182757507265,
      "grad_norm": 0.5921109318733215,
      "learning_rate": 0.0004251847861221851,
      "loss": 3.5301,
      "step": 27150
    },
    {
      "epoch": 2.9275643095468733,
      "grad_norm": 0.6075803637504578,
      "learning_rate": 0.00042486154509212363,
      "loss": 3.5377,
      "step": 27200
    },
    {
      "epoch": 2.9329458615864814,
      "grad_norm": 0.6090728044509888,
      "learning_rate": 0.0004245383040620622,
      "loss": 3.5595,
      "step": 27250
    },
    {
      "epoch": 2.93832741362609,
      "grad_norm": 0.6665115356445312,
      "learning_rate": 0.0004242150630320009,
      "loss": 3.5455,
      "step": 27300
    },
    {
      "epoch": 2.943708965665698,
      "grad_norm": 0.5654910802841187,
      "learning_rate": 0.0004238918220019394,
      "loss": 3.5366,
      "step": 27350
    },
    {
      "epoch": 2.949090517705306,
      "grad_norm": 0.5783439874649048,
      "learning_rate": 0.000423568580971878,
      "loss": 3.5253,
      "step": 27400
    },
    {
      "epoch": 2.9544720697449143,
      "grad_norm": 0.6025921702384949,
      "learning_rate": 0.00042324533994181655,
      "loss": 3.5239,
      "step": 27450
    },
    {
      "epoch": 2.9598536217845224,
      "grad_norm": 0.5583174228668213,
      "learning_rate": 0.00042292209891175514,
      "loss": 3.5343,
      "step": 27500
    },
    {
      "epoch": 2.965235173824131,
      "grad_norm": 0.6306706070899963,
      "learning_rate": 0.00042259885788169374,
      "loss": 3.5573,
      "step": 27550
    },
    {
      "epoch": 2.970616725863739,
      "grad_norm": 0.5936201214790344,
      "learning_rate": 0.00042227561685163233,
      "loss": 3.5174,
      "step": 27600
    },
    {
      "epoch": 2.975998277903347,
      "grad_norm": 0.6141100525856018,
      "learning_rate": 0.00042195237582157093,
      "loss": 3.5331,
      "step": 27650
    },
    {
      "epoch": 2.9813798299429557,
      "grad_norm": 0.6579440832138062,
      "learning_rate": 0.0004216291347915095,
      "loss": 3.5275,
      "step": 27700
    },
    {
      "epoch": 2.986761381982564,
      "grad_norm": 0.6137058138847351,
      "learning_rate": 0.00042130589376144806,
      "loss": 3.5242,
      "step": 27750
    },
    {
      "epoch": 2.992142934022172,
      "grad_norm": 0.5903951525688171,
      "learning_rate": 0.00042098265273138666,
      "loss": 3.5355,
      "step": 27800
    },
    {
      "epoch": 2.9975244860617805,
      "grad_norm": 0.627858579158783,
      "learning_rate": 0.0004206594117013252,
      "loss": 3.5307,
      "step": 27850
    },
    {
      "epoch": 3.0029060381013886,
      "grad_norm": 0.6203857660293579,
      "learning_rate": 0.00042033617067126385,
      "loss": 3.4698,
      "step": 27900
    },
    {
      "epoch": 3.0082875901409967,
      "grad_norm": 0.5455869436264038,
      "learning_rate": 0.00042001292964120244,
      "loss": 3.4417,
      "step": 27950
    },
    {
      "epoch": 3.0136691421806048,
      "grad_norm": 0.671680748462677,
      "learning_rate": 0.000419689688611141,
      "loss": 3.4515,
      "step": 28000
    },
    {
      "epoch": 3.0136691421806048,
      "eval_accuracy": 0.3697007949371231,
      "eval_loss": 3.5062358379364014,
      "eval_runtime": 184.6177,
      "eval_samples_per_second": 97.558,
      "eval_steps_per_second": 6.099,
      "step": 28000
    },
    {
      "epoch": 3.0190506942202133,
      "grad_norm": 0.57395339012146,
      "learning_rate": 0.0004193664475810796,
      "loss": 3.4514,
      "step": 28050
    },
    {
      "epoch": 3.0244322462598214,
      "grad_norm": 0.631781816482544,
      "learning_rate": 0.00041904320655101817,
      "loss": 3.4289,
      "step": 28100
    },
    {
      "epoch": 3.0298137982994295,
      "grad_norm": 0.602963387966156,
      "learning_rate": 0.0004187199655209567,
      "loss": 3.4534,
      "step": 28150
    },
    {
      "epoch": 3.0351953503390376,
      "grad_norm": 0.5869631767272949,
      "learning_rate": 0.00041839672449089536,
      "loss": 3.4418,
      "step": 28200
    },
    {
      "epoch": 3.040576902378646,
      "grad_norm": 0.6095043420791626,
      "learning_rate": 0.00041807348346083395,
      "loss": 3.4518,
      "step": 28250
    },
    {
      "epoch": 3.0459584544182543,
      "grad_norm": 0.5966224074363708,
      "learning_rate": 0.0004177502424307725,
      "loss": 3.4571,
      "step": 28300
    },
    {
      "epoch": 3.0513400064578624,
      "grad_norm": 0.5999679565429688,
      "learning_rate": 0.0004174270014007111,
      "loss": 3.4524,
      "step": 28350
    },
    {
      "epoch": 3.0567215584974705,
      "grad_norm": 0.5879393815994263,
      "learning_rate": 0.00041710376037064963,
      "loss": 3.4557,
      "step": 28400
    },
    {
      "epoch": 3.062103110537079,
      "grad_norm": 0.5999801754951477,
      "learning_rate": 0.0004167805193405883,
      "loss": 3.4413,
      "step": 28450
    },
    {
      "epoch": 3.067484662576687,
      "grad_norm": 0.6497672200202942,
      "learning_rate": 0.0004164572783105269,
      "loss": 3.4365,
      "step": 28500
    },
    {
      "epoch": 3.0728662146162953,
      "grad_norm": 0.6005390882492065,
      "learning_rate": 0.0004161340372804654,
      "loss": 3.4541,
      "step": 28550
    },
    {
      "epoch": 3.0782477666559034,
      "grad_norm": 0.5753721594810486,
      "learning_rate": 0.000415810796250404,
      "loss": 3.4559,
      "step": 28600
    },
    {
      "epoch": 3.083629318695512,
      "grad_norm": 0.61995929479599,
      "learning_rate": 0.0004154875552203426,
      "loss": 3.4692,
      "step": 28650
    },
    {
      "epoch": 3.08901087073512,
      "grad_norm": 0.6267808675765991,
      "learning_rate": 0.00041516431419028114,
      "loss": 3.4576,
      "step": 28700
    },
    {
      "epoch": 3.094392422774728,
      "grad_norm": 0.6189929842948914,
      "learning_rate": 0.0004148410731602198,
      "loss": 3.451,
      "step": 28750
    },
    {
      "epoch": 3.0997739748143363,
      "grad_norm": 0.5664248466491699,
      "learning_rate": 0.0004145178321301584,
      "loss": 3.4663,
      "step": 28800
    },
    {
      "epoch": 3.105155526853945,
      "grad_norm": 0.6724146604537964,
      "learning_rate": 0.00041419459110009693,
      "loss": 3.4505,
      "step": 28850
    },
    {
      "epoch": 3.110537078893553,
      "grad_norm": 0.6692029237747192,
      "learning_rate": 0.0004138713500700355,
      "loss": 3.4599,
      "step": 28900
    },
    {
      "epoch": 3.115918630933161,
      "grad_norm": 0.6136066317558289,
      "learning_rate": 0.00041354810903997406,
      "loss": 3.4539,
      "step": 28950
    },
    {
      "epoch": 3.121300182972769,
      "grad_norm": 0.5619696378707886,
      "learning_rate": 0.00041322486800991266,
      "loss": 3.4524,
      "step": 29000
    },
    {
      "epoch": 3.121300182972769,
      "eval_accuracy": 0.3702130933031684,
      "eval_loss": 3.5025534629821777,
      "eval_runtime": 184.3472,
      "eval_samples_per_second": 97.702,
      "eval_steps_per_second": 6.108,
      "step": 29000
    },
    {
      "epoch": 3.1266817350123777,
      "grad_norm": 0.6812556385993958,
      "learning_rate": 0.0004129016269798513,
      "loss": 3.4678,
      "step": 29050
    },
    {
      "epoch": 3.132063287051986,
      "grad_norm": 0.7284879088401794,
      "learning_rate": 0.00041257838594978985,
      "loss": 3.435,
      "step": 29100
    },
    {
      "epoch": 3.137444839091594,
      "grad_norm": 0.641926646232605,
      "learning_rate": 0.00041225514491972844,
      "loss": 3.4607,
      "step": 29150
    },
    {
      "epoch": 3.1428263911312024,
      "grad_norm": 0.6282857656478882,
      "learning_rate": 0.00041193836871026825,
      "loss": 3.4545,
      "step": 29200
    },
    {
      "epoch": 3.1482079431708105,
      "grad_norm": 0.5846517086029053,
      "learning_rate": 0.00041161512768020684,
      "loss": 3.464,
      "step": 29250
    },
    {
      "epoch": 3.1535894952104186,
      "grad_norm": 0.5944180488586426,
      "learning_rate": 0.0004112918866501454,
      "loss": 3.4727,
      "step": 29300
    },
    {
      "epoch": 3.1589710472500268,
      "grad_norm": 0.5835195779800415,
      "learning_rate": 0.000410968645620084,
      "loss": 3.4703,
      "step": 29350
    },
    {
      "epoch": 3.1643525992896353,
      "grad_norm": 0.5912026166915894,
      "learning_rate": 0.0004106454045900226,
      "loss": 3.4553,
      "step": 29400
    },
    {
      "epoch": 3.1697341513292434,
      "grad_norm": 0.5968257784843445,
      "learning_rate": 0.00041032216355996117,
      "loss": 3.4615,
      "step": 29450
    },
    {
      "epoch": 3.1751157033688515,
      "grad_norm": 0.6082326173782349,
      "learning_rate": 0.00040999892252989976,
      "loss": 3.4585,
      "step": 29500
    },
    {
      "epoch": 3.1804972554084596,
      "grad_norm": 0.6192669868469238,
      "learning_rate": 0.00040967568149983836,
      "loss": 3.4702,
      "step": 29550
    },
    {
      "epoch": 3.185878807448068,
      "grad_norm": 0.635516881942749,
      "learning_rate": 0.0004093524404697769,
      "loss": 3.4426,
      "step": 29600
    },
    {
      "epoch": 3.1912603594876763,
      "grad_norm": 0.598032534122467,
      "learning_rate": 0.0004090291994397155,
      "loss": 3.4888,
      "step": 29650
    },
    {
      "epoch": 3.1966419115272844,
      "grad_norm": 0.6301326751708984,
      "learning_rate": 0.00040870595840965414,
      "loss": 3.4404,
      "step": 29700
    },
    {
      "epoch": 3.2020234635668925,
      "grad_norm": 0.6526699662208557,
      "learning_rate": 0.0004083827173795927,
      "loss": 3.4758,
      "step": 29750
    },
    {
      "epoch": 3.207405015606501,
      "grad_norm": 0.593436062335968,
      "learning_rate": 0.0004080594763495313,
      "loss": 3.4732,
      "step": 29800
    },
    {
      "epoch": 3.212786567646109,
      "grad_norm": 0.6289575695991516,
      "learning_rate": 0.0004077362353194698,
      "loss": 3.4443,
      "step": 29850
    },
    {
      "epoch": 3.2181681196857173,
      "grad_norm": 0.6151435375213623,
      "learning_rate": 0.0004074129942894084,
      "loss": 3.4669,
      "step": 29900
    },
    {
      "epoch": 3.2235496717253254,
      "grad_norm": 0.6291882395744324,
      "learning_rate": 0.000407089753259347,
      "loss": 3.4347,
      "step": 29950
    },
    {
      "epoch": 3.228931223764934,
      "grad_norm": 0.662102460861206,
      "learning_rate": 0.0004067665122292856,
      "loss": 3.4785,
      "step": 30000
    },
    {
      "epoch": 3.228931223764934,
      "eval_accuracy": 0.3709005401243706,
      "eval_loss": 3.4953787326812744,
      "eval_runtime": 184.3909,
      "eval_samples_per_second": 97.678,
      "eval_steps_per_second": 6.107,
      "step": 30000
    },
    {
      "epoch": 3.234312775804542,
      "grad_norm": 0.5808469653129578,
      "learning_rate": 0.0004064432711992242,
      "loss": 3.4525,
      "step": 30050
    },
    {
      "epoch": 3.23969432784415,
      "grad_norm": 0.6059449911117554,
      "learning_rate": 0.0004061200301691628,
      "loss": 3.4748,
      "step": 30100
    },
    {
      "epoch": 3.2450758798837587,
      "grad_norm": 0.6339988112449646,
      "learning_rate": 0.00040579678913910133,
      "loss": 3.4601,
      "step": 30150
    },
    {
      "epoch": 3.250457431923367,
      "grad_norm": 0.6002784371376038,
      "learning_rate": 0.0004054735481090399,
      "loss": 3.4679,
      "step": 30200
    },
    {
      "epoch": 3.255838983962975,
      "grad_norm": 0.625385582447052,
      "learning_rate": 0.0004051503070789786,
      "loss": 3.4755,
      "step": 30250
    },
    {
      "epoch": 3.261220536002583,
      "grad_norm": 0.6317866444587708,
      "learning_rate": 0.0004048270660489171,
      "loss": 3.443,
      "step": 30300
    },
    {
      "epoch": 3.2666020880421915,
      "grad_norm": 0.6008287668228149,
      "learning_rate": 0.0004045038250188557,
      "loss": 3.4729,
      "step": 30350
    },
    {
      "epoch": 3.2719836400817996,
      "grad_norm": 0.6172404885292053,
      "learning_rate": 0.00040418058398879425,
      "loss": 3.4538,
      "step": 30400
    },
    {
      "epoch": 3.2773651921214078,
      "grad_norm": 0.592491090297699,
      "learning_rate": 0.00040385734295873284,
      "loss": 3.4644,
      "step": 30450
    },
    {
      "epoch": 3.282746744161016,
      "grad_norm": 0.580609142780304,
      "learning_rate": 0.00040353410192867144,
      "loss": 3.4757,
      "step": 30500
    },
    {
      "epoch": 3.2881282962006244,
      "grad_norm": 0.6322689056396484,
      "learning_rate": 0.00040321086089861003,
      "loss": 3.4678,
      "step": 30550
    },
    {
      "epoch": 3.2935098482402325,
      "grad_norm": 0.6234239339828491,
      "learning_rate": 0.00040288761986854863,
      "loss": 3.4806,
      "step": 30600
    },
    {
      "epoch": 3.2988914002798406,
      "grad_norm": 0.6772071719169617,
      "learning_rate": 0.0004025643788384872,
      "loss": 3.4769,
      "step": 30650
    },
    {
      "epoch": 3.304272952319449,
      "grad_norm": 0.5973711013793945,
      "learning_rate": 0.00040224113780842576,
      "loss": 3.4755,
      "step": 30700
    },
    {
      "epoch": 3.3096545043590573,
      "grad_norm": 0.6178754568099976,
      "learning_rate": 0.00040191789677836436,
      "loss": 3.4569,
      "step": 30750
    },
    {
      "epoch": 3.3150360563986654,
      "grad_norm": 0.5918617248535156,
      "learning_rate": 0.0004015946557483029,
      "loss": 3.4514,
      "step": 30800
    },
    {
      "epoch": 3.3204176084382735,
      "grad_norm": 0.6223157644271851,
      "learning_rate": 0.00040127141471824155,
      "loss": 3.4609,
      "step": 30850
    },
    {
      "epoch": 3.3257991604778816,
      "grad_norm": 0.6407801508903503,
      "learning_rate": 0.00040094817368818014,
      "loss": 3.4797,
      "step": 30900
    },
    {
      "epoch": 3.33118071251749,
      "grad_norm": 0.6502392888069153,
      "learning_rate": 0.0004006249326581187,
      "loss": 3.4808,
      "step": 30950
    },
    {
      "epoch": 3.3365622645570983,
      "grad_norm": 0.5981714129447937,
      "learning_rate": 0.0004003016916280573,
      "loss": 3.4572,
      "step": 31000
    },
    {
      "epoch": 3.3365622645570983,
      "eval_accuracy": 0.37158114181364055,
      "eval_loss": 3.4903347492218018,
      "eval_runtime": 184.3602,
      "eval_samples_per_second": 97.695,
      "eval_steps_per_second": 6.108,
      "step": 31000
    },
    {
      "epoch": 3.3419438165967064,
      "grad_norm": 0.618675947189331,
      "learning_rate": 0.00039997845059799587,
      "loss": 3.4476,
      "step": 31050
    },
    {
      "epoch": 3.347325368636315,
      "grad_norm": 0.6031677722930908,
      "learning_rate": 0.00039965520956793447,
      "loss": 3.4415,
      "step": 31100
    },
    {
      "epoch": 3.352706920675923,
      "grad_norm": 0.6784326434135437,
      "learning_rate": 0.00039933196853787306,
      "loss": 3.4655,
      "step": 31150
    },
    {
      "epoch": 3.358088472715531,
      "grad_norm": 0.62276691198349,
      "learning_rate": 0.00039901519232841287,
      "loss": 3.4656,
      "step": 31200
    },
    {
      "epoch": 3.3634700247551392,
      "grad_norm": 0.6125446557998657,
      "learning_rate": 0.00039869195129835146,
      "loss": 3.4757,
      "step": 31250
    },
    {
      "epoch": 3.368851576794748,
      "grad_norm": 0.5863272547721863,
      "learning_rate": 0.00039836871026829,
      "loss": 3.4704,
      "step": 31300
    },
    {
      "epoch": 3.374233128834356,
      "grad_norm": 0.6203173398971558,
      "learning_rate": 0.0003980454692382286,
      "loss": 3.472,
      "step": 31350
    },
    {
      "epoch": 3.379614680873964,
      "grad_norm": 0.6385954022407532,
      "learning_rate": 0.0003977286930287684,
      "loss": 3.4675,
      "step": 31400
    },
    {
      "epoch": 3.384996232913572,
      "grad_norm": 0.6075577735900879,
      "learning_rate": 0.000397405451998707,
      "loss": 3.4774,
      "step": 31450
    },
    {
      "epoch": 3.3903777849531807,
      "grad_norm": 0.6865385174751282,
      "learning_rate": 0.00039708221096864554,
      "loss": 3.4593,
      "step": 31500
    },
    {
      "epoch": 3.3957593369927888,
      "grad_norm": 0.6513367891311646,
      "learning_rate": 0.00039675896993858413,
      "loss": 3.462,
      "step": 31550
    },
    {
      "epoch": 3.401140889032397,
      "grad_norm": 0.6654554009437561,
      "learning_rate": 0.0003964357289085228,
      "loss": 3.4702,
      "step": 31600
    },
    {
      "epoch": 3.4065224410720054,
      "grad_norm": 0.6225518584251404,
      "learning_rate": 0.0003961124878784613,
      "loss": 3.4611,
      "step": 31650
    },
    {
      "epoch": 3.4119039931116135,
      "grad_norm": 0.6494869589805603,
      "learning_rate": 0.0003957892468483999,
      "loss": 3.4722,
      "step": 31700
    },
    {
      "epoch": 3.4172855451512216,
      "grad_norm": 0.6048783659934998,
      "learning_rate": 0.0003954660058183385,
      "loss": 3.4691,
      "step": 31750
    },
    {
      "epoch": 3.4226670971908297,
      "grad_norm": 0.584409236907959,
      "learning_rate": 0.00039514276478827705,
      "loss": 3.4793,
      "step": 31800
    },
    {
      "epoch": 3.428048649230438,
      "grad_norm": 0.598916232585907,
      "learning_rate": 0.0003948195237582157,
      "loss": 3.4576,
      "step": 31850
    },
    {
      "epoch": 3.4334302012700464,
      "grad_norm": 0.6095468997955322,
      "learning_rate": 0.0003944962827281543,
      "loss": 3.4661,
      "step": 31900
    },
    {
      "epoch": 3.4388117533096545,
      "grad_norm": 0.6535552144050598,
      "learning_rate": 0.00039417304169809284,
      "loss": 3.4583,
      "step": 31950
    },
    {
      "epoch": 3.4441933053492626,
      "grad_norm": 0.6212128400802612,
      "learning_rate": 0.00039384980066803143,
      "loss": 3.449,
      "step": 32000
    },
    {
      "epoch": 3.4441933053492626,
      "eval_accuracy": 0.3723594224490556,
      "eval_loss": 3.4843385219573975,
      "eval_runtime": 184.576,
      "eval_samples_per_second": 97.58,
      "eval_steps_per_second": 6.1,
      "step": 32000
    },
    {
      "epoch": 3.449574857388871,
      "grad_norm": 0.6409155130386353,
      "learning_rate": 0.00039352655963796997,
      "loss": 3.4843,
      "step": 32050
    },
    {
      "epoch": 3.4549564094284793,
      "grad_norm": 0.5992752313613892,
      "learning_rate": 0.00039320331860790857,
      "loss": 3.4573,
      "step": 32100
    },
    {
      "epoch": 3.4603379614680874,
      "grad_norm": 0.6120252013206482,
      "learning_rate": 0.0003928800775778472,
      "loss": 3.4697,
      "step": 32150
    },
    {
      "epoch": 3.4657195135076955,
      "grad_norm": 0.5995035767555237,
      "learning_rate": 0.00039255683654778576,
      "loss": 3.4483,
      "step": 32200
    },
    {
      "epoch": 3.471101065547304,
      "grad_norm": 0.5921107530593872,
      "learning_rate": 0.00039223359551772435,
      "loss": 3.4831,
      "step": 32250
    },
    {
      "epoch": 3.476482617586912,
      "grad_norm": 0.6000232100486755,
      "learning_rate": 0.00039191035448766294,
      "loss": 3.4643,
      "step": 32300
    },
    {
      "epoch": 3.4818641696265202,
      "grad_norm": 0.6294697523117065,
      "learning_rate": 0.0003915871134576015,
      "loss": 3.4659,
      "step": 32350
    },
    {
      "epoch": 3.4872457216661283,
      "grad_norm": 0.6156248450279236,
      "learning_rate": 0.0003912638724275401,
      "loss": 3.455,
      "step": 32400
    },
    {
      "epoch": 3.492627273705737,
      "grad_norm": 0.61265629529953,
      "learning_rate": 0.00039094063139747873,
      "loss": 3.4608,
      "step": 32450
    },
    {
      "epoch": 3.498008825745345,
      "grad_norm": 0.6132266521453857,
      "learning_rate": 0.00039061739036741727,
      "loss": 3.4685,
      "step": 32500
    },
    {
      "epoch": 3.503390377784953,
      "grad_norm": 0.6230223774909973,
      "learning_rate": 0.00039029414933735586,
      "loss": 3.4545,
      "step": 32550
    },
    {
      "epoch": 3.5087719298245617,
      "grad_norm": 0.6874709129333496,
      "learning_rate": 0.0003899709083072944,
      "loss": 3.4696,
      "step": 32600
    },
    {
      "epoch": 3.5141534818641698,
      "grad_norm": 0.6327680945396423,
      "learning_rate": 0.000389647667277233,
      "loss": 3.4681,
      "step": 32650
    },
    {
      "epoch": 3.519535033903778,
      "grad_norm": 0.5865985155105591,
      "learning_rate": 0.00038932442624717165,
      "loss": 3.4518,
      "step": 32700
    },
    {
      "epoch": 3.524916585943386,
      "grad_norm": 0.6492707133293152,
      "learning_rate": 0.0003890011852171102,
      "loss": 3.4696,
      "step": 32750
    },
    {
      "epoch": 3.530298137982994,
      "grad_norm": 0.6094464659690857,
      "learning_rate": 0.0003886779441870488,
      "loss": 3.4615,
      "step": 32800
    },
    {
      "epoch": 3.5356796900226026,
      "grad_norm": 0.6094421148300171,
      "learning_rate": 0.0003883547031569874,
      "loss": 3.4658,
      "step": 32850
    },
    {
      "epoch": 3.5410612420622107,
      "grad_norm": 0.6602780222892761,
      "learning_rate": 0.0003880314621269259,
      "loss": 3.4751,
      "step": 32900
    },
    {
      "epoch": 3.546442794101819,
      "grad_norm": 0.5812224745750427,
      "learning_rate": 0.0003877082210968645,
      "loss": 3.4663,
      "step": 32950
    },
    {
      "epoch": 3.5518243461414274,
      "grad_norm": 0.6093171834945679,
      "learning_rate": 0.00038738498006680316,
      "loss": 3.4582,
      "step": 33000
    },
    {
      "epoch": 3.5518243461414274,
      "eval_accuracy": 0.3728994273014936,
      "eval_loss": 3.4779298305511475,
      "eval_runtime": 184.3113,
      "eval_samples_per_second": 97.721,
      "eval_steps_per_second": 6.109,
      "step": 33000
    },
    {
      "epoch": 3.5572058981810355,
      "grad_norm": 0.6080589294433594,
      "learning_rate": 0.0003870617390367417,
      "loss": 3.4671,
      "step": 33050
    },
    {
      "epoch": 3.5625874502206436,
      "grad_norm": 0.6814197301864624,
      "learning_rate": 0.0003867384980066803,
      "loss": 3.4795,
      "step": 33100
    },
    {
      "epoch": 3.5679690022602517,
      "grad_norm": 0.6045950651168823,
      "learning_rate": 0.00038641525697661884,
      "loss": 3.4743,
      "step": 33150
    },
    {
      "epoch": 3.57335055429986,
      "grad_norm": 0.6404258608818054,
      "learning_rate": 0.00038609201594655743,
      "loss": 3.4725,
      "step": 33200
    },
    {
      "epoch": 3.5787321063394684,
      "grad_norm": 0.6214661598205566,
      "learning_rate": 0.000385768774916496,
      "loss": 3.4781,
      "step": 33250
    },
    {
      "epoch": 3.5841136583790765,
      "grad_norm": 0.6475256085395813,
      "learning_rate": 0.0003854455338864346,
      "loss": 3.4684,
      "step": 33300
    },
    {
      "epoch": 3.5894952104186846,
      "grad_norm": 0.6470777988433838,
      "learning_rate": 0.0003851222928563732,
      "loss": 3.4577,
      "step": 33350
    },
    {
      "epoch": 3.594876762458293,
      "grad_norm": 0.6511114835739136,
      "learning_rate": 0.0003847990518263118,
      "loss": 3.4776,
      "step": 33400
    },
    {
      "epoch": 3.6002583144979012,
      "grad_norm": 0.6190105080604553,
      "learning_rate": 0.00038447581079625035,
      "loss": 3.4907,
      "step": 33450
    },
    {
      "epoch": 3.6056398665375093,
      "grad_norm": 0.7136397361755371,
      "learning_rate": 0.00038415256976618895,
      "loss": 3.4707,
      "step": 33500
    },
    {
      "epoch": 3.611021418577118,
      "grad_norm": 0.5738139152526855,
      "learning_rate": 0.0003838293287361275,
      "loss": 3.4645,
      "step": 33550
    },
    {
      "epoch": 3.616402970616726,
      "grad_norm": 0.5961464643478394,
      "learning_rate": 0.00038350608770606613,
      "loss": 3.44,
      "step": 33600
    },
    {
      "epoch": 3.621784522656334,
      "grad_norm": 0.6128628253936768,
      "learning_rate": 0.00038318284667600473,
      "loss": 3.4563,
      "step": 33650
    },
    {
      "epoch": 3.627166074695942,
      "grad_norm": 0.6717644333839417,
      "learning_rate": 0.00038285960564594327,
      "loss": 3.4825,
      "step": 33700
    },
    {
      "epoch": 3.6325476267355503,
      "grad_norm": 0.6172165870666504,
      "learning_rate": 0.00038253636461588186,
      "loss": 3.4586,
      "step": 33750
    },
    {
      "epoch": 3.637929178775159,
      "grad_norm": 0.637100100517273,
      "learning_rate": 0.00038221312358582046,
      "loss": 3.4587,
      "step": 33800
    },
    {
      "epoch": 3.643310730814767,
      "grad_norm": 0.6230733394622803,
      "learning_rate": 0.00038188988255575905,
      "loss": 3.4602,
      "step": 33850
    },
    {
      "epoch": 3.648692282854375,
      "grad_norm": 0.5933191776275635,
      "learning_rate": 0.00038156664152569765,
      "loss": 3.4567,
      "step": 33900
    },
    {
      "epoch": 3.6540738348939836,
      "grad_norm": 0.6499582529067993,
      "learning_rate": 0.00038124340049563624,
      "loss": 3.4669,
      "step": 33950
    },
    {
      "epoch": 3.6594553869335917,
      "grad_norm": 0.6103061437606812,
      "learning_rate": 0.0003809201594655748,
      "loss": 3.4675,
      "step": 34000
    },
    {
      "epoch": 3.6594553869335917,
      "eval_accuracy": 0.3733750009914576,
      "eval_loss": 3.4707131385803223,
      "eval_runtime": 184.5759,
      "eval_samples_per_second": 97.58,
      "eval_steps_per_second": 6.1,
      "step": 34000
    },
    {
      "epoch": 3.6648369389732,
      "grad_norm": 0.5988022089004517,
      "learning_rate": 0.0003805969184355134,
      "loss": 3.4552,
      "step": 34050
    },
    {
      "epoch": 3.670218491012808,
      "grad_norm": 0.6203043460845947,
      "learning_rate": 0.0003802736774054519,
      "loss": 3.4822,
      "step": 34100
    },
    {
      "epoch": 3.675600043052416,
      "grad_norm": 0.6239579319953918,
      "learning_rate": 0.00037995043637539057,
      "loss": 3.4933,
      "step": 34150
    },
    {
      "epoch": 3.6809815950920246,
      "grad_norm": 0.6095141172409058,
      "learning_rate": 0.00037962719534532916,
      "loss": 3.4782,
      "step": 34200
    },
    {
      "epoch": 3.6863631471316327,
      "grad_norm": 0.6401554346084595,
      "learning_rate": 0.0003793039543152677,
      "loss": 3.4487,
      "step": 34250
    },
    {
      "epoch": 3.691744699171241,
      "grad_norm": 0.6030719876289368,
      "learning_rate": 0.0003789807132852063,
      "loss": 3.4573,
      "step": 34300
    },
    {
      "epoch": 3.6971262512108494,
      "grad_norm": 0.7953944802284241,
      "learning_rate": 0.0003786574722551449,
      "loss": 3.4491,
      "step": 34350
    },
    {
      "epoch": 3.7025078032504575,
      "grad_norm": 0.6190829277038574,
      "learning_rate": 0.00037833423122508343,
      "loss": 3.4743,
      "step": 34400
    },
    {
      "epoch": 3.7078893552900656,
      "grad_norm": 0.6722561717033386,
      "learning_rate": 0.0003780109901950221,
      "loss": 3.4517,
      "step": 34450
    },
    {
      "epoch": 3.713270907329674,
      "grad_norm": 0.6021807789802551,
      "learning_rate": 0.0003776877491649607,
      "loss": 3.4531,
      "step": 34500
    },
    {
      "epoch": 3.7186524593692822,
      "grad_norm": 0.6358321905136108,
      "learning_rate": 0.0003773645081348992,
      "loss": 3.4542,
      "step": 34550
    },
    {
      "epoch": 3.7240340114088903,
      "grad_norm": 0.5931634306907654,
      "learning_rate": 0.0003770412671048378,
      "loss": 3.4659,
      "step": 34600
    },
    {
      "epoch": 3.7294155634484984,
      "grad_norm": 0.649148166179657,
      "learning_rate": 0.00037671802607477635,
      "loss": 3.4773,
      "step": 34650
    },
    {
      "epoch": 3.7347971154881066,
      "grad_norm": 0.654477059841156,
      "learning_rate": 0.000376394785044715,
      "loss": 3.455,
      "step": 34700
    },
    {
      "epoch": 3.740178667527715,
      "grad_norm": 0.6332741975784302,
      "learning_rate": 0.0003760715440146536,
      "loss": 3.4623,
      "step": 34750
    },
    {
      "epoch": 3.745560219567323,
      "grad_norm": 0.6453436017036438,
      "learning_rate": 0.00037574830298459214,
      "loss": 3.461,
      "step": 34800
    },
    {
      "epoch": 3.7509417716069313,
      "grad_norm": 0.6450361013412476,
      "learning_rate": 0.00037542506195453073,
      "loss": 3.463,
      "step": 34850
    },
    {
      "epoch": 3.75632332364654,
      "grad_norm": 0.5885767936706543,
      "learning_rate": 0.0003751018209244693,
      "loss": 3.4671,
      "step": 34900
    },
    {
      "epoch": 3.761704875686148,
      "grad_norm": 0.6284133195877075,
      "learning_rate": 0.00037477857989440787,
      "loss": 3.4414,
      "step": 34950
    },
    {
      "epoch": 3.767086427725756,
      "grad_norm": 0.6213708519935608,
      "learning_rate": 0.0003744553388643465,
      "loss": 3.4701,
      "step": 35000
    },
    {
      "epoch": 3.767086427725756,
      "eval_accuracy": 0.3740369143840234,
      "eval_loss": 3.4637458324432373,
      "eval_runtime": 184.3099,
      "eval_samples_per_second": 97.721,
      "eval_steps_per_second": 6.109,
      "step": 35000
    },
    {
      "epoch": 3.772467979765364,
      "grad_norm": 0.5990517735481262,
      "learning_rate": 0.0003741320978342851,
      "loss": 3.4614,
      "step": 35050
    },
    {
      "epoch": 3.7778495318049723,
      "grad_norm": 0.6437912583351135,
      "learning_rate": 0.00037380885680422365,
      "loss": 3.4687,
      "step": 35100
    },
    {
      "epoch": 3.783231083844581,
      "grad_norm": 0.6233482360839844,
      "learning_rate": 0.00037348561577416224,
      "loss": 3.4596,
      "step": 35150
    },
    {
      "epoch": 3.788612635884189,
      "grad_norm": 0.6534203290939331,
      "learning_rate": 0.0003731623747441008,
      "loss": 3.4775,
      "step": 35200
    },
    {
      "epoch": 3.793994187923797,
      "grad_norm": 0.6427837610244751,
      "learning_rate": 0.0003728391337140394,
      "loss": 3.4657,
      "step": 35250
    },
    {
      "epoch": 3.7993757399634056,
      "grad_norm": 0.6595858335494995,
      "learning_rate": 0.00037251589268397803,
      "loss": 3.4605,
      "step": 35300
    },
    {
      "epoch": 3.8047572920030137,
      "grad_norm": 0.6350520849227905,
      "learning_rate": 0.00037219265165391657,
      "loss": 3.4658,
      "step": 35350
    },
    {
      "epoch": 3.810138844042622,
      "grad_norm": 0.6624295115470886,
      "learning_rate": 0.00037187587544445643,
      "loss": 3.4547,
      "step": 35400
    },
    {
      "epoch": 3.8155203960822304,
      "grad_norm": 0.6359967589378357,
      "learning_rate": 0.00037155263441439497,
      "loss": 3.4758,
      "step": 35450
    },
    {
      "epoch": 3.8209019481218385,
      "grad_norm": 0.6356232762336731,
      "learning_rate": 0.00037122939338433356,
      "loss": 3.4659,
      "step": 35500
    },
    {
      "epoch": 3.8262835001614466,
      "grad_norm": 0.6278019547462463,
      "learning_rate": 0.0003709061523542721,
      "loss": 3.4546,
      "step": 35550
    },
    {
      "epoch": 3.8316650522010547,
      "grad_norm": 0.6327377557754517,
      "learning_rate": 0.0003705829113242107,
      "loss": 3.4567,
      "step": 35600
    },
    {
      "epoch": 3.837046604240663,
      "grad_norm": 0.6610739231109619,
      "learning_rate": 0.00037025967029414935,
      "loss": 3.4564,
      "step": 35650
    },
    {
      "epoch": 3.8424281562802713,
      "grad_norm": 0.5995158553123474,
      "learning_rate": 0.0003699364292640879,
      "loss": 3.4601,
      "step": 35700
    },
    {
      "epoch": 3.8478097083198795,
      "grad_norm": 0.6885281205177307,
      "learning_rate": 0.0003696131882340265,
      "loss": 3.4575,
      "step": 35750
    },
    {
      "epoch": 3.8531912603594876,
      "grad_norm": 0.5718228816986084,
      "learning_rate": 0.0003692899472039651,
      "loss": 3.4668,
      "step": 35800
    },
    {
      "epoch": 3.858572812399096,
      "grad_norm": 0.6344829797744751,
      "learning_rate": 0.0003689667061739036,
      "loss": 3.4601,
      "step": 35850
    },
    {
      "epoch": 3.863954364438704,
      "grad_norm": 0.6443964838981628,
      "learning_rate": 0.0003686434651438422,
      "loss": 3.4453,
      "step": 35900
    },
    {
      "epoch": 3.8693359164783123,
      "grad_norm": 0.6165390610694885,
      "learning_rate": 0.00036832022411378086,
      "loss": 3.4414,
      "step": 35950
    },
    {
      "epoch": 3.8747174685179204,
      "grad_norm": 0.6989316344261169,
      "learning_rate": 0.0003679969830837194,
      "loss": 3.457,
      "step": 36000
    },
    {
      "epoch": 3.8747174685179204,
      "eval_accuracy": 0.37468611538871494,
      "eval_loss": 3.4592158794403076,
      "eval_runtime": 184.5704,
      "eval_samples_per_second": 97.583,
      "eval_steps_per_second": 6.101,
      "step": 36000
    },
    {
      "epoch": 3.8800990205575285,
      "grad_norm": 0.617153525352478,
      "learning_rate": 0.000367673742053658,
      "loss": 3.4632,
      "step": 36050
    },
    {
      "epoch": 3.885480572597137,
      "grad_norm": 0.6491139531135559,
      "learning_rate": 0.00036735050102359654,
      "loss": 3.4713,
      "step": 36100
    },
    {
      "epoch": 3.890862124636745,
      "grad_norm": 0.6319408416748047,
      "learning_rate": 0.00036702725999353513,
      "loss": 3.4462,
      "step": 36150
    },
    {
      "epoch": 3.8962436766763533,
      "grad_norm": 0.6488457322120667,
      "learning_rate": 0.0003667040189634737,
      "loss": 3.4514,
      "step": 36200
    },
    {
      "epoch": 3.901625228715962,
      "grad_norm": 0.623210608959198,
      "learning_rate": 0.0003663807779334123,
      "loss": 3.4429,
      "step": 36250
    },
    {
      "epoch": 3.90700678075557,
      "grad_norm": 0.6426396369934082,
      "learning_rate": 0.0003660575369033509,
      "loss": 3.4685,
      "step": 36300
    },
    {
      "epoch": 3.912388332795178,
      "grad_norm": 0.6298360228538513,
      "learning_rate": 0.0003657342958732895,
      "loss": 3.457,
      "step": 36350
    },
    {
      "epoch": 3.9177698848347866,
      "grad_norm": 0.6557319760322571,
      "learning_rate": 0.00036541105484322805,
      "loss": 3.4445,
      "step": 36400
    },
    {
      "epoch": 3.9231514368743947,
      "grad_norm": 0.6214303374290466,
      "learning_rate": 0.00036508781381316665,
      "loss": 3.4688,
      "step": 36450
    },
    {
      "epoch": 3.928532988914003,
      "grad_norm": 0.6472518444061279,
      "learning_rate": 0.0003647645727831053,
      "loss": 3.4611,
      "step": 36500
    },
    {
      "epoch": 3.933914540953611,
      "grad_norm": 0.588938295841217,
      "learning_rate": 0.00036444133175304384,
      "loss": 3.4494,
      "step": 36550
    },
    {
      "epoch": 3.939296092993219,
      "grad_norm": 0.6096231937408447,
      "learning_rate": 0.00036411809072298243,
      "loss": 3.4732,
      "step": 36600
    },
    {
      "epoch": 3.9446776450328276,
      "grad_norm": 0.7059899568557739,
      "learning_rate": 0.00036379484969292097,
      "loss": 3.4554,
      "step": 36650
    },
    {
      "epoch": 3.9500591970724357,
      "grad_norm": 0.6259297132492065,
      "learning_rate": 0.00036347160866285956,
      "loss": 3.4683,
      "step": 36700
    },
    {
      "epoch": 3.955440749112044,
      "grad_norm": 0.5921356678009033,
      "learning_rate": 0.0003631483676327981,
      "loss": 3.4667,
      "step": 36750
    },
    {
      "epoch": 3.9608223011516523,
      "grad_norm": 0.6173107028007507,
      "learning_rate": 0.00036282512660273675,
      "loss": 3.462,
      "step": 36800
    },
    {
      "epoch": 3.9662038531912605,
      "grad_norm": 0.6219081282615662,
      "learning_rate": 0.00036250188557267535,
      "loss": 3.4624,
      "step": 36850
    },
    {
      "epoch": 3.9715854052308686,
      "grad_norm": 0.598000705242157,
      "learning_rate": 0.0003621786445426139,
      "loss": 3.4539,
      "step": 36900
    },
    {
      "epoch": 3.9769669572704767,
      "grad_norm": 0.6490669846534729,
      "learning_rate": 0.0003618554035125525,
      "loss": 3.4456,
      "step": 36950
    },
    {
      "epoch": 3.9823485093100848,
      "grad_norm": 0.7354153394699097,
      "learning_rate": 0.0003615321624824911,
      "loss": 3.4467,
      "step": 37000
    },
    {
      "epoch": 3.9823485093100848,
      "eval_accuracy": 0.3750956281228877,
      "eval_loss": 3.454505443572998,
      "eval_runtime": 184.2543,
      "eval_samples_per_second": 97.751,
      "eval_steps_per_second": 6.111,
      "step": 37000
    },
    {
      "epoch": 3.9877300613496933,
      "grad_norm": 0.5946967005729675,
      "learning_rate": 0.0003612089214524296,
      "loss": 3.455,
      "step": 37050
    },
    {
      "epoch": 3.9931116133893014,
      "grad_norm": 0.6418635249137878,
      "learning_rate": 0.00036088568042236827,
      "loss": 3.4694,
      "step": 37100
    },
    {
      "epoch": 3.9984931654289095,
      "grad_norm": 0.636024534702301,
      "learning_rate": 0.00036056243939230686,
      "loss": 3.4698,
      "step": 37150
    },
    {
      "epoch": 4.003874717468518,
      "grad_norm": 0.6422595977783203,
      "learning_rate": 0.0003602391983622454,
      "loss": 3.3863,
      "step": 37200
    },
    {
      "epoch": 4.009256269508126,
      "grad_norm": 0.6733551621437073,
      "learning_rate": 0.000359915957332184,
      "loss": 3.3523,
      "step": 37250
    },
    {
      "epoch": 4.014637821547734,
      "grad_norm": 0.6378384232521057,
      "learning_rate": 0.00035959271630212254,
      "loss": 3.3541,
      "step": 37300
    },
    {
      "epoch": 4.020019373587343,
      "grad_norm": 0.6563474535942078,
      "learning_rate": 0.00035926947527206113,
      "loss": 3.3753,
      "step": 37350
    },
    {
      "epoch": 4.0254009256269505,
      "grad_norm": 0.6798022985458374,
      "learning_rate": 0.0003589462342419998,
      "loss": 3.3627,
      "step": 37400
    },
    {
      "epoch": 4.030782477666559,
      "grad_norm": 0.633373498916626,
      "learning_rate": 0.0003586294580325396,
      "loss": 3.3615,
      "step": 37450
    },
    {
      "epoch": 4.036164029706168,
      "grad_norm": 0.6773637533187866,
      "learning_rate": 0.0003583062170024782,
      "loss": 3.3792,
      "step": 37500
    },
    {
      "epoch": 4.041545581745775,
      "grad_norm": 0.664168119430542,
      "learning_rate": 0.0003579829759724167,
      "loss": 3.3762,
      "step": 37550
    },
    {
      "epoch": 4.046927133785384,
      "grad_norm": 0.6700920462608337,
      "learning_rate": 0.0003576597349423553,
      "loss": 3.3813,
      "step": 37600
    },
    {
      "epoch": 4.0523086858249915,
      "grad_norm": 0.6552506685256958,
      "learning_rate": 0.0003573364939122939,
      "loss": 3.3694,
      "step": 37650
    },
    {
      "epoch": 4.0576902378646,
      "grad_norm": 0.656863808631897,
      "learning_rate": 0.00035701325288223245,
      "loss": 3.3803,
      "step": 37700
    },
    {
      "epoch": 4.063071789904209,
      "grad_norm": 0.639150857925415,
      "learning_rate": 0.0003566900118521711,
      "loss": 3.3863,
      "step": 37750
    },
    {
      "epoch": 4.068453341943816,
      "grad_norm": 0.6749435663223267,
      "learning_rate": 0.0003563667708221097,
      "loss": 3.3896,
      "step": 37800
    },
    {
      "epoch": 4.073834893983425,
      "grad_norm": 0.6099902987480164,
      "learning_rate": 0.00035604352979204824,
      "loss": 3.3743,
      "step": 37850
    },
    {
      "epoch": 4.079216446023033,
      "grad_norm": 0.7206007838249207,
      "learning_rate": 0.00035572028876198683,
      "loss": 3.384,
      "step": 37900
    },
    {
      "epoch": 4.084597998062641,
      "grad_norm": 0.6344521641731262,
      "learning_rate": 0.00035539704773192537,
      "loss": 3.3811,
      "step": 37950
    },
    {
      "epoch": 4.08997955010225,
      "grad_norm": 0.6181900501251221,
      "learning_rate": 0.00035507380670186397,
      "loss": 3.3598,
      "step": 38000
    },
    {
      "epoch": 4.08997955010225,
      "eval_accuracy": 0.3759688713822527,
      "eval_loss": 3.456158399581909,
      "eval_runtime": 184.3784,
      "eval_samples_per_second": 97.685,
      "eval_steps_per_second": 6.107,
      "step": 38000
    },
    {
      "epoch": 4.095361102141858,
      "grad_norm": 0.5788723826408386,
      "learning_rate": 0.0003547505656718026,
      "loss": 3.3677,
      "step": 38050
    },
    {
      "epoch": 4.100742654181466,
      "grad_norm": 0.6603501439094543,
      "learning_rate": 0.00035442732464174116,
      "loss": 3.394,
      "step": 38100
    },
    {
      "epoch": 4.106124206221074,
      "grad_norm": 0.5997026562690735,
      "learning_rate": 0.00035410408361167975,
      "loss": 3.4038,
      "step": 38150
    },
    {
      "epoch": 4.111505758260682,
      "grad_norm": 0.6239067316055298,
      "learning_rate": 0.0003537808425816183,
      "loss": 3.3761,
      "step": 38200
    },
    {
      "epoch": 4.1168873103002905,
      "grad_norm": 0.653221070766449,
      "learning_rate": 0.0003534576015515569,
      "loss": 3.3777,
      "step": 38250
    },
    {
      "epoch": 4.122268862339899,
      "grad_norm": 0.6475557088851929,
      "learning_rate": 0.00035313436052149553,
      "loss": 3.3918,
      "step": 38300
    },
    {
      "epoch": 4.127650414379507,
      "grad_norm": 0.7148786187171936,
      "learning_rate": 0.0003528111194914341,
      "loss": 3.3838,
      "step": 38350
    },
    {
      "epoch": 4.133031966419115,
      "grad_norm": 0.6013211607933044,
      "learning_rate": 0.00035248787846137267,
      "loss": 3.3891,
      "step": 38400
    },
    {
      "epoch": 4.138413518458724,
      "grad_norm": 0.6526707410812378,
      "learning_rate": 0.00035216463743131126,
      "loss": 3.3767,
      "step": 38450
    },
    {
      "epoch": 4.1437950704983315,
      "grad_norm": 0.6397549510002136,
      "learning_rate": 0.0003518413964012498,
      "loss": 3.3754,
      "step": 38500
    },
    {
      "epoch": 4.14917662253794,
      "grad_norm": 0.6478105783462524,
      "learning_rate": 0.0003515181553711884,
      "loss": 3.4084,
      "step": 38550
    },
    {
      "epoch": 4.154558174577549,
      "grad_norm": 0.6594587564468384,
      "learning_rate": 0.00035119491434112705,
      "loss": 3.3689,
      "step": 38600
    },
    {
      "epoch": 4.159939726617156,
      "grad_norm": 0.695050835609436,
      "learning_rate": 0.0003508716733110656,
      "loss": 3.3758,
      "step": 38650
    },
    {
      "epoch": 4.165321278656765,
      "grad_norm": 0.6472200751304626,
      "learning_rate": 0.0003505484322810042,
      "loss": 3.3931,
      "step": 38700
    },
    {
      "epoch": 4.1707028306963725,
      "grad_norm": 0.6525875926017761,
      "learning_rate": 0.0003502251912509427,
      "loss": 3.3826,
      "step": 38750
    },
    {
      "epoch": 4.176084382735981,
      "grad_norm": 0.6336781978607178,
      "learning_rate": 0.0003499019502208813,
      "loss": 3.3954,
      "step": 38800
    },
    {
      "epoch": 4.18146593477559,
      "grad_norm": 0.6761037111282349,
      "learning_rate": 0.0003495787091908199,
      "loss": 3.4037,
      "step": 38850
    },
    {
      "epoch": 4.186847486815197,
      "grad_norm": 0.6590827703475952,
      "learning_rate": 0.0003492554681607585,
      "loss": 3.3931,
      "step": 38900
    },
    {
      "epoch": 4.192229038854806,
      "grad_norm": 0.6215708255767822,
      "learning_rate": 0.0003489322271306971,
      "loss": 3.3914,
      "step": 38950
    },
    {
      "epoch": 4.197610590894414,
      "grad_norm": 0.609784722328186,
      "learning_rate": 0.0003486089861006357,
      "loss": 3.3772,
      "step": 39000
    },
    {
      "epoch": 4.197610590894414,
      "eval_accuracy": 0.3758620655935311,
      "eval_loss": 3.451266288757324,
      "eval_runtime": 184.6044,
      "eval_samples_per_second": 97.565,
      "eval_steps_per_second": 6.1,
      "step": 39000
    },
    {
      "epoch": 4.202992142934022,
      "grad_norm": 0.6011910438537598,
      "learning_rate": 0.00034828574507057424,
      "loss": 3.3764,
      "step": 39050
    },
    {
      "epoch": 4.208373694973631,
      "grad_norm": 0.6473610401153564,
      "learning_rate": 0.00034796250404051283,
      "loss": 3.3889,
      "step": 39100
    },
    {
      "epoch": 4.213755247013238,
      "grad_norm": 0.6244720220565796,
      "learning_rate": 0.00034763926301045137,
      "loss": 3.3944,
      "step": 39150
    },
    {
      "epoch": 4.219136799052847,
      "grad_norm": 0.6898408532142639,
      "learning_rate": 0.00034731602198039,
      "loss": 3.3885,
      "step": 39200
    },
    {
      "epoch": 4.224518351092455,
      "grad_norm": 0.6290667653083801,
      "learning_rate": 0.0003469927809503286,
      "loss": 3.4119,
      "step": 39250
    },
    {
      "epoch": 4.229899903132063,
      "grad_norm": 0.6227483749389648,
      "learning_rate": 0.00034666953992026716,
      "loss": 3.3768,
      "step": 39300
    },
    {
      "epoch": 4.2352814551716715,
      "grad_norm": 0.6508646607398987,
      "learning_rate": 0.00034634629889020575,
      "loss": 3.3807,
      "step": 39350
    },
    {
      "epoch": 4.24066300721128,
      "grad_norm": 0.6218987703323364,
      "learning_rate": 0.00034602305786014435,
      "loss": 3.3786,
      "step": 39400
    },
    {
      "epoch": 4.246044559250888,
      "grad_norm": 0.6416997313499451,
      "learning_rate": 0.00034569981683008294,
      "loss": 3.3822,
      "step": 39450
    },
    {
      "epoch": 4.251426111290496,
      "grad_norm": 0.6093468070030212,
      "learning_rate": 0.0003453830406206227,
      "loss": 3.3974,
      "step": 39500
    },
    {
      "epoch": 4.256807663330104,
      "grad_norm": 0.6778056621551514,
      "learning_rate": 0.00034505979959056134,
      "loss": 3.378,
      "step": 39550
    },
    {
      "epoch": 4.2621892153697125,
      "grad_norm": 0.6332718729972839,
      "learning_rate": 0.00034473655856049994,
      "loss": 3.3894,
      "step": 39600
    },
    {
      "epoch": 4.267570767409321,
      "grad_norm": 0.6384563446044922,
      "learning_rate": 0.0003444133175304385,
      "loss": 3.407,
      "step": 39650
    },
    {
      "epoch": 4.272952319448929,
      "grad_norm": 0.6422023177146912,
      "learning_rate": 0.00034409007650037707,
      "loss": 3.3904,
      "step": 39700
    },
    {
      "epoch": 4.278333871488537,
      "grad_norm": 0.6447274684906006,
      "learning_rate": 0.00034376683547031567,
      "loss": 3.3862,
      "step": 39750
    },
    {
      "epoch": 4.283715423528146,
      "grad_norm": 0.6438936591148376,
      "learning_rate": 0.0003434435944402542,
      "loss": 3.4031,
      "step": 39800
    },
    {
      "epoch": 4.2890969755677535,
      "grad_norm": 0.6367171406745911,
      "learning_rate": 0.00034312035341019286,
      "loss": 3.4215,
      "step": 39850
    },
    {
      "epoch": 4.294478527607362,
      "grad_norm": 0.6519704461097717,
      "learning_rate": 0.00034279711238013145,
      "loss": 3.4026,
      "step": 39900
    },
    {
      "epoch": 4.299860079646971,
      "grad_norm": 0.6124309301376343,
      "learning_rate": 0.00034247387135007,
      "loss": 3.3958,
      "step": 39950
    },
    {
      "epoch": 4.305241631686578,
      "grad_norm": 0.6547208428382874,
      "learning_rate": 0.0003421506303200086,
      "loss": 3.389,
      "step": 40000
    },
    {
      "epoch": 4.305241631686578,
      "eval_accuracy": 0.37677290275207986,
      "eval_loss": 3.4454505443573,
      "eval_runtime": 184.2303,
      "eval_samples_per_second": 97.763,
      "eval_steps_per_second": 6.112,
      "step": 40000
    },
    {
      "epoch": 4.310623183726187,
      "grad_norm": 0.6543096899986267,
      "learning_rate": 0.0003418273892899471,
      "loss": 3.3993,
      "step": 40050
    },
    {
      "epoch": 4.3160047357657945,
      "grad_norm": 0.6158707737922668,
      "learning_rate": 0.0003415041482598858,
      "loss": 3.3985,
      "step": 40100
    },
    {
      "epoch": 4.321386287805403,
      "grad_norm": 0.6724017858505249,
      "learning_rate": 0.00034118090722982437,
      "loss": 3.4014,
      "step": 40150
    },
    {
      "epoch": 4.326767839845012,
      "grad_norm": 0.642092764377594,
      "learning_rate": 0.0003408576661997629,
      "loss": 3.4103,
      "step": 40200
    },
    {
      "epoch": 4.332149391884619,
      "grad_norm": 0.6737988591194153,
      "learning_rate": 0.0003405344251697015,
      "loss": 3.4029,
      "step": 40250
    },
    {
      "epoch": 4.337530943924228,
      "grad_norm": 0.7181668281555176,
      "learning_rate": 0.0003402111841396401,
      "loss": 3.4004,
      "step": 40300
    },
    {
      "epoch": 4.342912495963836,
      "grad_norm": 0.6153748631477356,
      "learning_rate": 0.00033988794310957864,
      "loss": 3.3895,
      "step": 40350
    },
    {
      "epoch": 4.348294048003444,
      "grad_norm": 0.6152939796447754,
      "learning_rate": 0.0003395647020795173,
      "loss": 3.4012,
      "step": 40400
    },
    {
      "epoch": 4.3536756000430525,
      "grad_norm": 0.6356926560401917,
      "learning_rate": 0.0003392414610494559,
      "loss": 3.3911,
      "step": 40450
    },
    {
      "epoch": 4.359057152082661,
      "grad_norm": 0.6974787712097168,
      "learning_rate": 0.0003389182200193944,
      "loss": 3.4039,
      "step": 40500
    },
    {
      "epoch": 4.364438704122269,
      "grad_norm": 0.6503478288650513,
      "learning_rate": 0.000338594978989333,
      "loss": 3.3963,
      "step": 40550
    },
    {
      "epoch": 4.369820256161877,
      "grad_norm": 0.6200667023658752,
      "learning_rate": 0.00033827173795927156,
      "loss": 3.3816,
      "step": 40600
    },
    {
      "epoch": 4.375201808201485,
      "grad_norm": 0.6516591310501099,
      "learning_rate": 0.00033794849692921015,
      "loss": 3.3725,
      "step": 40650
    },
    {
      "epoch": 4.3805833602410935,
      "grad_norm": 0.6140480637550354,
      "learning_rate": 0.0003376252558991488,
      "loss": 3.3972,
      "step": 40700
    },
    {
      "epoch": 4.385964912280702,
      "grad_norm": 0.7277047634124756,
      "learning_rate": 0.00033730201486908734,
      "loss": 3.3904,
      "step": 40750
    },
    {
      "epoch": 4.39134646432031,
      "grad_norm": 0.6539707779884338,
      "learning_rate": 0.00033697877383902594,
      "loss": 3.4112,
      "step": 40800
    },
    {
      "epoch": 4.396728016359918,
      "grad_norm": 0.6106005311012268,
      "learning_rate": 0.00033665553280896453,
      "loss": 3.3942,
      "step": 40850
    },
    {
      "epoch": 4.402109568399527,
      "grad_norm": 0.6393513679504395,
      "learning_rate": 0.00033633229177890307,
      "loss": 3.4147,
      "step": 40900
    },
    {
      "epoch": 4.4074911204391345,
      "grad_norm": 0.6651601791381836,
      "learning_rate": 0.00033600905074884167,
      "loss": 3.404,
      "step": 40950
    },
    {
      "epoch": 4.412872672478743,
      "grad_norm": 0.6897532343864441,
      "learning_rate": 0.0003356858097187803,
      "loss": 3.3824,
      "step": 41000
    },
    {
      "epoch": 4.412872672478743,
      "eval_accuracy": 0.37701650252655994,
      "eval_loss": 3.442161798477173,
      "eval_runtime": 184.7744,
      "eval_samples_per_second": 97.476,
      "eval_steps_per_second": 6.094,
      "step": 41000
    },
    {
      "epoch": 4.418254224518351,
      "grad_norm": 0.6652800440788269,
      "learning_rate": 0.00033536256868871886,
      "loss": 3.4106,
      "step": 41050
    },
    {
      "epoch": 4.423635776557959,
      "grad_norm": 0.6520366668701172,
      "learning_rate": 0.00033503932765865745,
      "loss": 3.4035,
      "step": 41100
    },
    {
      "epoch": 4.429017328597568,
      "grad_norm": 0.6537342071533203,
      "learning_rate": 0.000334716086628596,
      "loss": 3.3932,
      "step": 41150
    },
    {
      "epoch": 4.4343988806371755,
      "grad_norm": 0.6996263861656189,
      "learning_rate": 0.0003343928455985346,
      "loss": 3.4005,
      "step": 41200
    },
    {
      "epoch": 4.439780432676784,
      "grad_norm": 0.6283782720565796,
      "learning_rate": 0.00033406960456847324,
      "loss": 3.3904,
      "step": 41250
    },
    {
      "epoch": 4.445161984716393,
      "grad_norm": 0.6458320617675781,
      "learning_rate": 0.0003337463635384118,
      "loss": 3.4019,
      "step": 41300
    },
    {
      "epoch": 4.450543536756,
      "grad_norm": 0.6539556980133057,
      "learning_rate": 0.00033342312250835037,
      "loss": 3.4051,
      "step": 41350
    },
    {
      "epoch": 4.455925088795609,
      "grad_norm": 0.6024113893508911,
      "learning_rate": 0.00033309988147828896,
      "loss": 3.4083,
      "step": 41400
    },
    {
      "epoch": 4.461306640835216,
      "grad_norm": 0.6404975652694702,
      "learning_rate": 0.0003327766404482275,
      "loss": 3.3942,
      "step": 41450
    },
    {
      "epoch": 4.466688192874825,
      "grad_norm": 0.6734680533409119,
      "learning_rate": 0.0003324598642387673,
      "loss": 3.3891,
      "step": 41500
    },
    {
      "epoch": 4.4720697449144335,
      "grad_norm": 0.5948939323425293,
      "learning_rate": 0.0003321366232087059,
      "loss": 3.4035,
      "step": 41550
    },
    {
      "epoch": 4.477451296954041,
      "grad_norm": 0.6198262572288513,
      "learning_rate": 0.0003318133821786445,
      "loss": 3.41,
      "step": 41600
    },
    {
      "epoch": 4.48283284899365,
      "grad_norm": 0.634553849697113,
      "learning_rate": 0.0003314901411485831,
      "loss": 3.3812,
      "step": 41650
    },
    {
      "epoch": 4.488214401033258,
      "grad_norm": 0.6665427684783936,
      "learning_rate": 0.0003311669001185217,
      "loss": 3.3875,
      "step": 41700
    },
    {
      "epoch": 4.493595953072866,
      "grad_norm": 0.6211308836936951,
      "learning_rate": 0.0003308436590884603,
      "loss": 3.3919,
      "step": 41750
    },
    {
      "epoch": 4.4989775051124745,
      "grad_norm": 0.6392455697059631,
      "learning_rate": 0.0003305204180583988,
      "loss": 3.4245,
      "step": 41800
    },
    {
      "epoch": 4.504359057152083,
      "grad_norm": 0.6825820803642273,
      "learning_rate": 0.0003301971770283374,
      "loss": 3.4147,
      "step": 41850
    },
    {
      "epoch": 4.509740609191691,
      "grad_norm": 0.6289772391319275,
      "learning_rate": 0.00032987393599827607,
      "loss": 3.4194,
      "step": 41900
    },
    {
      "epoch": 4.515122161231299,
      "grad_norm": 0.6444154977798462,
      "learning_rate": 0.0003295506949682146,
      "loss": 3.4085,
      "step": 41950
    },
    {
      "epoch": 4.520503713270907,
      "grad_norm": 0.6935963034629822,
      "learning_rate": 0.0003292274539381532,
      "loss": 3.3907,
      "step": 42000
    },
    {
      "epoch": 4.520503713270907,
      "eval_accuracy": 0.3773419179255643,
      "eval_loss": 3.436394691467285,
      "eval_runtime": 184.3724,
      "eval_samples_per_second": 97.688,
      "eval_steps_per_second": 6.107,
      "step": 42000
    },
    {
      "epoch": 4.5258852653105155,
      "grad_norm": 0.6008836627006531,
      "learning_rate": 0.00032890421290809174,
      "loss": 3.4039,
      "step": 42050
    },
    {
      "epoch": 4.531266817350124,
      "grad_norm": 0.6186881065368652,
      "learning_rate": 0.00032858097187803034,
      "loss": 3.4019,
      "step": 42100
    },
    {
      "epoch": 4.536648369389732,
      "grad_norm": 0.6503677368164062,
      "learning_rate": 0.00032825773084796893,
      "loss": 3.4049,
      "step": 42150
    },
    {
      "epoch": 4.54202992142934,
      "grad_norm": 0.6397423148155212,
      "learning_rate": 0.00032793448981790753,
      "loss": 3.3936,
      "step": 42200
    },
    {
      "epoch": 4.547411473468949,
      "grad_norm": 0.6151939630508423,
      "learning_rate": 0.0003276112487878461,
      "loss": 3.3966,
      "step": 42250
    },
    {
      "epoch": 4.5527930255085565,
      "grad_norm": 0.6326496005058289,
      "learning_rate": 0.0003272880077577847,
      "loss": 3.4021,
      "step": 42300
    },
    {
      "epoch": 4.558174577548165,
      "grad_norm": 0.6877431273460388,
      "learning_rate": 0.00032696476672772326,
      "loss": 3.4072,
      "step": 42350
    },
    {
      "epoch": 4.563556129587774,
      "grad_norm": 0.6554200053215027,
      "learning_rate": 0.00032664152569766185,
      "loss": 3.4198,
      "step": 42400
    },
    {
      "epoch": 4.568937681627381,
      "grad_norm": 0.64201420545578,
      "learning_rate": 0.0003263182846676004,
      "loss": 3.4084,
      "step": 42450
    },
    {
      "epoch": 4.57431923366699,
      "grad_norm": 0.6482923626899719,
      "learning_rate": 0.00032599504363753904,
      "loss": 3.4052,
      "step": 42500
    },
    {
      "epoch": 4.579700785706597,
      "grad_norm": 0.6746305227279663,
      "learning_rate": 0.00032567180260747764,
      "loss": 3.4005,
      "step": 42550
    },
    {
      "epoch": 4.585082337746206,
      "grad_norm": 0.6510485410690308,
      "learning_rate": 0.0003253485615774162,
      "loss": 3.3892,
      "step": 42600
    },
    {
      "epoch": 4.5904638897858145,
      "grad_norm": 0.7001817226409912,
      "learning_rate": 0.00032502532054735477,
      "loss": 3.3951,
      "step": 42650
    },
    {
      "epoch": 4.595845441825422,
      "grad_norm": 0.6700904965400696,
      "learning_rate": 0.00032470207951729337,
      "loss": 3.4037,
      "step": 42700
    },
    {
      "epoch": 4.601226993865031,
      "grad_norm": 0.6753993034362793,
      "learning_rate": 0.0003243788384872319,
      "loss": 3.4072,
      "step": 42750
    },
    {
      "epoch": 4.606608545904638,
      "grad_norm": 0.6279345750808716,
      "learning_rate": 0.00032405559745717056,
      "loss": 3.3958,
      "step": 42800
    },
    {
      "epoch": 4.611990097944247,
      "grad_norm": 0.6324381828308105,
      "learning_rate": 0.00032373235642710915,
      "loss": 3.3918,
      "step": 42850
    },
    {
      "epoch": 4.6173716499838555,
      "grad_norm": 0.6431390643119812,
      "learning_rate": 0.0003234091153970477,
      "loss": 3.3907,
      "step": 42900
    },
    {
      "epoch": 4.622753202023463,
      "grad_norm": 0.6529428362846375,
      "learning_rate": 0.0003230858743669863,
      "loss": 3.3876,
      "step": 42950
    },
    {
      "epoch": 4.628134754063072,
      "grad_norm": 0.6419126391410828,
      "learning_rate": 0.0003227626333369248,
      "loss": 3.3859,
      "step": 43000
    },
    {
      "epoch": 4.628134754063072,
      "eval_accuracy": 0.3782851336446818,
      "eval_loss": 3.4323384761810303,
      "eval_runtime": 184.4945,
      "eval_samples_per_second": 97.624,
      "eval_steps_per_second": 6.103,
      "step": 43000
    },
    {
      "epoch": 4.63351630610268,
      "grad_norm": 0.6662230491638184,
      "learning_rate": 0.0003224393923068635,
      "loss": 3.3911,
      "step": 43050
    },
    {
      "epoch": 4.638897858142288,
      "grad_norm": 0.7196651697158813,
      "learning_rate": 0.00032211615127680207,
      "loss": 3.4014,
      "step": 43100
    },
    {
      "epoch": 4.6442794101818965,
      "grad_norm": 0.6619535088539124,
      "learning_rate": 0.0003217929102467406,
      "loss": 3.3932,
      "step": 43150
    },
    {
      "epoch": 4.649660962221505,
      "grad_norm": 0.6407357454299927,
      "learning_rate": 0.0003214696692166792,
      "loss": 3.3983,
      "step": 43200
    },
    {
      "epoch": 4.655042514261113,
      "grad_norm": 0.7635655403137207,
      "learning_rate": 0.0003211464281866178,
      "loss": 3.4164,
      "step": 43250
    },
    {
      "epoch": 4.660424066300721,
      "grad_norm": 0.6684760451316833,
      "learning_rate": 0.00032082318715655634,
      "loss": 3.4197,
      "step": 43300
    },
    {
      "epoch": 4.665805618340329,
      "grad_norm": 0.673147976398468,
      "learning_rate": 0.000320499946126495,
      "loss": 3.392,
      "step": 43350
    },
    {
      "epoch": 4.6711871703799375,
      "grad_norm": 0.6477641463279724,
      "learning_rate": 0.0003201767050964336,
      "loss": 3.4001,
      "step": 43400
    },
    {
      "epoch": 4.676568722419546,
      "grad_norm": 0.6571009159088135,
      "learning_rate": 0.0003198534640663721,
      "loss": 3.393,
      "step": 43450
    },
    {
      "epoch": 4.681950274459154,
      "grad_norm": 0.7050467133522034,
      "learning_rate": 0.00031953668785691193,
      "loss": 3.4098,
      "step": 43500
    },
    {
      "epoch": 4.687331826498762,
      "grad_norm": 0.711652398109436,
      "learning_rate": 0.0003192134468268505,
      "loss": 3.4034,
      "step": 43550
    },
    {
      "epoch": 4.692713378538371,
      "grad_norm": 0.6692236661911011,
      "learning_rate": 0.0003188902057967891,
      "loss": 3.3815,
      "step": 43600
    },
    {
      "epoch": 4.6980949305779784,
      "grad_norm": 0.6892833709716797,
      "learning_rate": 0.00031856696476672766,
      "loss": 3.3971,
      "step": 43650
    },
    {
      "epoch": 4.703476482617587,
      "grad_norm": 0.6452851295471191,
      "learning_rate": 0.0003182437237366663,
      "loss": 3.3964,
      "step": 43700
    },
    {
      "epoch": 4.7088580346571955,
      "grad_norm": 0.683089017868042,
      "learning_rate": 0.0003179204827066049,
      "loss": 3.3904,
      "step": 43750
    },
    {
      "epoch": 4.714239586696803,
      "grad_norm": 0.6352059841156006,
      "learning_rate": 0.00031759724167654344,
      "loss": 3.37,
      "step": 43800
    },
    {
      "epoch": 4.719621138736412,
      "grad_norm": 0.6610687375068665,
      "learning_rate": 0.00031727400064648204,
      "loss": 3.398,
      "step": 43850
    },
    {
      "epoch": 4.725002690776019,
      "grad_norm": 0.615999162197113,
      "learning_rate": 0.0003169507596164206,
      "loss": 3.4015,
      "step": 43900
    },
    {
      "epoch": 4.730384242815628,
      "grad_norm": 0.6555367112159729,
      "learning_rate": 0.0003166275185863592,
      "loss": 3.3844,
      "step": 43950
    },
    {
      "epoch": 4.7357657948552365,
      "grad_norm": 0.6653710603713989,
      "learning_rate": 0.0003163042775562978,
      "loss": 3.4025,
      "step": 44000
    },
    {
      "epoch": 4.7357657948552365,
      "eval_accuracy": 0.3785725205329511,
      "eval_loss": 3.4251272678375244,
      "eval_runtime": 196.7837,
      "eval_samples_per_second": 91.527,
      "eval_steps_per_second": 5.722,
      "step": 44000
    },
    {
      "epoch": 4.741147346894844,
      "grad_norm": 0.6428149342536926,
      "learning_rate": 0.00031598103652623636,
      "loss": 3.4011,
      "step": 44050
    },
    {
      "epoch": 4.746528898934453,
      "grad_norm": 0.7112248539924622,
      "learning_rate": 0.00031565779549617496,
      "loss": 3.3996,
      "step": 44100
    },
    {
      "epoch": 4.751910450974061,
      "grad_norm": 0.6282157897949219,
      "learning_rate": 0.00031533455446611355,
      "loss": 3.3918,
      "step": 44150
    },
    {
      "epoch": 4.757292003013669,
      "grad_norm": 0.6741077899932861,
      "learning_rate": 0.0003150113134360521,
      "loss": 3.3868,
      "step": 44200
    },
    {
      "epoch": 4.7626735550532775,
      "grad_norm": 0.6872155666351318,
      "learning_rate": 0.0003146880724059907,
      "loss": 3.3892,
      "step": 44250
    },
    {
      "epoch": 4.768055107092886,
      "grad_norm": 0.6378316283226013,
      "learning_rate": 0.00031436483137592934,
      "loss": 3.4156,
      "step": 44300
    },
    {
      "epoch": 4.773436659132494,
      "grad_norm": 0.6506353616714478,
      "learning_rate": 0.0003140415903458679,
      "loss": 3.4003,
      "step": 44350
    },
    {
      "epoch": 4.778818211172102,
      "grad_norm": 0.6535550355911255,
      "learning_rate": 0.00031371834931580647,
      "loss": 3.3857,
      "step": 44400
    },
    {
      "epoch": 4.78419976321171,
      "grad_norm": 0.6900771260261536,
      "learning_rate": 0.000313395108285745,
      "loss": 3.407,
      "step": 44450
    },
    {
      "epoch": 4.7895813152513185,
      "grad_norm": 0.6276595592498779,
      "learning_rate": 0.0003130718672556836,
      "loss": 3.4002,
      "step": 44500
    },
    {
      "epoch": 4.794962867290927,
      "grad_norm": 0.6330270171165466,
      "learning_rate": 0.0003127486262256222,
      "loss": 3.3925,
      "step": 44550
    },
    {
      "epoch": 4.800344419330535,
      "grad_norm": 0.6442872881889343,
      "learning_rate": 0.0003124253851955608,
      "loss": 3.4055,
      "step": 44600
    },
    {
      "epoch": 4.805725971370143,
      "grad_norm": 0.6644704341888428,
      "learning_rate": 0.0003121021441654994,
      "loss": 3.3818,
      "step": 44650
    },
    {
      "epoch": 4.811107523409751,
      "grad_norm": 0.6481800079345703,
      "learning_rate": 0.000311778903135438,
      "loss": 3.3919,
      "step": 44700
    },
    {
      "epoch": 4.8164890754493594,
      "grad_norm": 0.6482080221176147,
      "learning_rate": 0.0003114556621053765,
      "loss": 3.3941,
      "step": 44750
    },
    {
      "epoch": 4.821870627488968,
      "grad_norm": 0.6653525829315186,
      "learning_rate": 0.0003111324210753151,
      "loss": 3.3933,
      "step": 44800
    },
    {
      "epoch": 4.827252179528576,
      "grad_norm": 0.6746211051940918,
      "learning_rate": 0.00031080918004525377,
      "loss": 3.3999,
      "step": 44850
    },
    {
      "epoch": 4.832633731568184,
      "grad_norm": 0.70958012342453,
      "learning_rate": 0.0003104859390151923,
      "loss": 3.3952,
      "step": 44900
    },
    {
      "epoch": 4.838015283607793,
      "grad_norm": 0.6676791906356812,
      "learning_rate": 0.0003101626979851309,
      "loss": 3.3887,
      "step": 44950
    },
    {
      "epoch": 4.8433968356474,
      "grad_norm": 0.6802822947502136,
      "learning_rate": 0.00030983945695506945,
      "loss": 3.4024,
      "step": 45000
    },
    {
      "epoch": 4.8433968356474,
      "eval_accuracy": 0.37923084838021903,
      "eval_loss": 3.4191226959228516,
      "eval_runtime": 205.1901,
      "eval_samples_per_second": 87.777,
      "eval_steps_per_second": 5.488,
      "step": 45000
    },
    {
      "epoch": 4.848778387687009,
      "grad_norm": 0.6934207677841187,
      "learning_rate": 0.00030951621592500804,
      "loss": 3.4028,
      "step": 45050
    },
    {
      "epoch": 4.8541599397266175,
      "grad_norm": 0.6621261239051819,
      "learning_rate": 0.00030919297489494663,
      "loss": 3.4001,
      "step": 45100
    },
    {
      "epoch": 4.859541491766225,
      "grad_norm": 0.6766626834869385,
      "learning_rate": 0.00030886973386488523,
      "loss": 3.3941,
      "step": 45150
    },
    {
      "epoch": 4.864923043805834,
      "grad_norm": 0.6346762180328369,
      "learning_rate": 0.0003085464928348238,
      "loss": 3.4054,
      "step": 45200
    },
    {
      "epoch": 4.870304595845441,
      "grad_norm": 0.7209916114807129,
      "learning_rate": 0.0003082232518047624,
      "loss": 3.4034,
      "step": 45250
    },
    {
      "epoch": 4.87568614788505,
      "grad_norm": 0.6274827718734741,
      "learning_rate": 0.00030790001077470096,
      "loss": 3.3991,
      "step": 45300
    },
    {
      "epoch": 4.8810676999246585,
      "grad_norm": 0.7232449650764465,
      "learning_rate": 0.00030757676974463955,
      "loss": 3.3938,
      "step": 45350
    },
    {
      "epoch": 4.886449251964266,
      "grad_norm": 0.7160061597824097,
      "learning_rate": 0.0003072535287145781,
      "loss": 3.3982,
      "step": 45400
    },
    {
      "epoch": 4.891830804003875,
      "grad_norm": 0.6695621013641357,
      "learning_rate": 0.00030693028768451674,
      "loss": 3.4003,
      "step": 45450
    },
    {
      "epoch": 4.897212356043483,
      "grad_norm": 0.7052890062332153,
      "learning_rate": 0.00030660704665445534,
      "loss": 3.4032,
      "step": 45500
    },
    {
      "epoch": 4.902593908083091,
      "grad_norm": 0.6913229823112488,
      "learning_rate": 0.00030629027044499514,
      "loss": 3.3954,
      "step": 45550
    },
    {
      "epoch": 4.9079754601226995,
      "grad_norm": 0.6375316381454468,
      "learning_rate": 0.00030596702941493374,
      "loss": 3.3993,
      "step": 45600
    },
    {
      "epoch": 4.913357012162308,
      "grad_norm": 0.6715528964996338,
      "learning_rate": 0.0003056437883848723,
      "loss": 3.4047,
      "step": 45650
    },
    {
      "epoch": 4.918738564201916,
      "grad_norm": 0.6626393795013428,
      "learning_rate": 0.0003053205473548109,
      "loss": 3.4155,
      "step": 45700
    },
    {
      "epoch": 4.924120116241524,
      "grad_norm": 0.6647288203239441,
      "learning_rate": 0.0003049973063247494,
      "loss": 3.4019,
      "step": 45750
    },
    {
      "epoch": 4.929501668281132,
      "grad_norm": 0.6812551021575928,
      "learning_rate": 0.00030467406529468806,
      "loss": 3.4047,
      "step": 45800
    },
    {
      "epoch": 4.9348832203207404,
      "grad_norm": 0.6454948782920837,
      "learning_rate": 0.00030435082426462666,
      "loss": 3.3905,
      "step": 45850
    },
    {
      "epoch": 4.940264772360349,
      "grad_norm": 0.6466948390007019,
      "learning_rate": 0.0003040275832345652,
      "loss": 3.3993,
      "step": 45900
    },
    {
      "epoch": 4.945646324399957,
      "grad_norm": 0.6519392728805542,
      "learning_rate": 0.0003037043422045038,
      "loss": 3.3947,
      "step": 45950
    },
    {
      "epoch": 4.951027876439565,
      "grad_norm": 0.6693304777145386,
      "learning_rate": 0.0003033811011744424,
      "loss": 3.3857,
      "step": 46000
    },
    {
      "epoch": 4.951027876439565,
      "eval_accuracy": 0.37971044222703215,
      "eval_loss": 3.4141387939453125,
      "eval_runtime": 191.8362,
      "eval_samples_per_second": 93.887,
      "eval_steps_per_second": 5.87,
      "step": 46000
    },
    {
      "epoch": 4.956409428479174,
      "grad_norm": 0.7195205688476562,
      "learning_rate": 0.00030305786014438093,
      "loss": 3.3972,
      "step": 46050
    },
    {
      "epoch": 4.961790980518781,
      "grad_norm": 0.6687961220741272,
      "learning_rate": 0.0003027346191143196,
      "loss": 3.3996,
      "step": 46100
    },
    {
      "epoch": 4.96717253255839,
      "grad_norm": 0.6921252012252808,
      "learning_rate": 0.00030241137808425817,
      "loss": 3.38,
      "step": 46150
    },
    {
      "epoch": 4.9725540845979985,
      "grad_norm": 0.7453935742378235,
      "learning_rate": 0.0003020881370541967,
      "loss": 3.4135,
      "step": 46200
    },
    {
      "epoch": 4.977935636637606,
      "grad_norm": 0.6929162740707397,
      "learning_rate": 0.0003017648960241353,
      "loss": 3.4043,
      "step": 46250
    },
    {
      "epoch": 4.983317188677215,
      "grad_norm": 0.6497056484222412,
      "learning_rate": 0.00030144165499407385,
      "loss": 3.3844,
      "step": 46300
    },
    {
      "epoch": 4.988698740716822,
      "grad_norm": 0.6445100903511047,
      "learning_rate": 0.00030111841396401244,
      "loss": 3.3977,
      "step": 46350
    },
    {
      "epoch": 4.994080292756431,
      "grad_norm": 0.6586089730262756,
      "learning_rate": 0.0003007951729339511,
      "loss": 3.3954,
      "step": 46400
    },
    {
      "epoch": 4.9994618447960395,
      "grad_norm": 0.6948351263999939,
      "learning_rate": 0.00030047193190388963,
      "loss": 3.4021,
      "step": 46450
    },
    {
      "epoch": 5.004843396835647,
      "grad_norm": 0.6887161135673523,
      "learning_rate": 0.0003001486908738282,
      "loss": 3.3117,
      "step": 46500
    },
    {
      "epoch": 5.010224948875256,
      "grad_norm": 0.6700156331062317,
      "learning_rate": 0.0002998254498437668,
      "loss": 3.315,
      "step": 46550
    },
    {
      "epoch": 5.015606500914864,
      "grad_norm": 0.7294987440109253,
      "learning_rate": 0.0002995022088137054,
      "loss": 3.3054,
      "step": 46600
    },
    {
      "epoch": 5.020988052954472,
      "grad_norm": 0.6779057383537292,
      "learning_rate": 0.00029917896778364396,
      "loss": 3.3014,
      "step": 46650
    },
    {
      "epoch": 5.0263696049940805,
      "grad_norm": 0.7391521334648132,
      "learning_rate": 0.00029885572675358255,
      "loss": 3.3047,
      "step": 46700
    },
    {
      "epoch": 5.031751157033688,
      "grad_norm": 0.6654561758041382,
      "learning_rate": 0.00029853248572352114,
      "loss": 3.3142,
      "step": 46750
    },
    {
      "epoch": 5.037132709073297,
      "grad_norm": 0.6845083236694336,
      "learning_rate": 0.00029820924469345974,
      "loss": 3.3233,
      "step": 46800
    },
    {
      "epoch": 5.042514261112905,
      "grad_norm": 0.6730548739433289,
      "learning_rate": 0.0002978860036633983,
      "loss": 3.2988,
      "step": 46850
    },
    {
      "epoch": 5.047895813152513,
      "grad_norm": 0.7016841769218445,
      "learning_rate": 0.00029756276263333693,
      "loss": 3.29,
      "step": 46900
    },
    {
      "epoch": 5.0532773651921215,
      "grad_norm": 0.70191490650177,
      "learning_rate": 0.00029723952160327547,
      "loss": 3.3019,
      "step": 46950
    },
    {
      "epoch": 5.05865891723173,
      "grad_norm": 0.6608459949493408,
      "learning_rate": 0.00029691628057321406,
      "loss": 3.3326,
      "step": 47000
    },
    {
      "epoch": 5.05865891723173,
      "eval_accuracy": 0.37991134141660116,
      "eval_loss": 3.419304609298706,
      "eval_runtime": 196.8607,
      "eval_samples_per_second": 91.491,
      "eval_steps_per_second": 5.72,
      "step": 47000
    },
    {
      "epoch": 5.064040469271338,
      "grad_norm": 0.6859248280525208,
      "learning_rate": 0.00029659303954315266,
      "loss": 3.3176,
      "step": 47050
    },
    {
      "epoch": 5.069422021310946,
      "grad_norm": 0.6877291202545166,
      "learning_rate": 0.00029626979851309125,
      "loss": 3.3087,
      "step": 47100
    },
    {
      "epoch": 5.074803573350554,
      "grad_norm": 0.6915717720985413,
      "learning_rate": 0.00029594655748302985,
      "loss": 3.3038,
      "step": 47150
    },
    {
      "epoch": 5.080185125390162,
      "grad_norm": 0.6920211911201477,
      "learning_rate": 0.0002956233164529684,
      "loss": 3.3085,
      "step": 47200
    },
    {
      "epoch": 5.085566677429771,
      "grad_norm": 0.6401144862174988,
      "learning_rate": 0.000295300075422907,
      "loss": 3.3258,
      "step": 47250
    },
    {
      "epoch": 5.090948229469379,
      "grad_norm": 0.6390669345855713,
      "learning_rate": 0.0002949768343928456,
      "loss": 3.3241,
      "step": 47300
    },
    {
      "epoch": 5.096329781508987,
      "grad_norm": 0.7305752038955688,
      "learning_rate": 0.00029465359336278417,
      "loss": 3.3095,
      "step": 47350
    },
    {
      "epoch": 5.101711333548596,
      "grad_norm": 0.7214723825454712,
      "learning_rate": 0.0002943303523327227,
      "loss": 3.3288,
      "step": 47400
    },
    {
      "epoch": 5.107092885588203,
      "grad_norm": 0.6987432241439819,
      "learning_rate": 0.00029400711130266136,
      "loss": 3.3194,
      "step": 47450
    },
    {
      "epoch": 5.112474437627812,
      "grad_norm": 0.6675676107406616,
      "learning_rate": 0.0002936838702725999,
      "loss": 3.3093,
      "step": 47500
    },
    {
      "epoch": 5.1178559896674205,
      "grad_norm": 0.6785469055175781,
      "learning_rate": 0.0002933606292425385,
      "loss": 3.3254,
      "step": 47550
    },
    {
      "epoch": 5.123237541707028,
      "grad_norm": 0.6891974210739136,
      "learning_rate": 0.0002930438530330783,
      "loss": 3.3209,
      "step": 47600
    },
    {
      "epoch": 5.128619093746637,
      "grad_norm": 0.7184481024742126,
      "learning_rate": 0.0002927206120030169,
      "loss": 3.3365,
      "step": 47650
    },
    {
      "epoch": 5.134000645786244,
      "grad_norm": 0.6881241798400879,
      "learning_rate": 0.0002923973709729555,
      "loss": 3.3414,
      "step": 47700
    },
    {
      "epoch": 5.139382197825853,
      "grad_norm": 0.666351318359375,
      "learning_rate": 0.00029207412994289403,
      "loss": 3.3402,
      "step": 47750
    },
    {
      "epoch": 5.1447637498654615,
      "grad_norm": 0.6807634234428406,
      "learning_rate": 0.0002917508889128327,
      "loss": 3.3304,
      "step": 47800
    },
    {
      "epoch": 5.150145301905069,
      "grad_norm": 0.7165495753288269,
      "learning_rate": 0.0002914276478827712,
      "loss": 3.3519,
      "step": 47850
    },
    {
      "epoch": 5.155526853944678,
      "grad_norm": 0.7941805124282837,
      "learning_rate": 0.0002911044068527098,
      "loss": 3.3313,
      "step": 47900
    },
    {
      "epoch": 5.160908405984286,
      "grad_norm": 0.744206428527832,
      "learning_rate": 0.0002907811658226484,
      "loss": 3.3305,
      "step": 47950
    },
    {
      "epoch": 5.166289958023894,
      "grad_norm": 0.7201210856437683,
      "learning_rate": 0.000290457924792587,
      "loss": 3.3355,
      "step": 48000
    },
    {
      "epoch": 5.166289958023894,
      "eval_accuracy": 0.38018818897475243,
      "eval_loss": 3.415891170501709,
      "eval_runtime": 187.2337,
      "eval_samples_per_second": 96.195,
      "eval_steps_per_second": 6.014,
      "step": 48000
    },
    {
      "epoch": 5.1716715100635025,
      "grad_norm": 0.7774996757507324,
      "learning_rate": 0.00029013468376252555,
      "loss": 3.313,
      "step": 48050
    },
    {
      "epoch": 5.17705306210311,
      "grad_norm": 1.036620855331421,
      "learning_rate": 0.00028981144273246414,
      "loss": 3.3172,
      "step": 48100
    },
    {
      "epoch": 5.182434614142719,
      "grad_norm": 0.6704376935958862,
      "learning_rate": 0.00028948820170240274,
      "loss": 3.331,
      "step": 48150
    },
    {
      "epoch": 5.187816166182327,
      "grad_norm": 0.6944007873535156,
      "learning_rate": 0.00028916496067234133,
      "loss": 3.3377,
      "step": 48200
    },
    {
      "epoch": 5.193197718221935,
      "grad_norm": 0.7034952640533447,
      "learning_rate": 0.0002888417196422799,
      "loss": 3.3392,
      "step": 48250
    },
    {
      "epoch": 5.198579270261543,
      "grad_norm": 0.709521472454071,
      "learning_rate": 0.00028851847861221847,
      "loss": 3.3437,
      "step": 48300
    },
    {
      "epoch": 5.203960822301152,
      "grad_norm": 0.6958163976669312,
      "learning_rate": 0.00028819523758215706,
      "loss": 3.3251,
      "step": 48350
    },
    {
      "epoch": 5.20934237434076,
      "grad_norm": 0.6938237547874451,
      "learning_rate": 0.00028787199655209566,
      "loss": 3.3402,
      "step": 48400
    },
    {
      "epoch": 5.214723926380368,
      "grad_norm": 0.7004581093788147,
      "learning_rate": 0.00028754875552203425,
      "loss": 3.3431,
      "step": 48450
    },
    {
      "epoch": 5.220105478419977,
      "grad_norm": 0.7187439799308777,
      "learning_rate": 0.0002872255144919728,
      "loss": 3.3325,
      "step": 48500
    },
    {
      "epoch": 5.225487030459584,
      "grad_norm": 0.7515068650245667,
      "learning_rate": 0.00028690227346191144,
      "loss": 3.3313,
      "step": 48550
    },
    {
      "epoch": 5.230868582499193,
      "grad_norm": 0.6710924506187439,
      "learning_rate": 0.00028657903243185,
      "loss": 3.3406,
      "step": 48600
    },
    {
      "epoch": 5.236250134538801,
      "grad_norm": 0.7324944138526917,
      "learning_rate": 0.0002862557914017886,
      "loss": 3.3281,
      "step": 48650
    },
    {
      "epoch": 5.241631686578409,
      "grad_norm": 0.7170085906982422,
      "learning_rate": 0.00028593255037172717,
      "loss": 3.3183,
      "step": 48700
    },
    {
      "epoch": 5.247013238618018,
      "grad_norm": 0.69121253490448,
      "learning_rate": 0.00028560930934166576,
      "loss": 3.3427,
      "step": 48750
    },
    {
      "epoch": 5.252394790657625,
      "grad_norm": 0.7113257050514221,
      "learning_rate": 0.00028528606831160436,
      "loss": 3.3261,
      "step": 48800
    },
    {
      "epoch": 5.257776342697234,
      "grad_norm": 0.6930692195892334,
      "learning_rate": 0.0002849628272815429,
      "loss": 3.33,
      "step": 48850
    },
    {
      "epoch": 5.2631578947368425,
      "grad_norm": 0.6934365630149841,
      "learning_rate": 0.00028464605107208276,
      "loss": 3.3259,
      "step": 48900
    },
    {
      "epoch": 5.26853944677645,
      "grad_norm": 0.7453590631484985,
      "learning_rate": 0.0002843228100420213,
      "loss": 3.3446,
      "step": 48950
    },
    {
      "epoch": 5.273920998816059,
      "grad_norm": 0.6774255037307739,
      "learning_rate": 0.0002839995690119599,
      "loss": 3.3281,
      "step": 49000
    },
    {
      "epoch": 5.273920998816059,
      "eval_accuracy": 0.38062334379044943,
      "eval_loss": 3.4131619930267334,
      "eval_runtime": 198.2275,
      "eval_samples_per_second": 90.86,
      "eval_steps_per_second": 5.68,
      "step": 49000
    },
    {
      "epoch": 5.279302550855666,
      "grad_norm": 0.7351974248886108,
      "learning_rate": 0.0002836763279818985,
      "loss": 3.3381,
      "step": 49050
    },
    {
      "epoch": 5.284684102895275,
      "grad_norm": 0.6695067286491394,
      "learning_rate": 0.0002833530869518371,
      "loss": 3.3311,
      "step": 49100
    },
    {
      "epoch": 5.2900656549348835,
      "grad_norm": 0.694674015045166,
      "learning_rate": 0.0002830298459217756,
      "loss": 3.3274,
      "step": 49150
    },
    {
      "epoch": 5.295447206974491,
      "grad_norm": 0.6997494101524353,
      "learning_rate": 0.0002827066048917142,
      "loss": 3.3367,
      "step": 49200
    },
    {
      "epoch": 5.3008287590141,
      "grad_norm": 0.6941072344779968,
      "learning_rate": 0.0002823833638616528,
      "loss": 3.3414,
      "step": 49250
    },
    {
      "epoch": 5.306210311053708,
      "grad_norm": 0.7511026263237,
      "learning_rate": 0.0002820601228315914,
      "loss": 3.3472,
      "step": 49300
    },
    {
      "epoch": 5.311591863093316,
      "grad_norm": 0.7501461505889893,
      "learning_rate": 0.00028173688180153,
      "loss": 3.3496,
      "step": 49350
    },
    {
      "epoch": 5.316973415132924,
      "grad_norm": 0.7130686044692993,
      "learning_rate": 0.00028141364077146854,
      "loss": 3.3316,
      "step": 49400
    },
    {
      "epoch": 5.322354967172533,
      "grad_norm": 0.757497251033783,
      "learning_rate": 0.00028109039974140714,
      "loss": 3.3416,
      "step": 49450
    },
    {
      "epoch": 5.327736519212141,
      "grad_norm": 0.6970386505126953,
      "learning_rate": 0.00028076715871134573,
      "loss": 3.326,
      "step": 49500
    },
    {
      "epoch": 5.333118071251749,
      "grad_norm": 0.6531247496604919,
      "learning_rate": 0.00028044391768128433,
      "loss": 3.3432,
      "step": 49550
    },
    {
      "epoch": 5.338499623291357,
      "grad_norm": 0.6707568168640137,
      "learning_rate": 0.0002801206766512229,
      "loss": 3.3348,
      "step": 49600
    },
    {
      "epoch": 5.343881175330965,
      "grad_norm": 0.6901975274085999,
      "learning_rate": 0.0002797974356211615,
      "loss": 3.3318,
      "step": 49650
    },
    {
      "epoch": 5.349262727370574,
      "grad_norm": 0.700663685798645,
      "learning_rate": 0.00027947419459110006,
      "loss": 3.3511,
      "step": 49700
    },
    {
      "epoch": 5.354644279410182,
      "grad_norm": 0.7497619986534119,
      "learning_rate": 0.00027915095356103865,
      "loss": 3.3416,
      "step": 49750
    },
    {
      "epoch": 5.36002583144979,
      "grad_norm": 0.7329772114753723,
      "learning_rate": 0.00027882771253097725,
      "loss": 3.3256,
      "step": 49800
    },
    {
      "epoch": 5.365407383489399,
      "grad_norm": 0.6941583156585693,
      "learning_rate": 0.00027850447150091584,
      "loss": 3.3327,
      "step": 49850
    },
    {
      "epoch": 5.370788935529006,
      "grad_norm": 0.6678893566131592,
      "learning_rate": 0.00027818123047085444,
      "loss": 3.3587,
      "step": 49900
    },
    {
      "epoch": 5.376170487568615,
      "grad_norm": 0.7521560192108154,
      "learning_rate": 0.000277857989440793,
      "loss": 3.3422,
      "step": 49950
    },
    {
      "epoch": 5.3815520396082235,
      "grad_norm": 0.7073588371276855,
      "learning_rate": 0.00027753474841073157,
      "loss": 3.3539,
      "step": 50000
    },
    {
      "epoch": 5.3815520396082235,
      "eval_accuracy": 0.38081207385658317,
      "eval_loss": 3.408217191696167,
      "eval_runtime": 190.7868,
      "eval_samples_per_second": 94.404,
      "eval_steps_per_second": 5.902,
      "step": 50000
    },
    {
      "epoch": 5.386933591647831,
      "grad_norm": 0.7438055276870728,
      "learning_rate": 0.00027721150738067017,
      "loss": 3.3488,
      "step": 50050
    },
    {
      "epoch": 5.39231514368744,
      "grad_norm": 0.7055433988571167,
      "learning_rate": 0.00027688826635060876,
      "loss": 3.3172,
      "step": 50100
    },
    {
      "epoch": 5.397696695727047,
      "grad_norm": 0.7161540389060974,
      "learning_rate": 0.0002765650253205473,
      "loss": 3.344,
      "step": 50150
    },
    {
      "epoch": 5.403078247766656,
      "grad_norm": 0.6727215051651001,
      "learning_rate": 0.00027624178429048595,
      "loss": 3.3333,
      "step": 50200
    },
    {
      "epoch": 5.4084597998062645,
      "grad_norm": 0.7575907707214355,
      "learning_rate": 0.0002759185432604245,
      "loss": 3.3384,
      "step": 50250
    },
    {
      "epoch": 5.413841351845872,
      "grad_norm": 0.7084335684776306,
      "learning_rate": 0.0002755953022303631,
      "loss": 3.333,
      "step": 50300
    },
    {
      "epoch": 5.419222903885481,
      "grad_norm": 0.7179964780807495,
      "learning_rate": 0.0002752720612003017,
      "loss": 3.3441,
      "step": 50350
    },
    {
      "epoch": 5.424604455925088,
      "grad_norm": 0.7183540463447571,
      "learning_rate": 0.0002749488201702403,
      "loss": 3.3433,
      "step": 50400
    },
    {
      "epoch": 5.429986007964697,
      "grad_norm": 0.7511764168739319,
      "learning_rate": 0.0002746255791401788,
      "loss": 3.3451,
      "step": 50450
    },
    {
      "epoch": 5.435367560004305,
      "grad_norm": 0.7309442162513733,
      "learning_rate": 0.0002743023381101174,
      "loss": 3.3431,
      "step": 50500
    },
    {
      "epoch": 5.440749112043913,
      "grad_norm": 0.7652359008789062,
      "learning_rate": 0.000273979097080056,
      "loss": 3.3379,
      "step": 50550
    },
    {
      "epoch": 5.446130664083522,
      "grad_norm": 0.7862690687179565,
      "learning_rate": 0.0002736558560499946,
      "loss": 3.3495,
      "step": 50600
    },
    {
      "epoch": 5.45151221612313,
      "grad_norm": 0.7119584679603577,
      "learning_rate": 0.0002733326150199332,
      "loss": 3.3392,
      "step": 50650
    },
    {
      "epoch": 5.456893768162738,
      "grad_norm": 0.7276889085769653,
      "learning_rate": 0.00027300937398987173,
      "loss": 3.3436,
      "step": 50700
    },
    {
      "epoch": 5.462275320202346,
      "grad_norm": 0.7200219035148621,
      "learning_rate": 0.0002726861329598104,
      "loss": 3.3449,
      "step": 50750
    },
    {
      "epoch": 5.467656872241955,
      "grad_norm": 0.7543548941612244,
      "learning_rate": 0.0002723628919297489,
      "loss": 3.3471,
      "step": 50800
    },
    {
      "epoch": 5.473038424281563,
      "grad_norm": 0.7325814366340637,
      "learning_rate": 0.0002720396508996875,
      "loss": 3.3482,
      "step": 50850
    },
    {
      "epoch": 5.478419976321171,
      "grad_norm": 0.7134089469909668,
      "learning_rate": 0.0002717164098696261,
      "loss": 3.3414,
      "step": 50900
    },
    {
      "epoch": 5.483801528360779,
      "grad_norm": 0.7114308476448059,
      "learning_rate": 0.00027139316883956465,
      "loss": 3.3497,
      "step": 50950
    },
    {
      "epoch": 5.489183080400387,
      "grad_norm": 0.7253952026367188,
      "learning_rate": 0.00027106992780950325,
      "loss": 3.3443,
      "step": 51000
    },
    {
      "epoch": 5.489183080400387,
      "eval_accuracy": 0.3813965177401373,
      "eval_loss": 3.4036450386047363,
      "eval_runtime": 190.4642,
      "eval_samples_per_second": 94.564,
      "eval_steps_per_second": 5.912,
      "step": 51000
    },
    {
      "epoch": 5.494564632439996,
      "grad_norm": 0.7643466591835022,
      "learning_rate": 0.00027074668677944184,
      "loss": 3.3332,
      "step": 51050
    },
    {
      "epoch": 5.499946184479604,
      "grad_norm": 0.6971955299377441,
      "learning_rate": 0.00027042344574938044,
      "loss": 3.3611,
      "step": 51100
    },
    {
      "epoch": 5.505327736519212,
      "grad_norm": 0.7099331617355347,
      "learning_rate": 0.000270100204719319,
      "loss": 3.3225,
      "step": 51150
    },
    {
      "epoch": 5.510709288558821,
      "grad_norm": 0.711465060710907,
      "learning_rate": 0.0002697769636892576,
      "loss": 3.3339,
      "step": 51200
    },
    {
      "epoch": 5.516090840598428,
      "grad_norm": 0.7545268535614014,
      "learning_rate": 0.00026945372265919617,
      "loss": 3.3341,
      "step": 51250
    },
    {
      "epoch": 5.521472392638037,
      "grad_norm": 0.7730751037597656,
      "learning_rate": 0.00026913048162913476,
      "loss": 3.3339,
      "step": 51300
    },
    {
      "epoch": 5.5268539446776455,
      "grad_norm": 0.7452085614204407,
      "learning_rate": 0.00026880724059907336,
      "loss": 3.3474,
      "step": 51350
    },
    {
      "epoch": 5.532235496717253,
      "grad_norm": 0.7253414392471313,
      "learning_rate": 0.00026848399956901195,
      "loss": 3.3399,
      "step": 51400
    },
    {
      "epoch": 5.537617048756862,
      "grad_norm": 0.7186622023582458,
      "learning_rate": 0.0002681607585389505,
      "loss": 3.3568,
      "step": 51450
    },
    {
      "epoch": 5.542998600796469,
      "grad_norm": 0.697465717792511,
      "learning_rate": 0.0002678375175088891,
      "loss": 3.3473,
      "step": 51500
    },
    {
      "epoch": 5.548380152836078,
      "grad_norm": 0.7634124159812927,
      "learning_rate": 0.0002675142764788277,
      "loss": 3.3487,
      "step": 51550
    },
    {
      "epoch": 5.553761704875686,
      "grad_norm": 0.7234585285186768,
      "learning_rate": 0.0002671910354487663,
      "loss": 3.3414,
      "step": 51600
    },
    {
      "epoch": 5.559143256915294,
      "grad_norm": 0.7003602981567383,
      "learning_rate": 0.00026686779441870487,
      "loss": 3.3439,
      "step": 51650
    },
    {
      "epoch": 5.564524808954903,
      "grad_norm": 0.6854865550994873,
      "learning_rate": 0.0002665445533886434,
      "loss": 3.3688,
      "step": 51700
    },
    {
      "epoch": 5.569906360994511,
      "grad_norm": 0.7288345098495483,
      "learning_rate": 0.00026622131235858206,
      "loss": 3.3356,
      "step": 51750
    },
    {
      "epoch": 5.575287913034119,
      "grad_norm": 0.7311908006668091,
      "learning_rate": 0.0002659045361491218,
      "loss": 3.3338,
      "step": 51800
    },
    {
      "epoch": 5.580669465073727,
      "grad_norm": 0.7358433604240417,
      "learning_rate": 0.00026558129511906046,
      "loss": 3.3398,
      "step": 51850
    },
    {
      "epoch": 5.586051017113336,
      "grad_norm": 0.7444213032722473,
      "learning_rate": 0.000265258054088999,
      "loss": 3.3355,
      "step": 51900
    },
    {
      "epoch": 5.591432569152944,
      "grad_norm": 0.7090075612068176,
      "learning_rate": 0.0002649348130589376,
      "loss": 3.3651,
      "step": 51950
    },
    {
      "epoch": 5.596814121192552,
      "grad_norm": 0.7260854840278625,
      "learning_rate": 0.0002646115720288762,
      "loss": 3.3474,
      "step": 52000
    },
    {
      "epoch": 5.596814121192552,
      "eval_accuracy": 0.38173562340300626,
      "eval_loss": 3.4010379314422607,
      "eval_runtime": 198.0447,
      "eval_samples_per_second": 90.944,
      "eval_steps_per_second": 5.686,
      "step": 52000
    },
    {
      "epoch": 5.60219567323216,
      "grad_norm": 0.7559796571731567,
      "learning_rate": 0.00026428833099881473,
      "loss": 3.3582,
      "step": 52050
    },
    {
      "epoch": 5.607577225271768,
      "grad_norm": 0.7076489925384521,
      "learning_rate": 0.0002639650899687533,
      "loss": 3.3544,
      "step": 52100
    },
    {
      "epoch": 5.612958777311377,
      "grad_norm": 0.7647249698638916,
      "learning_rate": 0.0002636418489386919,
      "loss": 3.3548,
      "step": 52150
    },
    {
      "epoch": 5.618340329350985,
      "grad_norm": 0.7382960915565491,
      "learning_rate": 0.0002633186079086305,
      "loss": 3.3521,
      "step": 52200
    },
    {
      "epoch": 5.623721881390593,
      "grad_norm": 0.7489807605743408,
      "learning_rate": 0.00026299536687856905,
      "loss": 3.3359,
      "step": 52250
    },
    {
      "epoch": 5.629103433430201,
      "grad_norm": 0.7721214294433594,
      "learning_rate": 0.0002626721258485077,
      "loss": 3.3407,
      "step": 52300
    },
    {
      "epoch": 5.634484985469809,
      "grad_norm": 0.715048611164093,
      "learning_rate": 0.00026234888481844624,
      "loss": 3.3591,
      "step": 52350
    },
    {
      "epoch": 5.639866537509418,
      "grad_norm": 0.7328721880912781,
      "learning_rate": 0.00026202564378838484,
      "loss": 3.3324,
      "step": 52400
    },
    {
      "epoch": 5.645248089549026,
      "grad_norm": 0.8084849119186401,
      "learning_rate": 0.00026170240275832343,
      "loss": 3.3537,
      "step": 52450
    },
    {
      "epoch": 5.650629641588634,
      "grad_norm": 0.687576413154602,
      "learning_rate": 0.00026137916172826203,
      "loss": 3.3392,
      "step": 52500
    },
    {
      "epoch": 5.656011193628243,
      "grad_norm": 0.7627485990524292,
      "learning_rate": 0.0002610559206982006,
      "loss": 3.3461,
      "step": 52550
    },
    {
      "epoch": 5.66139274566785,
      "grad_norm": 0.7329762578010559,
      "learning_rate": 0.00026073267966813916,
      "loss": 3.357,
      "step": 52600
    },
    {
      "epoch": 5.666774297707459,
      "grad_norm": 0.7457464933395386,
      "learning_rate": 0.00026040943863807776,
      "loss": 3.3515,
      "step": 52650
    },
    {
      "epoch": 5.672155849747067,
      "grad_norm": 0.7168912291526794,
      "learning_rate": 0.00026008619760801635,
      "loss": 3.3289,
      "step": 52700
    },
    {
      "epoch": 5.677537401786675,
      "grad_norm": 0.7807959318161011,
      "learning_rate": 0.00025976295657795495,
      "loss": 3.3489,
      "step": 52750
    },
    {
      "epoch": 5.682918953826284,
      "grad_norm": 0.7521499395370483,
      "learning_rate": 0.0002594397155478935,
      "loss": 3.3381,
      "step": 52800
    },
    {
      "epoch": 5.688300505865891,
      "grad_norm": 0.734977662563324,
      "learning_rate": 0.00025911647451783214,
      "loss": 3.3315,
      "step": 52850
    },
    {
      "epoch": 5.6936820579055,
      "grad_norm": 0.7627432942390442,
      "learning_rate": 0.0002587932334877707,
      "loss": 3.3351,
      "step": 52900
    },
    {
      "epoch": 5.699063609945108,
      "grad_norm": 0.7547647953033447,
      "learning_rate": 0.00025846999245770927,
      "loss": 3.3308,
      "step": 52950
    },
    {
      "epoch": 5.704445161984716,
      "grad_norm": 0.750952422618866,
      "learning_rate": 0.00025814675142764787,
      "loss": 3.3324,
      "step": 53000
    },
    {
      "epoch": 5.704445161984716,
      "eval_accuracy": 0.382410466489222,
      "eval_loss": 3.3959882259368896,
      "eval_runtime": 208.0153,
      "eval_samples_per_second": 86.585,
      "eval_steps_per_second": 5.413,
      "step": 53000
    },
    {
      "epoch": 5.709826714024325,
      "grad_norm": 0.7652902603149414,
      "learning_rate": 0.00025782351039758646,
      "loss": 3.3478,
      "step": 53050
    },
    {
      "epoch": 5.715208266063933,
      "grad_norm": 0.6935646533966064,
      "learning_rate": 0.000257500269367525,
      "loss": 3.3445,
      "step": 53100
    },
    {
      "epoch": 5.720589818103541,
      "grad_norm": 0.776350200176239,
      "learning_rate": 0.0002571770283374636,
      "loss": 3.3534,
      "step": 53150
    },
    {
      "epoch": 5.725971370143149,
      "grad_norm": 0.677318274974823,
      "learning_rate": 0.0002568537873074022,
      "loss": 3.3401,
      "step": 53200
    },
    {
      "epoch": 5.731352922182758,
      "grad_norm": 0.7982649803161621,
      "learning_rate": 0.0002565305462773408,
      "loss": 3.3499,
      "step": 53250
    },
    {
      "epoch": 5.736734474222366,
      "grad_norm": 0.6893640756607056,
      "learning_rate": 0.0002562073052472794,
      "loss": 3.3459,
      "step": 53300
    },
    {
      "epoch": 5.742116026261974,
      "grad_norm": 0.7793111205101013,
      "learning_rate": 0.0002558840642172179,
      "loss": 3.344,
      "step": 53350
    },
    {
      "epoch": 5.747497578301582,
      "grad_norm": 0.7277651429176331,
      "learning_rate": 0.00025556082318715657,
      "loss": 3.3437,
      "step": 53400
    },
    {
      "epoch": 5.75287913034119,
      "grad_norm": 0.7123382091522217,
      "learning_rate": 0.0002552375821570951,
      "loss": 3.3577,
      "step": 53450
    },
    {
      "epoch": 5.758260682380799,
      "grad_norm": 0.6870895624160767,
      "learning_rate": 0.0002549143411270337,
      "loss": 3.3442,
      "step": 53500
    },
    {
      "epoch": 5.763642234420407,
      "grad_norm": 0.7414118051528931,
      "learning_rate": 0.0002545911000969723,
      "loss": 3.3424,
      "step": 53550
    },
    {
      "epoch": 5.769023786460015,
      "grad_norm": 0.7082991600036621,
      "learning_rate": 0.0002542678590669109,
      "loss": 3.3339,
      "step": 53600
    },
    {
      "epoch": 5.774405338499624,
      "grad_norm": 0.7212987542152405,
      "learning_rate": 0.00025394461803684943,
      "loss": 3.3527,
      "step": 53650
    },
    {
      "epoch": 5.779786890539231,
      "grad_norm": 0.8128365278244019,
      "learning_rate": 0.00025362137700678803,
      "loss": 3.342,
      "step": 53700
    },
    {
      "epoch": 5.78516844257884,
      "grad_norm": 0.8344599008560181,
      "learning_rate": 0.0002532981359767266,
      "loss": 3.3232,
      "step": 53750
    },
    {
      "epoch": 5.790549994618448,
      "grad_norm": 0.7615811228752136,
      "learning_rate": 0.0002529748949466652,
      "loss": 3.3454,
      "step": 53800
    },
    {
      "epoch": 5.795931546658056,
      "grad_norm": 0.698824405670166,
      "learning_rate": 0.0002526516539166038,
      "loss": 3.3487,
      "step": 53850
    },
    {
      "epoch": 5.801313098697665,
      "grad_norm": 0.7221081852912903,
      "learning_rate": 0.00025232841288654235,
      "loss": 3.3569,
      "step": 53900
    },
    {
      "epoch": 5.806694650737272,
      "grad_norm": 0.724270224571228,
      "learning_rate": 0.00025200517185648095,
      "loss": 3.335,
      "step": 53950
    },
    {
      "epoch": 5.812076202776881,
      "grad_norm": 0.7062095999717712,
      "learning_rate": 0.00025168193082641954,
      "loss": 3.3392,
      "step": 54000
    },
    {
      "epoch": 5.812076202776881,
      "eval_accuracy": 0.3827846670348551,
      "eval_loss": 3.3915441036224365,
      "eval_runtime": 186.351,
      "eval_samples_per_second": 96.651,
      "eval_steps_per_second": 6.042,
      "step": 54000
    },
    {
      "epoch": 5.817457754816489,
      "grad_norm": 0.7363617420196533,
      "learning_rate": 0.00025135868979635814,
      "loss": 3.3573,
      "step": 54050
    },
    {
      "epoch": 5.822839306856097,
      "grad_norm": 0.7413693070411682,
      "learning_rate": 0.0002510354487662967,
      "loss": 3.3362,
      "step": 54100
    },
    {
      "epoch": 5.828220858895706,
      "grad_norm": 0.7135999798774719,
      "learning_rate": 0.00025071867255683654,
      "loss": 3.3498,
      "step": 54150
    },
    {
      "epoch": 5.833602410935313,
      "grad_norm": 0.9513452649116516,
      "learning_rate": 0.00025039543152677513,
      "loss": 3.3495,
      "step": 54200
    },
    {
      "epoch": 5.838983962974922,
      "grad_norm": 0.764833927154541,
      "learning_rate": 0.0002500721904967137,
      "loss": 3.3478,
      "step": 54250
    },
    {
      "epoch": 5.84436551501453,
      "grad_norm": 0.7110487222671509,
      "learning_rate": 0.00024974894946665227,
      "loss": 3.3415,
      "step": 54300
    },
    {
      "epoch": 5.849747067054138,
      "grad_norm": 0.857364296913147,
      "learning_rate": 0.00024942570843659086,
      "loss": 3.3493,
      "step": 54350
    },
    {
      "epoch": 5.855128619093747,
      "grad_norm": 0.7012077569961548,
      "learning_rate": 0.00024910246740652946,
      "loss": 3.3413,
      "step": 54400
    },
    {
      "epoch": 5.860510171133355,
      "grad_norm": 0.727841317653656,
      "learning_rate": 0.000248779226376468,
      "loss": 3.3412,
      "step": 54450
    },
    {
      "epoch": 5.865891723172963,
      "grad_norm": 0.7530490159988403,
      "learning_rate": 0.00024845598534640665,
      "loss": 3.3456,
      "step": 54500
    },
    {
      "epoch": 5.871273275212571,
      "grad_norm": 0.7999284267425537,
      "learning_rate": 0.0002481327443163452,
      "loss": 3.3584,
      "step": 54550
    },
    {
      "epoch": 5.87665482725218,
      "grad_norm": 0.7562769055366516,
      "learning_rate": 0.0002478095032862838,
      "loss": 3.3568,
      "step": 54600
    },
    {
      "epoch": 5.882036379291788,
      "grad_norm": 0.7588623762130737,
      "learning_rate": 0.0002474862622562224,
      "loss": 3.3355,
      "step": 54650
    },
    {
      "epoch": 5.887417931331396,
      "grad_norm": 0.7253016829490662,
      "learning_rate": 0.00024716302122616097,
      "loss": 3.3541,
      "step": 54700
    },
    {
      "epoch": 5.892799483371004,
      "grad_norm": 0.6772181987762451,
      "learning_rate": 0.0002468397801960995,
      "loss": 3.3558,
      "step": 54750
    },
    {
      "epoch": 5.898181035410612,
      "grad_norm": 0.7588209509849548,
      "learning_rate": 0.0002465165391660381,
      "loss": 3.3446,
      "step": 54800
    },
    {
      "epoch": 5.903562587450221,
      "grad_norm": 0.7439490556716919,
      "learning_rate": 0.0002461932981359767,
      "loss": 3.3588,
      "step": 54850
    },
    {
      "epoch": 5.9089441394898286,
      "grad_norm": 0.773601770401001,
      "learning_rate": 0.0002458700571059153,
      "loss": 3.357,
      "step": 54900
    },
    {
      "epoch": 5.914325691529437,
      "grad_norm": 0.7919325828552246,
      "learning_rate": 0.0002455468160758539,
      "loss": 3.3206,
      "step": 54950
    },
    {
      "epoch": 5.919707243569046,
      "grad_norm": 0.7080097794532776,
      "learning_rate": 0.00024522357504579243,
      "loss": 3.3439,
      "step": 55000
    },
    {
      "epoch": 5.919707243569046,
      "eval_accuracy": 0.38304847624646865,
      "eval_loss": 3.386930227279663,
      "eval_runtime": 216.5783,
      "eval_samples_per_second": 83.162,
      "eval_steps_per_second": 5.199,
      "step": 55000
    },
    {
      "epoch": 5.925088795608653,
      "grad_norm": 0.7515069842338562,
      "learning_rate": 0.000244900334015731,
      "loss": 3.3414,
      "step": 55050
    },
    {
      "epoch": 5.930470347648262,
      "grad_norm": 0.7621107697486877,
      "learning_rate": 0.0002445770929856696,
      "loss": 3.3442,
      "step": 55100
    },
    {
      "epoch": 5.93585189968787,
      "grad_norm": 0.7559869289398193,
      "learning_rate": 0.0002442538519556082,
      "loss": 3.333,
      "step": 55150
    },
    {
      "epoch": 5.941233451727478,
      "grad_norm": 0.7848432064056396,
      "learning_rate": 0.0002439306109255468,
      "loss": 3.345,
      "step": 55200
    },
    {
      "epoch": 5.946615003767087,
      "grad_norm": 0.7558146119117737,
      "learning_rate": 0.00024360736989548538,
      "loss": 3.3435,
      "step": 55250
    },
    {
      "epoch": 5.951996555806694,
      "grad_norm": 0.7186861038208008,
      "learning_rate": 0.00024328412886542394,
      "loss": 3.3485,
      "step": 55300
    },
    {
      "epoch": 5.957378107846303,
      "grad_norm": 0.770593523979187,
      "learning_rate": 0.00024296088783536257,
      "loss": 3.3395,
      "step": 55350
    },
    {
      "epoch": 5.962759659885911,
      "grad_norm": 0.8355637788772583,
      "learning_rate": 0.00024263764680530113,
      "loss": 3.3321,
      "step": 55400
    },
    {
      "epoch": 5.968141211925519,
      "grad_norm": 0.7618758678436279,
      "learning_rate": 0.0002423144057752397,
      "loss": 3.3341,
      "step": 55450
    },
    {
      "epoch": 5.973522763965128,
      "grad_norm": 0.7331770062446594,
      "learning_rate": 0.00024199116474517832,
      "loss": 3.3438,
      "step": 55500
    },
    {
      "epoch": 5.978904316004736,
      "grad_norm": 0.764519453048706,
      "learning_rate": 0.0002416679237151169,
      "loss": 3.3471,
      "step": 55550
    },
    {
      "epoch": 5.984285868044344,
      "grad_norm": 0.7753000855445862,
      "learning_rate": 0.00024134468268505546,
      "loss": 3.3175,
      "step": 55600
    },
    {
      "epoch": 5.989667420083952,
      "grad_norm": 0.7588565349578857,
      "learning_rate": 0.00024102144165499405,
      "loss": 3.3349,
      "step": 55650
    },
    {
      "epoch": 5.995048972123561,
      "grad_norm": 0.7396625280380249,
      "learning_rate": 0.00024069820062493265,
      "loss": 3.3549,
      "step": 55700
    },
    {
      "epoch": 6.000430524163169,
      "grad_norm": 0.7639598250389099,
      "learning_rate": 0.00024037495959487121,
      "loss": 3.3327,
      "step": 55750
    },
    {
      "epoch": 6.005812076202777,
      "grad_norm": 0.7728949785232544,
      "learning_rate": 0.0002400517185648098,
      "loss": 3.2475,
      "step": 55800
    },
    {
      "epoch": 6.011193628242385,
      "grad_norm": 0.7817671895027161,
      "learning_rate": 0.00023972847753474838,
      "loss": 3.2563,
      "step": 55850
    },
    {
      "epoch": 6.016575180281993,
      "grad_norm": 0.7547463178634644,
      "learning_rate": 0.00023940523650468697,
      "loss": 3.2627,
      "step": 55900
    },
    {
      "epoch": 6.021956732321602,
      "grad_norm": 0.7595130205154419,
      "learning_rate": 0.00023908199547462557,
      "loss": 3.2555,
      "step": 55950
    },
    {
      "epoch": 6.0273382843612096,
      "grad_norm": 0.7184428572654724,
      "learning_rate": 0.00023875875444456413,
      "loss": 3.2672,
      "step": 56000
    },
    {
      "epoch": 6.0273382843612096,
      "eval_accuracy": 0.3834293046182584,
      "eval_loss": 3.389963150024414,
      "eval_runtime": 188.2488,
      "eval_samples_per_second": 95.677,
      "eval_steps_per_second": 5.981,
      "step": 56000
    },
    {
      "epoch": 6.032719836400818,
      "grad_norm": 0.7633922696113586,
      "learning_rate": 0.0002384355134145027,
      "loss": 3.2507,
      "step": 56050
    },
    {
      "epoch": 6.038101388440427,
      "grad_norm": 0.7777019739151001,
      "learning_rate": 0.00023811227238444132,
      "loss": 3.275,
      "step": 56100
    },
    {
      "epoch": 6.043482940480034,
      "grad_norm": 0.7740806341171265,
      "learning_rate": 0.00023779549617498113,
      "loss": 3.2677,
      "step": 56150
    },
    {
      "epoch": 6.048864492519643,
      "grad_norm": 0.7871271371841431,
      "learning_rate": 0.0002374722551449197,
      "loss": 3.2406,
      "step": 56200
    },
    {
      "epoch": 6.0542460445592505,
      "grad_norm": 0.8018182516098022,
      "learning_rate": 0.0002371490141148583,
      "loss": 3.2704,
      "step": 56250
    },
    {
      "epoch": 6.059627596598859,
      "grad_norm": 0.7371104955673218,
      "learning_rate": 0.00023682577308479689,
      "loss": 3.2754,
      "step": 56300
    },
    {
      "epoch": 6.065009148638468,
      "grad_norm": 0.7586576342582703,
      "learning_rate": 0.00023650253205473545,
      "loss": 3.2574,
      "step": 56350
    },
    {
      "epoch": 6.070390700678075,
      "grad_norm": 0.7714259624481201,
      "learning_rate": 0.00023617929102467402,
      "loss": 3.2755,
      "step": 56400
    },
    {
      "epoch": 6.075772252717684,
      "grad_norm": 0.7666574716567993,
      "learning_rate": 0.00023585604999461264,
      "loss": 3.2716,
      "step": 56450
    },
    {
      "epoch": 6.081153804757292,
      "grad_norm": 0.7932227849960327,
      "learning_rate": 0.0002355328089645512,
      "loss": 3.249,
      "step": 56500
    },
    {
      "epoch": 6.0865353567969,
      "grad_norm": 0.7671225070953369,
      "learning_rate": 0.00023520956793448978,
      "loss": 3.2637,
      "step": 56550
    },
    {
      "epoch": 6.091916908836509,
      "grad_norm": 0.7642271518707275,
      "learning_rate": 0.0002348863269044284,
      "loss": 3.2717,
      "step": 56600
    },
    {
      "epoch": 6.097298460876116,
      "grad_norm": 0.8770667314529419,
      "learning_rate": 0.00023456308587436697,
      "loss": 3.2782,
      "step": 56650
    },
    {
      "epoch": 6.102680012915725,
      "grad_norm": 0.803207278251648,
      "learning_rate": 0.00023423984484430554,
      "loss": 3.2724,
      "step": 56700
    },
    {
      "epoch": 6.108061564955333,
      "grad_norm": 0.7858288288116455,
      "learning_rate": 0.00023391660381424413,
      "loss": 3.2772,
      "step": 56750
    },
    {
      "epoch": 6.113443116994941,
      "grad_norm": 0.7991230487823486,
      "learning_rate": 0.00023359336278418272,
      "loss": 3.2718,
      "step": 56800
    },
    {
      "epoch": 6.11882466903455,
      "grad_norm": 0.755598783493042,
      "learning_rate": 0.0002332701217541213,
      "loss": 3.261,
      "step": 56850
    },
    {
      "epoch": 6.124206221074158,
      "grad_norm": 0.760961651802063,
      "learning_rate": 0.0002329468807240599,
      "loss": 3.2824,
      "step": 56900
    },
    {
      "epoch": 6.129587773113766,
      "grad_norm": 0.7972965836524963,
      "learning_rate": 0.00023262363969399845,
      "loss": 3.2669,
      "step": 56950
    },
    {
      "epoch": 6.134969325153374,
      "grad_norm": 0.8029502034187317,
      "learning_rate": 0.00023230039866393708,
      "loss": 3.2598,
      "step": 57000
    },
    {
      "epoch": 6.134969325153374,
      "eval_accuracy": 0.38348210992173626,
      "eval_loss": 3.388145923614502,
      "eval_runtime": 197.7941,
      "eval_samples_per_second": 91.059,
      "eval_steps_per_second": 5.693,
      "step": 57000
    },
    {
      "epoch": 6.140350877192983,
      "grad_norm": 0.7810432314872742,
      "learning_rate": 0.00023197715763387564,
      "loss": 3.2652,
      "step": 57050
    },
    {
      "epoch": 6.1457324292325906,
      "grad_norm": 0.7250264286994934,
      "learning_rate": 0.0002316539166038142,
      "loss": 3.2647,
      "step": 57100
    },
    {
      "epoch": 6.151113981272199,
      "grad_norm": 0.7505423426628113,
      "learning_rate": 0.00023133067557375283,
      "loss": 3.2731,
      "step": 57150
    },
    {
      "epoch": 6.156495533311807,
      "grad_norm": 0.7808586955070496,
      "learning_rate": 0.0002310074345436914,
      "loss": 3.2849,
      "step": 57200
    },
    {
      "epoch": 6.161877085351415,
      "grad_norm": 0.7675476670265198,
      "learning_rate": 0.00023068419351362997,
      "loss": 3.2968,
      "step": 57250
    },
    {
      "epoch": 6.167258637391024,
      "grad_norm": 0.7670409083366394,
      "learning_rate": 0.00023036095248356856,
      "loss": 3.2719,
      "step": 57300
    },
    {
      "epoch": 6.1726401894306315,
      "grad_norm": 0.7464594841003418,
      "learning_rate": 0.00023003771145350716,
      "loss": 3.2811,
      "step": 57350
    },
    {
      "epoch": 6.17802174147024,
      "grad_norm": 0.7713546752929688,
      "learning_rate": 0.00022971447042344572,
      "loss": 3.272,
      "step": 57400
    },
    {
      "epoch": 6.183403293509849,
      "grad_norm": 0.7970213890075684,
      "learning_rate": 0.00022939122939338432,
      "loss": 3.29,
      "step": 57450
    },
    {
      "epoch": 6.188784845549456,
      "grad_norm": 0.8031085133552551,
      "learning_rate": 0.0002290679883633229,
      "loss": 3.2677,
      "step": 57500
    },
    {
      "epoch": 6.194166397589065,
      "grad_norm": 0.7508841156959534,
      "learning_rate": 0.00022874474733326148,
      "loss": 3.2894,
      "step": 57550
    },
    {
      "epoch": 6.1995479496286725,
      "grad_norm": 0.7949566841125488,
      "learning_rate": 0.00022842150630320008,
      "loss": 3.2722,
      "step": 57600
    },
    {
      "epoch": 6.204929501668281,
      "grad_norm": 0.7332485318183899,
      "learning_rate": 0.00022809826527313864,
      "loss": 3.3003,
      "step": 57650
    },
    {
      "epoch": 6.21031105370789,
      "grad_norm": 0.8173457980155945,
      "learning_rate": 0.0002277750242430772,
      "loss": 3.2929,
      "step": 57700
    },
    {
      "epoch": 6.215692605747497,
      "grad_norm": 0.7237012386322021,
      "learning_rate": 0.00022745178321301583,
      "loss": 3.3013,
      "step": 57750
    },
    {
      "epoch": 6.221074157787106,
      "grad_norm": 0.7500722408294678,
      "learning_rate": 0.0002271285421829544,
      "loss": 3.2881,
      "step": 57800
    },
    {
      "epoch": 6.226455709826714,
      "grad_norm": 0.8080673217773438,
      "learning_rate": 0.00022680530115289297,
      "loss": 3.2831,
      "step": 57850
    },
    {
      "epoch": 6.231837261866322,
      "grad_norm": 0.7517146468162537,
      "learning_rate": 0.0002264820601228316,
      "loss": 3.2824,
      "step": 57900
    },
    {
      "epoch": 6.237218813905931,
      "grad_norm": 0.7838613390922546,
      "learning_rate": 0.00022615881909277016,
      "loss": 3.2785,
      "step": 57950
    },
    {
      "epoch": 6.242600365945538,
      "grad_norm": 0.7424094080924988,
      "learning_rate": 0.00022583557806270875,
      "loss": 3.2672,
      "step": 58000
    },
    {
      "epoch": 6.242600365945538,
      "eval_accuracy": 0.38380850319673093,
      "eval_loss": 3.3856444358825684,
      "eval_runtime": 205.4001,
      "eval_samples_per_second": 87.687,
      "eval_steps_per_second": 5.482,
      "step": 58000
    },
    {
      "epoch": 6.247981917985147,
      "grad_norm": 0.7983281016349792,
      "learning_rate": 0.00022551233703264732,
      "loss": 3.2943,
      "step": 58050
    },
    {
      "epoch": 6.253363470024755,
      "grad_norm": 0.7585350275039673,
      "learning_rate": 0.0002251890960025859,
      "loss": 3.2705,
      "step": 58100
    },
    {
      "epoch": 6.258745022064363,
      "grad_norm": 0.784182608127594,
      "learning_rate": 0.0002248658549725245,
      "loss": 3.284,
      "step": 58150
    },
    {
      "epoch": 6.264126574103972,
      "grad_norm": 0.7657437324523926,
      "learning_rate": 0.00022454261394246308,
      "loss": 3.2911,
      "step": 58200
    },
    {
      "epoch": 6.26950812614358,
      "grad_norm": 0.7738139629364014,
      "learning_rate": 0.0002242258377330029,
      "loss": 3.2663,
      "step": 58250
    },
    {
      "epoch": 6.274889678183188,
      "grad_norm": 0.7893441915512085,
      "learning_rate": 0.00022390259670294148,
      "loss": 3.2952,
      "step": 58300
    },
    {
      "epoch": 6.280271230222796,
      "grad_norm": 0.7671812176704407,
      "learning_rate": 0.00022357935567288005,
      "loss": 3.2895,
      "step": 58350
    },
    {
      "epoch": 6.285652782262405,
      "grad_norm": 0.7413412928581238,
      "learning_rate": 0.00022325611464281864,
      "loss": 3.2955,
      "step": 58400
    },
    {
      "epoch": 6.2910343343020125,
      "grad_norm": 0.7650637626647949,
      "learning_rate": 0.00022293287361275723,
      "loss": 3.3065,
      "step": 58450
    },
    {
      "epoch": 6.296415886341621,
      "grad_norm": 0.7891427278518677,
      "learning_rate": 0.0002226096325826958,
      "loss": 3.2731,
      "step": 58500
    },
    {
      "epoch": 6.301797438381229,
      "grad_norm": 0.8112223148345947,
      "learning_rate": 0.0002222863915526344,
      "loss": 3.2942,
      "step": 58550
    },
    {
      "epoch": 6.307178990420837,
      "grad_norm": 0.7923247814178467,
      "learning_rate": 0.00022196315052257296,
      "loss": 3.2622,
      "step": 58600
    },
    {
      "epoch": 6.312560542460446,
      "grad_norm": 0.7752324342727661,
      "learning_rate": 0.00022163990949251156,
      "loss": 3.2741,
      "step": 58650
    },
    {
      "epoch": 6.3179420945000535,
      "grad_norm": 0.8026335835456848,
      "learning_rate": 0.00022131666846245015,
      "loss": 3.2845,
      "step": 58700
    },
    {
      "epoch": 6.323323646539662,
      "grad_norm": 0.7663998603820801,
      "learning_rate": 0.00022099342743238872,
      "loss": 3.2661,
      "step": 58750
    },
    {
      "epoch": 6.328705198579271,
      "grad_norm": 0.7961952090263367,
      "learning_rate": 0.00022067018640232734,
      "loss": 3.2958,
      "step": 58800
    },
    {
      "epoch": 6.334086750618878,
      "grad_norm": 0.7418253421783447,
      "learning_rate": 0.0002203469453722659,
      "loss": 3.3066,
      "step": 58850
    },
    {
      "epoch": 6.339468302658487,
      "grad_norm": 0.8229352831840515,
      "learning_rate": 0.00022002370434220448,
      "loss": 3.3054,
      "step": 58900
    },
    {
      "epoch": 6.344849854698095,
      "grad_norm": 0.75184565782547,
      "learning_rate": 0.00021970046331214307,
      "loss": 3.2975,
      "step": 58950
    },
    {
      "epoch": 6.350231406737703,
      "grad_norm": 0.7392463684082031,
      "learning_rate": 0.00021937722228208167,
      "loss": 3.2765,
      "step": 59000
    },
    {
      "epoch": 6.350231406737703,
      "eval_accuracy": 0.38428353362225587,
      "eval_loss": 3.3835694789886475,
      "eval_runtime": 200.8239,
      "eval_samples_per_second": 89.686,
      "eval_steps_per_second": 5.607,
      "step": 59000
    },
    {
      "epoch": 6.355612958777312,
      "grad_norm": 0.8165404796600342,
      "learning_rate": 0.00021905398125202024,
      "loss": 3.2919,
      "step": 59050
    },
    {
      "epoch": 6.360994510816919,
      "grad_norm": 0.7926778197288513,
      "learning_rate": 0.00021873074022195883,
      "loss": 3.284,
      "step": 59100
    },
    {
      "epoch": 6.366376062856528,
      "grad_norm": 0.7367433309555054,
      "learning_rate": 0.0002184074991918974,
      "loss": 3.2727,
      "step": 59150
    },
    {
      "epoch": 6.371757614896136,
      "grad_norm": 0.8262997269630432,
      "learning_rate": 0.00021808425816183597,
      "loss": 3.2878,
      "step": 59200
    },
    {
      "epoch": 6.377139166935744,
      "grad_norm": 0.7632347941398621,
      "learning_rate": 0.0002177610171317746,
      "loss": 3.2828,
      "step": 59250
    },
    {
      "epoch": 6.382520718975353,
      "grad_norm": 0.841763973236084,
      "learning_rate": 0.00021743777610171315,
      "loss": 3.3098,
      "step": 59300
    },
    {
      "epoch": 6.387902271014961,
      "grad_norm": 0.7442952990531921,
      "learning_rate": 0.00021711453507165172,
      "loss": 3.2836,
      "step": 59350
    },
    {
      "epoch": 6.393283823054569,
      "grad_norm": 0.8750033974647522,
      "learning_rate": 0.00021679129404159034,
      "loss": 3.2941,
      "step": 59400
    },
    {
      "epoch": 6.398665375094177,
      "grad_norm": 0.7784764170646667,
      "learning_rate": 0.0002164680530115289,
      "loss": 3.2964,
      "step": 59450
    },
    {
      "epoch": 6.404046927133785,
      "grad_norm": 0.7881698608398438,
      "learning_rate": 0.00021614481198146748,
      "loss": 3.286,
      "step": 59500
    },
    {
      "epoch": 6.4094284791733935,
      "grad_norm": 0.8030485510826111,
      "learning_rate": 0.00021582157095140607,
      "loss": 3.2949,
      "step": 59550
    },
    {
      "epoch": 6.414810031213002,
      "grad_norm": 0.7736039161682129,
      "learning_rate": 0.00021549832992134467,
      "loss": 3.2819,
      "step": 59600
    },
    {
      "epoch": 6.42019158325261,
      "grad_norm": 0.7619981169700623,
      "learning_rate": 0.00021517508889128324,
      "loss": 3.2705,
      "step": 59650
    },
    {
      "epoch": 6.425573135292218,
      "grad_norm": 0.7922276258468628,
      "learning_rate": 0.00021485184786122183,
      "loss": 3.2897,
      "step": 59700
    },
    {
      "epoch": 6.430954687331827,
      "grad_norm": 0.8027983903884888,
      "learning_rate": 0.0002145286068311604,
      "loss": 3.2993,
      "step": 59750
    },
    {
      "epoch": 6.4363362393714345,
      "grad_norm": 0.732873797416687,
      "learning_rate": 0.00021420536580109902,
      "loss": 3.2883,
      "step": 59800
    },
    {
      "epoch": 6.441717791411043,
      "grad_norm": 0.7547596096992493,
      "learning_rate": 0.0002138821247710376,
      "loss": 3.2883,
      "step": 59850
    },
    {
      "epoch": 6.447099343450651,
      "grad_norm": 0.7760118842124939,
      "learning_rate": 0.00021355888374097615,
      "loss": 3.2854,
      "step": 59900
    },
    {
      "epoch": 6.452480895490259,
      "grad_norm": 0.8054875135421753,
      "learning_rate": 0.00021323564271091478,
      "loss": 3.2997,
      "step": 59950
    },
    {
      "epoch": 6.457862447529868,
      "grad_norm": 0.7943117618560791,
      "learning_rate": 0.00021291240168085334,
      "loss": 3.2792,
      "step": 60000
    },
    {
      "epoch": 6.457862447529868,
      "eval_accuracy": 0.3847438959079258,
      "eval_loss": 3.3782360553741455,
      "eval_runtime": 193.4445,
      "eval_samples_per_second": 93.107,
      "eval_steps_per_second": 5.821,
      "step": 60000
    },
    {
      "epoch": 6.4632439995694755,
      "grad_norm": 0.7851802110671997,
      "learning_rate": 0.0002125891606507919,
      "loss": 3.2994,
      "step": 60050
    },
    {
      "epoch": 6.468625551609084,
      "grad_norm": 0.7804414629936218,
      "learning_rate": 0.0002122659196207305,
      "loss": 3.2957,
      "step": 60100
    },
    {
      "epoch": 6.474007103648693,
      "grad_norm": 0.749247670173645,
      "learning_rate": 0.0002119426785906691,
      "loss": 3.2969,
      "step": 60150
    },
    {
      "epoch": 6.4793886556883,
      "grad_norm": 0.7385237216949463,
      "learning_rate": 0.00021161943756060767,
      "loss": 3.2831,
      "step": 60200
    },
    {
      "epoch": 6.484770207727909,
      "grad_norm": 0.7636567950248718,
      "learning_rate": 0.00021129619653054626,
      "loss": 3.2885,
      "step": 60250
    },
    {
      "epoch": 6.490151759767517,
      "grad_norm": 0.8201603889465332,
      "learning_rate": 0.00021097295550048483,
      "loss": 3.2877,
      "step": 60300
    },
    {
      "epoch": 6.495533311807125,
      "grad_norm": 0.8199951648712158,
      "learning_rate": 0.00021064971447042343,
      "loss": 3.2965,
      "step": 60350
    },
    {
      "epoch": 6.500914863846734,
      "grad_norm": 0.8492313027381897,
      "learning_rate": 0.00021032647344036202,
      "loss": 3.2852,
      "step": 60400
    },
    {
      "epoch": 6.506296415886341,
      "grad_norm": 0.8086408972740173,
      "learning_rate": 0.0002100032324103006,
      "loss": 3.2796,
      "step": 60450
    },
    {
      "epoch": 6.51167796792595,
      "grad_norm": 0.7892538905143738,
      "learning_rate": 0.00020967999138023916,
      "loss": 3.2794,
      "step": 60500
    },
    {
      "epoch": 6.517059519965558,
      "grad_norm": 0.8353956937789917,
      "learning_rate": 0.00020935675035017778,
      "loss": 3.2903,
      "step": 60550
    },
    {
      "epoch": 6.522441072005166,
      "grad_norm": 0.7824742197990417,
      "learning_rate": 0.00020903350932011634,
      "loss": 3.3005,
      "step": 60600
    },
    {
      "epoch": 6.5278226240447745,
      "grad_norm": 0.7651792764663696,
      "learning_rate": 0.0002087102682900549,
      "loss": 3.2913,
      "step": 60650
    },
    {
      "epoch": 6.533204176084383,
      "grad_norm": 0.7593153715133667,
      "learning_rate": 0.00020838702725999353,
      "loss": 3.2911,
      "step": 60700
    },
    {
      "epoch": 6.538585728123991,
      "grad_norm": 0.8071369528770447,
      "learning_rate": 0.0002080637862299321,
      "loss": 3.2957,
      "step": 60750
    },
    {
      "epoch": 6.543967280163599,
      "grad_norm": 0.7907211780548096,
      "learning_rate": 0.0002077405451998707,
      "loss": 3.2991,
      "step": 60800
    },
    {
      "epoch": 6.549348832203208,
      "grad_norm": 0.7981623411178589,
      "learning_rate": 0.00020741730416980926,
      "loss": 3.295,
      "step": 60850
    },
    {
      "epoch": 6.5547303842428155,
      "grad_norm": 0.8324753046035767,
      "learning_rate": 0.00020709406313974786,
      "loss": 3.2947,
      "step": 60900
    },
    {
      "epoch": 6.560111936282424,
      "grad_norm": 0.8198930621147156,
      "learning_rate": 0.00020677082210968645,
      "loss": 3.2791,
      "step": 60950
    },
    {
      "epoch": 6.565493488322032,
      "grad_norm": 0.7891588807106018,
      "learning_rate": 0.00020644758107962502,
      "loss": 3.2942,
      "step": 61000
    },
    {
      "epoch": 6.565493488322032,
      "eval_accuracy": 0.38520664855712766,
      "eval_loss": 3.3762881755828857,
      "eval_runtime": 219.3121,
      "eval_samples_per_second": 82.125,
      "eval_steps_per_second": 5.134,
      "step": 61000
    },
    {
      "epoch": 6.57087504036164,
      "grad_norm": 0.7924179434776306,
      "learning_rate": 0.0002061243400495636,
      "loss": 3.2871,
      "step": 61050
    },
    {
      "epoch": 6.576256592401249,
      "grad_norm": 0.8246179819107056,
      "learning_rate": 0.0002058010990195022,
      "loss": 3.2972,
      "step": 61100
    },
    {
      "epoch": 6.5816381444408565,
      "grad_norm": 0.8679706454277039,
      "learning_rate": 0.00020547785798944078,
      "loss": 3.2756,
      "step": 61150
    },
    {
      "epoch": 6.587019696480465,
      "grad_norm": 0.8301821947097778,
      "learning_rate": 0.00020515461695937934,
      "loss": 3.2851,
      "step": 61200
    },
    {
      "epoch": 6.592401248520073,
      "grad_norm": 0.7571713924407959,
      "learning_rate": 0.00020483784074991918,
      "loss": 3.2959,
      "step": 61250
    },
    {
      "epoch": 6.597782800559681,
      "grad_norm": 0.7680508494377136,
      "learning_rate": 0.00020451459971985775,
      "loss": 3.2983,
      "step": 61300
    },
    {
      "epoch": 6.60316435259929,
      "grad_norm": 0.7844907641410828,
      "learning_rate": 0.00020419135868979634,
      "loss": 3.2993,
      "step": 61350
    },
    {
      "epoch": 6.608545904638898,
      "grad_norm": 0.7639943957328796,
      "learning_rate": 0.0002038681176597349,
      "loss": 3.2966,
      "step": 61400
    },
    {
      "epoch": 6.613927456678506,
      "grad_norm": 0.8657926917076111,
      "learning_rate": 0.0002035448766296735,
      "loss": 3.3063,
      "step": 61450
    },
    {
      "epoch": 6.619309008718115,
      "grad_norm": 0.7783991098403931,
      "learning_rate": 0.0002032216355996121,
      "loss": 3.2774,
      "step": 61500
    },
    {
      "epoch": 6.624690560757722,
      "grad_norm": 0.7940557599067688,
      "learning_rate": 0.00020289839456955067,
      "loss": 3.306,
      "step": 61550
    },
    {
      "epoch": 6.630072112797331,
      "grad_norm": 0.8108448386192322,
      "learning_rate": 0.0002025751535394893,
      "loss": 3.3069,
      "step": 61600
    },
    {
      "epoch": 6.635453664836939,
      "grad_norm": 1.010061502456665,
      "learning_rate": 0.00020225191250942785,
      "loss": 3.2884,
      "step": 61650
    },
    {
      "epoch": 6.640835216876547,
      "grad_norm": 0.8749879598617554,
      "learning_rate": 0.00020192867147936642,
      "loss": 3.2978,
      "step": 61700
    },
    {
      "epoch": 6.6462167689161555,
      "grad_norm": 0.7392527461051941,
      "learning_rate": 0.00020160543044930502,
      "loss": 3.2911,
      "step": 61750
    },
    {
      "epoch": 6.651598320955763,
      "grad_norm": 0.8178379535675049,
      "learning_rate": 0.0002012821894192436,
      "loss": 3.2743,
      "step": 61800
    },
    {
      "epoch": 6.656979872995372,
      "grad_norm": 0.7475403547286987,
      "learning_rate": 0.00020095894838918218,
      "loss": 3.2989,
      "step": 61850
    },
    {
      "epoch": 6.66236142503498,
      "grad_norm": 0.7366738319396973,
      "learning_rate": 0.00020063570735912077,
      "loss": 3.3109,
      "step": 61900
    },
    {
      "epoch": 6.667742977074588,
      "grad_norm": 0.7561606168746948,
      "learning_rate": 0.00020031246632905934,
      "loss": 3.2986,
      "step": 61950
    },
    {
      "epoch": 6.6731245291141965,
      "grad_norm": 0.773314893245697,
      "learning_rate": 0.00019998922529899794,
      "loss": 3.2948,
      "step": 62000
    },
    {
      "epoch": 6.6731245291141965,
      "eval_accuracy": 0.385528587063722,
      "eval_loss": 3.3701326847076416,
      "eval_runtime": 195.6055,
      "eval_samples_per_second": 92.078,
      "eval_steps_per_second": 5.756,
      "step": 62000
    },
    {
      "epoch": 6.678506081153805,
      "grad_norm": 0.8271602988243103,
      "learning_rate": 0.00019966598426893653,
      "loss": 3.3028,
      "step": 62050
    },
    {
      "epoch": 6.683887633193413,
      "grad_norm": 0.8517950177192688,
      "learning_rate": 0.0001993427432388751,
      "loss": 3.2839,
      "step": 62100
    },
    {
      "epoch": 6.689269185233021,
      "grad_norm": 0.7976472973823547,
      "learning_rate": 0.00019901950220881367,
      "loss": 3.2911,
      "step": 62150
    },
    {
      "epoch": 6.69465073727263,
      "grad_norm": 0.7941126227378845,
      "learning_rate": 0.0001986962611787523,
      "loss": 3.3087,
      "step": 62200
    },
    {
      "epoch": 6.7000322893122375,
      "grad_norm": 0.7907029986381531,
      "learning_rate": 0.00019837302014869085,
      "loss": 3.2913,
      "step": 62250
    },
    {
      "epoch": 6.705413841351846,
      "grad_norm": 0.7937203645706177,
      "learning_rate": 0.00019804977911862942,
      "loss": 3.3019,
      "step": 62300
    },
    {
      "epoch": 6.710795393391454,
      "grad_norm": 0.8378528356552124,
      "learning_rate": 0.00019772653808856804,
      "loss": 3.3048,
      "step": 62350
    },
    {
      "epoch": 6.716176945431062,
      "grad_norm": 0.7811477780342102,
      "learning_rate": 0.0001974032970585066,
      "loss": 3.2996,
      "step": 62400
    },
    {
      "epoch": 6.721558497470671,
      "grad_norm": 0.8000446557998657,
      "learning_rate": 0.00019708005602844518,
      "loss": 3.3094,
      "step": 62450
    },
    {
      "epoch": 6.7269400495102785,
      "grad_norm": 0.7956240177154541,
      "learning_rate": 0.00019675681499838377,
      "loss": 3.2969,
      "step": 62500
    },
    {
      "epoch": 6.732321601549887,
      "grad_norm": 0.805371880531311,
      "learning_rate": 0.00019643357396832237,
      "loss": 3.2944,
      "step": 62550
    },
    {
      "epoch": 6.737703153589496,
      "grad_norm": 0.839552640914917,
      "learning_rate": 0.00019611033293826096,
      "loss": 3.2921,
      "step": 62600
    },
    {
      "epoch": 6.743084705629103,
      "grad_norm": 0.8785794973373413,
      "learning_rate": 0.00019578709190819953,
      "loss": 3.3007,
      "step": 62650
    },
    {
      "epoch": 6.748466257668712,
      "grad_norm": 0.7905896306037903,
      "learning_rate": 0.0001954638508781381,
      "loss": 3.2843,
      "step": 62700
    },
    {
      "epoch": 6.75384780970832,
      "grad_norm": 0.7786391973495483,
      "learning_rate": 0.00019514060984807672,
      "loss": 3.305,
      "step": 62750
    },
    {
      "epoch": 6.759229361747928,
      "grad_norm": 0.7951213121414185,
      "learning_rate": 0.0001948173688180153,
      "loss": 3.2966,
      "step": 62800
    },
    {
      "epoch": 6.7646109137875365,
      "grad_norm": 0.7984376549720764,
      "learning_rate": 0.00019449412778795386,
      "loss": 3.2844,
      "step": 62850
    },
    {
      "epoch": 6.769992465827144,
      "grad_norm": 0.7717624306678772,
      "learning_rate": 0.00019417088675789245,
      "loss": 3.2955,
      "step": 62900
    },
    {
      "epoch": 6.775374017866753,
      "grad_norm": 0.796658456325531,
      "learning_rate": 0.00019384764572783104,
      "loss": 3.2949,
      "step": 62950
    },
    {
      "epoch": 6.780755569906361,
      "grad_norm": 0.7735686302185059,
      "learning_rate": 0.0001935244046977696,
      "loss": 3.286,
      "step": 63000
    },
    {
      "epoch": 6.780755569906361,
      "eval_accuracy": 0.3860607689082373,
      "eval_loss": 3.367265462875366,
      "eval_runtime": 196.6474,
      "eval_samples_per_second": 91.59,
      "eval_steps_per_second": 5.726,
      "step": 63000
    },
    {
      "epoch": 6.786137121945969,
      "grad_norm": 0.8305283188819885,
      "learning_rate": 0.0001932011636677082,
      "loss": 3.2906,
      "step": 63050
    },
    {
      "epoch": 6.7915186739855775,
      "grad_norm": 0.8509591817855835,
      "learning_rate": 0.00019287792263764677,
      "loss": 3.3117,
      "step": 63100
    },
    {
      "epoch": 6.796900226025185,
      "grad_norm": 0.7963206171989441,
      "learning_rate": 0.00019255468160758537,
      "loss": 3.311,
      "step": 63150
    },
    {
      "epoch": 6.802281778064794,
      "grad_norm": 0.8255854249000549,
      "learning_rate": 0.00019223144057752396,
      "loss": 3.2965,
      "step": 63200
    },
    {
      "epoch": 6.807663330104402,
      "grad_norm": 0.8133097887039185,
      "learning_rate": 0.00019190819954746253,
      "loss": 3.3035,
      "step": 63250
    },
    {
      "epoch": 6.813044882144011,
      "grad_norm": 0.7953236103057861,
      "learning_rate": 0.0001915849585174011,
      "loss": 3.2804,
      "step": 63300
    },
    {
      "epoch": 6.8184264341836185,
      "grad_norm": 0.8067494034767151,
      "learning_rate": 0.00019126171748733972,
      "loss": 3.2853,
      "step": 63350
    },
    {
      "epoch": 6.823807986223227,
      "grad_norm": 0.7870157361030579,
      "learning_rate": 0.0001909384764572783,
      "loss": 3.2959,
      "step": 63400
    },
    {
      "epoch": 6.829189538262835,
      "grad_norm": 0.8003214597702026,
      "learning_rate": 0.00019061523542721688,
      "loss": 3.3108,
      "step": 63450
    },
    {
      "epoch": 6.834571090302443,
      "grad_norm": 0.8102846145629883,
      "learning_rate": 0.00019029199439715548,
      "loss": 3.2868,
      "step": 63500
    },
    {
      "epoch": 6.839952642342052,
      "grad_norm": 0.7991217374801636,
      "learning_rate": 0.00018996875336709404,
      "loss": 3.2897,
      "step": 63550
    },
    {
      "epoch": 6.8453341943816595,
      "grad_norm": 0.7900161743164062,
      "learning_rate": 0.00018965197715763385,
      "loss": 3.2843,
      "step": 63600
    },
    {
      "epoch": 6.850715746421268,
      "grad_norm": 0.8041602969169617,
      "learning_rate": 0.00018932873612757245,
      "loss": 3.296,
      "step": 63650
    },
    {
      "epoch": 6.856097298460876,
      "grad_norm": 0.805846631526947,
      "learning_rate": 0.00018900549509751104,
      "loss": 3.2927,
      "step": 63700
    },
    {
      "epoch": 6.861478850500484,
      "grad_norm": 0.7478847503662109,
      "learning_rate": 0.0001886822540674496,
      "loss": 3.2971,
      "step": 63750
    },
    {
      "epoch": 6.866860402540093,
      "grad_norm": 0.7645144462585449,
      "learning_rate": 0.00018835901303738818,
      "loss": 3.3148,
      "step": 63800
    },
    {
      "epoch": 6.8722419545797,
      "grad_norm": 0.7955029606819153,
      "learning_rate": 0.0001880357720073268,
      "loss": 3.2849,
      "step": 63850
    },
    {
      "epoch": 6.877623506619309,
      "grad_norm": 0.7824018001556396,
      "learning_rate": 0.00018771253097726537,
      "loss": 3.3064,
      "step": 63900
    },
    {
      "epoch": 6.8830050586589175,
      "grad_norm": 0.7687212824821472,
      "learning_rate": 0.00018738928994720393,
      "loss": 3.3133,
      "step": 63950
    },
    {
      "epoch": 6.888386610698525,
      "grad_norm": 0.8769058585166931,
      "learning_rate": 0.00018706604891714255,
      "loss": 3.3041,
      "step": 64000
    },
    {
      "epoch": 6.888386610698525,
      "eval_accuracy": 0.38647223739439074,
      "eval_loss": 3.36236310005188,
      "eval_runtime": 184.6628,
      "eval_samples_per_second": 97.535,
      "eval_steps_per_second": 6.098,
      "step": 64000
    },
    {
      "epoch": 6.893768162738134,
      "grad_norm": 0.7766793966293335,
      "learning_rate": 0.00018674280788708112,
      "loss": 3.2853,
      "step": 64050
    },
    {
      "epoch": 6.899149714777742,
      "grad_norm": 0.7845059037208557,
      "learning_rate": 0.0001864195668570197,
      "loss": 3.2917,
      "step": 64100
    },
    {
      "epoch": 6.90453126681735,
      "grad_norm": 0.7686408162117004,
      "learning_rate": 0.00018609632582695828,
      "loss": 3.32,
      "step": 64150
    },
    {
      "epoch": 6.9099128188569585,
      "grad_norm": 0.8262222409248352,
      "learning_rate": 0.00018577308479689685,
      "loss": 3.305,
      "step": 64200
    },
    {
      "epoch": 6.915294370896566,
      "grad_norm": 0.7985914945602417,
      "learning_rate": 0.00018544984376683545,
      "loss": 3.2843,
      "step": 64250
    },
    {
      "epoch": 6.920675922936175,
      "grad_norm": 0.8286968469619751,
      "learning_rate": 0.00018512660273677404,
      "loss": 3.2902,
      "step": 64300
    },
    {
      "epoch": 6.926057474975783,
      "grad_norm": 0.8497870564460754,
      "learning_rate": 0.0001848033617067126,
      "loss": 3.2835,
      "step": 64350
    },
    {
      "epoch": 6.931439027015391,
      "grad_norm": 0.8027312159538269,
      "learning_rate": 0.00018448012067665123,
      "loss": 3.2844,
      "step": 64400
    },
    {
      "epoch": 6.9368205790549995,
      "grad_norm": 0.8358915448188782,
      "learning_rate": 0.0001841568796465898,
      "loss": 3.292,
      "step": 64450
    },
    {
      "epoch": 6.942202131094608,
      "grad_norm": 0.9708573222160339,
      "learning_rate": 0.00018383363861652837,
      "loss": 3.2703,
      "step": 64500
    },
    {
      "epoch": 6.947583683134216,
      "grad_norm": 1.3864351511001587,
      "learning_rate": 0.00018351039758646696,
      "loss": 3.2984,
      "step": 64550
    },
    {
      "epoch": 6.952965235173824,
      "grad_norm": 0.7738952040672302,
      "learning_rate": 0.00018318715655640555,
      "loss": 3.2896,
      "step": 64600
    },
    {
      "epoch": 6.958346787213433,
      "grad_norm": 0.7555449604988098,
      "learning_rate": 0.00018286391552634412,
      "loss": 3.3056,
      "step": 64650
    },
    {
      "epoch": 6.9637283392530405,
      "grad_norm": 0.7992121577262878,
      "learning_rate": 0.00018254067449628272,
      "loss": 3.2951,
      "step": 64700
    },
    {
      "epoch": 6.969109891292649,
      "grad_norm": 0.8251198530197144,
      "learning_rate": 0.00018221743346622128,
      "loss": 3.2933,
      "step": 64750
    },
    {
      "epoch": 6.974491443332257,
      "grad_norm": 0.8252221345901489,
      "learning_rate": 0.00018189419243615988,
      "loss": 3.3013,
      "step": 64800
    },
    {
      "epoch": 6.979872995371865,
      "grad_norm": 0.791930615901947,
      "learning_rate": 0.00018157095140609847,
      "loss": 3.2822,
      "step": 64850
    },
    {
      "epoch": 6.985254547411474,
      "grad_norm": 0.7784399390220642,
      "learning_rate": 0.00018124771037603704,
      "loss": 3.3015,
      "step": 64900
    },
    {
      "epoch": 6.990636099451081,
      "grad_norm": 0.7943699955940247,
      "learning_rate": 0.0001809244693459756,
      "loss": 3.3046,
      "step": 64950
    },
    {
      "epoch": 6.99601765149069,
      "grad_norm": 0.8186669945716858,
      "learning_rate": 0.00018060122831591423,
      "loss": 3.3019,
      "step": 65000
    },
    {
      "epoch": 6.99601765149069,
      "eval_accuracy": 0.38673506873001395,
      "eval_loss": 3.3579602241516113,
      "eval_runtime": 184.5292,
      "eval_samples_per_second": 97.605,
      "eval_steps_per_second": 6.102,
      "step": 65000
    },
    {
      "epoch": 7.0013992035302985,
      "grad_norm": 0.7564951777458191,
      "learning_rate": 0.0001802779872858528,
      "loss": 3.2627,
      "step": 65050
    },
    {
      "epoch": 7.006780755569906,
      "grad_norm": 0.762776255607605,
      "learning_rate": 0.00017995474625579137,
      "loss": 3.2103,
      "step": 65100
    },
    {
      "epoch": 7.012162307609515,
      "grad_norm": 0.8338350653648376,
      "learning_rate": 0.00017963150522573,
      "loss": 3.1969,
      "step": 65150
    },
    {
      "epoch": 7.017543859649122,
      "grad_norm": 0.7760246992111206,
      "learning_rate": 0.00017930826419566856,
      "loss": 3.2035,
      "step": 65200
    },
    {
      "epoch": 7.022925411688731,
      "grad_norm": 0.8103695511817932,
      "learning_rate": 0.00017898502316560712,
      "loss": 3.2175,
      "step": 65250
    },
    {
      "epoch": 7.0283069637283395,
      "grad_norm": 0.751948893070221,
      "learning_rate": 0.00017866178213554572,
      "loss": 3.2099,
      "step": 65300
    },
    {
      "epoch": 7.033688515767947,
      "grad_norm": 0.8307621479034424,
      "learning_rate": 0.0001783385411054843,
      "loss": 3.2178,
      "step": 65350
    },
    {
      "epoch": 7.039070067807556,
      "grad_norm": 0.8598138689994812,
      "learning_rate": 0.0001780153000754229,
      "loss": 3.2043,
      "step": 65400
    },
    {
      "epoch": 7.044451619847164,
      "grad_norm": 0.8418580293655396,
      "learning_rate": 0.00017769205904536147,
      "loss": 3.2226,
      "step": 65450
    },
    {
      "epoch": 7.049833171886772,
      "grad_norm": 0.8170554637908936,
      "learning_rate": 0.00017736881801530004,
      "loss": 3.2055,
      "step": 65500
    },
    {
      "epoch": 7.0552147239263805,
      "grad_norm": 0.8064137697219849,
      "learning_rate": 0.00017704557698523866,
      "loss": 3.2164,
      "step": 65550
    },
    {
      "epoch": 7.060596275965988,
      "grad_norm": 0.8077121376991272,
      "learning_rate": 0.00017672233595517723,
      "loss": 3.2216,
      "step": 65600
    },
    {
      "epoch": 7.065977828005597,
      "grad_norm": 0.8185051679611206,
      "learning_rate": 0.0001763990949251158,
      "loss": 3.2117,
      "step": 65650
    },
    {
      "epoch": 7.071359380045205,
      "grad_norm": 0.8072046041488647,
      "learning_rate": 0.00017607585389505442,
      "loss": 3.2012,
      "step": 65700
    },
    {
      "epoch": 7.076740932084813,
      "grad_norm": 0.793249249458313,
      "learning_rate": 0.000175752612864993,
      "loss": 3.2168,
      "step": 65750
    },
    {
      "epoch": 7.0821224841244215,
      "grad_norm": 0.8135785460472107,
      "learning_rate": 0.00017542937183493156,
      "loss": 3.2165,
      "step": 65800
    },
    {
      "epoch": 7.08750403616403,
      "grad_norm": 0.809556782245636,
      "learning_rate": 0.00017510613080487015,
      "loss": 3.2119,
      "step": 65850
    },
    {
      "epoch": 7.092885588203638,
      "grad_norm": 0.7963404655456543,
      "learning_rate": 0.00017478288977480874,
      "loss": 3.2168,
      "step": 65900
    },
    {
      "epoch": 7.098267140243246,
      "grad_norm": 0.7925771474838257,
      "learning_rate": 0.0001744596487447473,
      "loss": 3.2152,
      "step": 65950
    },
    {
      "epoch": 7.103648692282855,
      "grad_norm": 0.8133032917976379,
      "learning_rate": 0.00017414287253528712,
      "loss": 3.2336,
      "step": 66000
    },
    {
      "epoch": 7.103648692282855,
      "eval_accuracy": 0.38642106188423014,
      "eval_loss": 3.365020275115967,
      "eval_runtime": 184.5648,
      "eval_samples_per_second": 97.586,
      "eval_steps_per_second": 6.101,
      "step": 66000
    },
    {
      "epoch": 7.109030244322462,
      "grad_norm": 0.8315716981887817,
      "learning_rate": 0.00017381963150522569,
      "loss": 3.2107,
      "step": 66050
    },
    {
      "epoch": 7.114411796362071,
      "grad_norm": 0.8249379992485046,
      "learning_rate": 0.0001734963904751643,
      "loss": 3.2303,
      "step": 66100
    },
    {
      "epoch": 7.119793348401679,
      "grad_norm": 0.8179910182952881,
      "learning_rate": 0.00017317314944510288,
      "loss": 3.2117,
      "step": 66150
    },
    {
      "epoch": 7.125174900441287,
      "grad_norm": 0.8034934401512146,
      "learning_rate": 0.00017284990841504147,
      "loss": 3.2288,
      "step": 66200
    },
    {
      "epoch": 7.130556452480896,
      "grad_norm": 0.8077306747436523,
      "learning_rate": 0.00017252666738498007,
      "loss": 3.2157,
      "step": 66250
    },
    {
      "epoch": 7.135938004520503,
      "grad_norm": 0.8018556833267212,
      "learning_rate": 0.00017220342635491863,
      "loss": 3.2173,
      "step": 66300
    },
    {
      "epoch": 7.141319556560112,
      "grad_norm": 0.8551734685897827,
      "learning_rate": 0.00017188018532485723,
      "loss": 3.2285,
      "step": 66350
    },
    {
      "epoch": 7.1467011085997205,
      "grad_norm": 0.8494272232055664,
      "learning_rate": 0.0001715569442947958,
      "loss": 3.2397,
      "step": 66400
    },
    {
      "epoch": 7.152082660639328,
      "grad_norm": 0.8257672190666199,
      "learning_rate": 0.0001712337032647344,
      "loss": 3.2463,
      "step": 66450
    },
    {
      "epoch": 7.157464212678937,
      "grad_norm": 0.8906263113021851,
      "learning_rate": 0.00017091046223467298,
      "loss": 3.2295,
      "step": 66500
    },
    {
      "epoch": 7.162845764718545,
      "grad_norm": 0.7685968279838562,
      "learning_rate": 0.00017058722120461155,
      "loss": 3.2235,
      "step": 66550
    },
    {
      "epoch": 7.168227316758153,
      "grad_norm": 0.8113622665405273,
      "learning_rate": 0.00017026398017455012,
      "loss": 3.2251,
      "step": 66600
    },
    {
      "epoch": 7.1736088687977615,
      "grad_norm": 0.8102627992630005,
      "learning_rate": 0.00016994073914448874,
      "loss": 3.2373,
      "step": 66650
    },
    {
      "epoch": 7.178990420837369,
      "grad_norm": 0.8016330003738403,
      "learning_rate": 0.0001696174981144273,
      "loss": 3.2408,
      "step": 66700
    },
    {
      "epoch": 7.184371972876978,
      "grad_norm": 0.7886762619018555,
      "learning_rate": 0.00016929425708436588,
      "loss": 3.2278,
      "step": 66750
    },
    {
      "epoch": 7.189753524916586,
      "grad_norm": 0.8240224719047546,
      "learning_rate": 0.0001689710160543045,
      "loss": 3.2209,
      "step": 66800
    },
    {
      "epoch": 7.195135076956194,
      "grad_norm": 0.8243261575698853,
      "learning_rate": 0.00016864777502424307,
      "loss": 3.2263,
      "step": 66850
    },
    {
      "epoch": 7.2005166289958025,
      "grad_norm": 0.8698069453239441,
      "learning_rate": 0.00016832453399418163,
      "loss": 3.2499,
      "step": 66900
    },
    {
      "epoch": 7.205898181035411,
      "grad_norm": 0.7939157485961914,
      "learning_rate": 0.00016800129296412023,
      "loss": 3.2308,
      "step": 66950
    },
    {
      "epoch": 7.211279733075019,
      "grad_norm": 0.8176946640014648,
      "learning_rate": 0.00016767805193405882,
      "loss": 3.2303,
      "step": 67000
    },
    {
      "epoch": 7.211279733075019,
      "eval_accuracy": 0.3865887132901279,
      "eval_loss": 3.365469217300415,
      "eval_runtime": 184.3158,
      "eval_samples_per_second": 97.718,
      "eval_steps_per_second": 6.109,
      "step": 67000
    },
    {
      "epoch": 7.216661285114627,
      "grad_norm": 0.8914434313774109,
      "learning_rate": 0.0001673548109039974,
      "loss": 3.2291,
      "step": 67050
    },
    {
      "epoch": 7.222042837154235,
      "grad_norm": 0.8209969997406006,
      "learning_rate": 0.00016703156987393598,
      "loss": 3.2302,
      "step": 67100
    },
    {
      "epoch": 7.2274243891938434,
      "grad_norm": 0.7916122674942017,
      "learning_rate": 0.00016670832884387455,
      "loss": 3.2484,
      "step": 67150
    },
    {
      "epoch": 7.232805941233452,
      "grad_norm": 0.7980721592903137,
      "learning_rate": 0.00016638508781381317,
      "loss": 3.237,
      "step": 67200
    },
    {
      "epoch": 7.23818749327306,
      "grad_norm": 0.7647969722747803,
      "learning_rate": 0.00016606184678375174,
      "loss": 3.2169,
      "step": 67250
    },
    {
      "epoch": 7.243569045312668,
      "grad_norm": 0.7874657511711121,
      "learning_rate": 0.0001657386057536903,
      "loss": 3.2475,
      "step": 67300
    },
    {
      "epoch": 7.248950597352277,
      "grad_norm": 0.7803701162338257,
      "learning_rate": 0.00016541536472362893,
      "loss": 3.2368,
      "step": 67350
    },
    {
      "epoch": 7.254332149391884,
      "grad_norm": 0.7929939031600952,
      "learning_rate": 0.0001650921236935675,
      "loss": 3.2343,
      "step": 67400
    },
    {
      "epoch": 7.259713701431493,
      "grad_norm": 0.8952374458312988,
      "learning_rate": 0.00016476888266350607,
      "loss": 3.2381,
      "step": 67450
    },
    {
      "epoch": 7.265095253471101,
      "grad_norm": 0.8030776381492615,
      "learning_rate": 0.00016444564163344466,
      "loss": 3.2286,
      "step": 67500
    },
    {
      "epoch": 7.270476805510709,
      "grad_norm": 0.7860260605812073,
      "learning_rate": 0.00016412240060338326,
      "loss": 3.2233,
      "step": 67550
    },
    {
      "epoch": 7.275858357550318,
      "grad_norm": 0.8729909062385559,
      "learning_rate": 0.00016379915957332182,
      "loss": 3.2317,
      "step": 67600
    },
    {
      "epoch": 7.281239909589925,
      "grad_norm": 0.8101193904876709,
      "learning_rate": 0.00016347591854326042,
      "loss": 3.2376,
      "step": 67650
    },
    {
      "epoch": 7.286621461629534,
      "grad_norm": 1.127354621887207,
      "learning_rate": 0.00016315267751319898,
      "loss": 3.2439,
      "step": 67700
    },
    {
      "epoch": 7.2920030136691425,
      "grad_norm": 0.8090687990188599,
      "learning_rate": 0.00016282943648313755,
      "loss": 3.2474,
      "step": 67750
    },
    {
      "epoch": 7.29738456570875,
      "grad_norm": 0.8139429092407227,
      "learning_rate": 0.00016250619545307617,
      "loss": 3.2462,
      "step": 67800
    },
    {
      "epoch": 7.302766117748359,
      "grad_norm": 0.8385265469551086,
      "learning_rate": 0.00016218295442301474,
      "loss": 3.226,
      "step": 67850
    },
    {
      "epoch": 7.308147669787967,
      "grad_norm": 0.8168327212333679,
      "learning_rate": 0.0001618597133929533,
      "loss": 3.2407,
      "step": 67900
    },
    {
      "epoch": 7.313529221827575,
      "grad_norm": 0.8468223214149475,
      "learning_rate": 0.00016153647236289193,
      "loss": 3.2289,
      "step": 67950
    },
    {
      "epoch": 7.3189107738671835,
      "grad_norm": 0.858020007610321,
      "learning_rate": 0.0001612132313328305,
      "loss": 3.2548,
      "step": 68000
    },
    {
      "epoch": 7.3189107738671835,
      "eval_accuracy": 0.3870435342785193,
      "eval_loss": 3.3603622913360596,
      "eval_runtime": 184.5567,
      "eval_samples_per_second": 97.591,
      "eval_steps_per_second": 6.101,
      "step": 68000
    },
    {
      "epoch": 7.324292325906791,
      "grad_norm": 0.8167481422424316,
      "learning_rate": 0.0001608899903027691,
      "loss": 3.2454,
      "step": 68050
    },
    {
      "epoch": 7.3296738779464,
      "grad_norm": 0.8074829578399658,
      "learning_rate": 0.00016056674927270766,
      "loss": 3.2456,
      "step": 68100
    },
    {
      "epoch": 7.335055429986008,
      "grad_norm": 0.8418408632278442,
      "learning_rate": 0.00016024350824264626,
      "loss": 3.2299,
      "step": 68150
    },
    {
      "epoch": 7.340436982025616,
      "grad_norm": 0.8448289632797241,
      "learning_rate": 0.00015992026721258485,
      "loss": 3.2531,
      "step": 68200
    },
    {
      "epoch": 7.3458185340652244,
      "grad_norm": 0.8528227210044861,
      "learning_rate": 0.00015959702618252342,
      "loss": 3.2318,
      "step": 68250
    },
    {
      "epoch": 7.351200086104833,
      "grad_norm": 0.8260256052017212,
      "learning_rate": 0.00015927378515246199,
      "loss": 3.2515,
      "step": 68300
    },
    {
      "epoch": 7.356581638144441,
      "grad_norm": 0.8155575394630432,
      "learning_rate": 0.0001589505441224006,
      "loss": 3.2469,
      "step": 68350
    },
    {
      "epoch": 7.361963190184049,
      "grad_norm": 1.0006415843963623,
      "learning_rate": 0.00015862730309233917,
      "loss": 3.2245,
      "step": 68400
    },
    {
      "epoch": 7.367344742223658,
      "grad_norm": 0.8224972486495972,
      "learning_rate": 0.00015830406206227774,
      "loss": 3.2367,
      "step": 68450
    },
    {
      "epoch": 7.372726294263265,
      "grad_norm": 0.8263415694236755,
      "learning_rate": 0.00015798082103221636,
      "loss": 3.237,
      "step": 68500
    },
    {
      "epoch": 7.378107846302874,
      "grad_norm": 0.8941337466239929,
      "learning_rate": 0.00015765758000215493,
      "loss": 3.2423,
      "step": 68550
    },
    {
      "epoch": 7.383489398342482,
      "grad_norm": 0.8227181434631348,
      "learning_rate": 0.0001573343389720935,
      "loss": 3.2604,
      "step": 68600
    },
    {
      "epoch": 7.38887095038209,
      "grad_norm": 0.8496069312095642,
      "learning_rate": 0.0001570110979420321,
      "loss": 3.2618,
      "step": 68650
    },
    {
      "epoch": 7.394252502421699,
      "grad_norm": 0.7936899065971375,
      "learning_rate": 0.0001566878569119707,
      "loss": 3.2644,
      "step": 68700
    },
    {
      "epoch": 7.399634054461306,
      "grad_norm": 0.8838838934898376,
      "learning_rate": 0.00015636461588190926,
      "loss": 3.2346,
      "step": 68750
    },
    {
      "epoch": 7.405015606500915,
      "grad_norm": 0.8155412077903748,
      "learning_rate": 0.00015604137485184785,
      "loss": 3.2479,
      "step": 68800
    },
    {
      "epoch": 7.4103971585405235,
      "grad_norm": 0.7817298173904419,
      "learning_rate": 0.00015571813382178642,
      "loss": 3.2348,
      "step": 68850
    },
    {
      "epoch": 7.415778710580131,
      "grad_norm": 0.8149805068969727,
      "learning_rate": 0.000155394892791725,
      "loss": 3.2466,
      "step": 68900
    },
    {
      "epoch": 7.42116026261974,
      "grad_norm": 0.8292676210403442,
      "learning_rate": 0.0001550716517616636,
      "loss": 3.2491,
      "step": 68950
    },
    {
      "epoch": 7.426541814659347,
      "grad_norm": 0.8644334673881531,
      "learning_rate": 0.00015474841073160217,
      "loss": 3.2587,
      "step": 69000
    },
    {
      "epoch": 7.426541814659347,
      "eval_accuracy": 0.3875042225228527,
      "eval_loss": 3.355698585510254,
      "eval_runtime": 184.7566,
      "eval_samples_per_second": 97.485,
      "eval_steps_per_second": 6.095,
      "step": 69000
    },
    {
      "epoch": 7.431923366698956,
      "grad_norm": 0.7975975275039673,
      "learning_rate": 0.0001544251697015408,
      "loss": 3.2362,
      "step": 69050
    },
    {
      "epoch": 7.4373049187385645,
      "grad_norm": 0.8595061898231506,
      "learning_rate": 0.00015410839349208058,
      "loss": 3.2405,
      "step": 69100
    },
    {
      "epoch": 7.442686470778172,
      "grad_norm": 0.7921493649482727,
      "learning_rate": 0.00015378515246201917,
      "loss": 3.221,
      "step": 69150
    },
    {
      "epoch": 7.448068022817781,
      "grad_norm": 0.8106430768966675,
      "learning_rate": 0.00015346191143195774,
      "loss": 3.2534,
      "step": 69200
    },
    {
      "epoch": 7.453449574857389,
      "grad_norm": 0.8793022036552429,
      "learning_rate": 0.00015313867040189633,
      "loss": 3.2193,
      "step": 69250
    },
    {
      "epoch": 7.458831126896997,
      "grad_norm": 0.8052279949188232,
      "learning_rate": 0.00015281542937183493,
      "loss": 3.2463,
      "step": 69300
    },
    {
      "epoch": 7.4642126789366054,
      "grad_norm": 0.8353609442710876,
      "learning_rate": 0.0001524921883417735,
      "loss": 3.2417,
      "step": 69350
    },
    {
      "epoch": 7.469594230976213,
      "grad_norm": 0.7944355010986328,
      "learning_rate": 0.00015216894731171206,
      "loss": 3.2429,
      "step": 69400
    },
    {
      "epoch": 7.474975783015822,
      "grad_norm": 0.8002538084983826,
      "learning_rate": 0.00015184570628165068,
      "loss": 3.2468,
      "step": 69450
    },
    {
      "epoch": 7.48035733505543,
      "grad_norm": 0.8184480667114258,
      "learning_rate": 0.00015152246525158925,
      "loss": 3.251,
      "step": 69500
    },
    {
      "epoch": 7.485738887095038,
      "grad_norm": 0.8724619746208191,
      "learning_rate": 0.00015119922422152782,
      "loss": 3.2654,
      "step": 69550
    },
    {
      "epoch": 7.491120439134646,
      "grad_norm": 0.8161394596099854,
      "learning_rate": 0.00015087598319146644,
      "loss": 3.2362,
      "step": 69600
    },
    {
      "epoch": 7.496501991174255,
      "grad_norm": 0.831038773059845,
      "learning_rate": 0.000150552742161405,
      "loss": 3.2455,
      "step": 69650
    },
    {
      "epoch": 7.501883543213863,
      "grad_norm": 0.810804545879364,
      "learning_rate": 0.00015022950113134358,
      "loss": 3.2377,
      "step": 69700
    },
    {
      "epoch": 7.507265095253471,
      "grad_norm": 0.8601387143135071,
      "learning_rate": 0.00014990626010128217,
      "loss": 3.2384,
      "step": 69750
    },
    {
      "epoch": 7.51264664729308,
      "grad_norm": 0.8107192516326904,
      "learning_rate": 0.00014958301907122077,
      "loss": 3.2399,
      "step": 69800
    },
    {
      "epoch": 7.518028199332687,
      "grad_norm": 0.8003073930740356,
      "learning_rate": 0.00014925977804115933,
      "loss": 3.2443,
      "step": 69850
    },
    {
      "epoch": 7.523409751372296,
      "grad_norm": 0.8305342197418213,
      "learning_rate": 0.00014893653701109793,
      "loss": 3.2546,
      "step": 69900
    },
    {
      "epoch": 7.528791303411904,
      "grad_norm": 0.8125633597373962,
      "learning_rate": 0.0001486132959810365,
      "loss": 3.2267,
      "step": 69950
    },
    {
      "epoch": 7.534172855451512,
      "grad_norm": 0.8240832686424255,
      "learning_rate": 0.0001482900549509751,
      "loss": 3.2523,
      "step": 70000
    },
    {
      "epoch": 7.534172855451512,
      "eval_accuracy": 0.38784832621856113,
      "eval_loss": 3.3515408039093018,
      "eval_runtime": 184.3158,
      "eval_samples_per_second": 97.718,
      "eval_steps_per_second": 6.109,
      "step": 70000
    },
    {
      "epoch": 7.539554407491121,
      "grad_norm": 0.8337522745132446,
      "learning_rate": 0.00014796681392091368,
      "loss": 3.2599,
      "step": 70050
    },
    {
      "epoch": 7.544935959530728,
      "grad_norm": 0.805530309677124,
      "learning_rate": 0.00014764357289085228,
      "loss": 3.2561,
      "step": 70100
    },
    {
      "epoch": 7.550317511570337,
      "grad_norm": 0.8399182558059692,
      "learning_rate": 0.00014732033186079085,
      "loss": 3.2355,
      "step": 70150
    },
    {
      "epoch": 7.5556990636099455,
      "grad_norm": 0.8107239603996277,
      "learning_rate": 0.00014699709083072944,
      "loss": 3.252,
      "step": 70200
    },
    {
      "epoch": 7.561080615649553,
      "grad_norm": 0.8374863266944885,
      "learning_rate": 0.00014667384980066804,
      "loss": 3.2429,
      "step": 70250
    },
    {
      "epoch": 7.566462167689162,
      "grad_norm": 0.840305507183075,
      "learning_rate": 0.0001463506087706066,
      "loss": 3.2472,
      "step": 70300
    },
    {
      "epoch": 7.57184371972877,
      "grad_norm": 0.7951100468635559,
      "learning_rate": 0.0001460273677405452,
      "loss": 3.2496,
      "step": 70350
    },
    {
      "epoch": 7.577225271768378,
      "grad_norm": 0.7982338666915894,
      "learning_rate": 0.00014570412671048377,
      "loss": 3.236,
      "step": 70400
    },
    {
      "epoch": 7.5826068238079865,
      "grad_norm": 0.8277232050895691,
      "learning_rate": 0.00014538088568042236,
      "loss": 3.2402,
      "step": 70450
    },
    {
      "epoch": 7.587988375847594,
      "grad_norm": 0.7989823222160339,
      "learning_rate": 0.00014505764465036093,
      "loss": 3.2383,
      "step": 70500
    },
    {
      "epoch": 7.593369927887203,
      "grad_norm": 0.8105563521385193,
      "learning_rate": 0.00014473440362029952,
      "loss": 3.2492,
      "step": 70550
    },
    {
      "epoch": 7.598751479926811,
      "grad_norm": 0.8283807039260864,
      "learning_rate": 0.00014441116259023812,
      "loss": 3.2417,
      "step": 70600
    },
    {
      "epoch": 7.604133031966419,
      "grad_norm": 0.8422099351882935,
      "learning_rate": 0.00014408792156017669,
      "loss": 3.238,
      "step": 70650
    },
    {
      "epoch": 7.609514584006027,
      "grad_norm": 0.7672217488288879,
      "learning_rate": 0.00014376468053011528,
      "loss": 3.2308,
      "step": 70700
    },
    {
      "epoch": 7.614896136045635,
      "grad_norm": 0.9077163934707642,
      "learning_rate": 0.00014344143950005387,
      "loss": 3.235,
      "step": 70750
    },
    {
      "epoch": 7.620277688085244,
      "grad_norm": 0.8130533695220947,
      "learning_rate": 0.00014311819846999244,
      "loss": 3.2493,
      "step": 70800
    },
    {
      "epoch": 7.625659240124852,
      "grad_norm": 0.8436557650566101,
      "learning_rate": 0.00014279495743993104,
      "loss": 3.2475,
      "step": 70850
    },
    {
      "epoch": 7.63104079216446,
      "grad_norm": 0.8015080094337463,
      "learning_rate": 0.00014247171640986963,
      "loss": 3.2492,
      "step": 70900
    },
    {
      "epoch": 7.636422344204068,
      "grad_norm": 0.8413164019584656,
      "learning_rate": 0.0001421484753798082,
      "loss": 3.2481,
      "step": 70950
    },
    {
      "epoch": 7.641803896243677,
      "grad_norm": 0.8565871119499207,
      "learning_rate": 0.0001418252343497468,
      "loss": 3.24,
      "step": 71000
    },
    {
      "epoch": 7.641803896243677,
      "eval_accuracy": 0.3882562091594167,
      "eval_loss": 3.3485300540924072,
      "eval_runtime": 184.4566,
      "eval_samples_per_second": 97.644,
      "eval_steps_per_second": 6.104,
      "step": 71000
    },
    {
      "epoch": 7.647185448283285,
      "grad_norm": 0.8073666095733643,
      "learning_rate": 0.00014150199331968536,
      "loss": 3.2495,
      "step": 71050
    },
    {
      "epoch": 7.652567000322893,
      "grad_norm": 0.8320879936218262,
      "learning_rate": 0.00014117875228962396,
      "loss": 3.2487,
      "step": 71100
    },
    {
      "epoch": 7.657948552362502,
      "grad_norm": 0.8725606799125671,
      "learning_rate": 0.00014085551125956252,
      "loss": 3.2487,
      "step": 71150
    },
    {
      "epoch": 7.663330104402109,
      "grad_norm": 0.8702536225318909,
      "learning_rate": 0.00014053227022950112,
      "loss": 3.2489,
      "step": 71200
    },
    {
      "epoch": 7.668711656441718,
      "grad_norm": 0.8725094199180603,
      "learning_rate": 0.0001402090291994397,
      "loss": 3.2368,
      "step": 71250
    },
    {
      "epoch": 7.674093208481326,
      "grad_norm": 0.8177581429481506,
      "learning_rate": 0.00013988578816937828,
      "loss": 3.2444,
      "step": 71300
    },
    {
      "epoch": 7.679474760520934,
      "grad_norm": 0.8418097496032715,
      "learning_rate": 0.00013956254713931687,
      "loss": 3.2463,
      "step": 71350
    },
    {
      "epoch": 7.684856312560543,
      "grad_norm": 0.8426458239555359,
      "learning_rate": 0.00013923930610925547,
      "loss": 3.2393,
      "step": 71400
    },
    {
      "epoch": 7.69023786460015,
      "grad_norm": 0.8500837087631226,
      "learning_rate": 0.00013891606507919404,
      "loss": 3.246,
      "step": 71450
    },
    {
      "epoch": 7.695619416639759,
      "grad_norm": 0.8315359950065613,
      "learning_rate": 0.00013859282404913263,
      "loss": 3.2535,
      "step": 71500
    },
    {
      "epoch": 7.7010009686793675,
      "grad_norm": 0.8654466867446899,
      "learning_rate": 0.0001382695830190712,
      "loss": 3.2521,
      "step": 71550
    },
    {
      "epoch": 7.706382520718975,
      "grad_norm": 0.8135007619857788,
      "learning_rate": 0.000137952806809611,
      "loss": 3.2289,
      "step": 71600
    },
    {
      "epoch": 7.711764072758584,
      "grad_norm": 0.8058249950408936,
      "learning_rate": 0.0001376295657795496,
      "loss": 3.233,
      "step": 71650
    },
    {
      "epoch": 7.717145624798192,
      "grad_norm": 0.816959023475647,
      "learning_rate": 0.0001373063247494882,
      "loss": 3.2568,
      "step": 71700
    },
    {
      "epoch": 7.7225271768378,
      "grad_norm": 0.8437349796295166,
      "learning_rate": 0.00013698308371942676,
      "loss": 3.261,
      "step": 71750
    },
    {
      "epoch": 7.727908728877408,
      "grad_norm": 0.895237147808075,
      "learning_rate": 0.00013665984268936536,
      "loss": 3.2422,
      "step": 71800
    },
    {
      "epoch": 7.733290280917016,
      "grad_norm": 0.8356442451477051,
      "learning_rate": 0.00013633660165930395,
      "loss": 3.244,
      "step": 71850
    },
    {
      "epoch": 7.738671832956625,
      "grad_norm": 0.8297794461250305,
      "learning_rate": 0.00013601336062924255,
      "loss": 3.2611,
      "step": 71900
    },
    {
      "epoch": 7.744053384996233,
      "grad_norm": 0.8997371792793274,
      "learning_rate": 0.00013569011959918111,
      "loss": 3.2341,
      "step": 71950
    },
    {
      "epoch": 7.749434937035841,
      "grad_norm": 0.8430255055427551,
      "learning_rate": 0.0001353668785691197,
      "loss": 3.2709,
      "step": 72000
    },
    {
      "epoch": 7.749434937035841,
      "eval_accuracy": 0.38879458421853746,
      "eval_loss": 3.343496799468994,
      "eval_runtime": 184.8717,
      "eval_samples_per_second": 97.424,
      "eval_steps_per_second": 6.091,
      "step": 72000
    },
    {
      "epoch": 7.754816489075449,
      "grad_norm": 1.0251821279525757,
      "learning_rate": 0.00013504363753905828,
      "loss": 3.2467,
      "step": 72050
    },
    {
      "epoch": 7.760198041115058,
      "grad_norm": 0.864314079284668,
      "learning_rate": 0.00013472039650899687,
      "loss": 3.2458,
      "step": 72100
    },
    {
      "epoch": 7.765579593154666,
      "grad_norm": 0.885066032409668,
      "learning_rate": 0.00013439715547893544,
      "loss": 3.273,
      "step": 72150
    },
    {
      "epoch": 7.770961145194274,
      "grad_norm": 0.8064603209495544,
      "learning_rate": 0.00013407391444887403,
      "loss": 3.2389,
      "step": 72200
    },
    {
      "epoch": 7.776342697233883,
      "grad_norm": 0.8416740894317627,
      "learning_rate": 0.0001337506734188126,
      "loss": 3.253,
      "step": 72250
    },
    {
      "epoch": 7.78172424927349,
      "grad_norm": 0.8319293260574341,
      "learning_rate": 0.0001334274323887512,
      "loss": 3.2619,
      "step": 72300
    },
    {
      "epoch": 7.787105801313099,
      "grad_norm": 0.8383718132972717,
      "learning_rate": 0.0001331041913586898,
      "loss": 3.239,
      "step": 72350
    },
    {
      "epoch": 7.792487353352707,
      "grad_norm": 0.8045030236244202,
      "learning_rate": 0.00013278095032862838,
      "loss": 3.2418,
      "step": 72400
    },
    {
      "epoch": 7.797868905392315,
      "grad_norm": 0.8513164520263672,
      "learning_rate": 0.00013245770929856695,
      "loss": 3.2499,
      "step": 72450
    },
    {
      "epoch": 7.803250457431924,
      "grad_norm": 0.8449521064758301,
      "learning_rate": 0.00013213446826850555,
      "loss": 3.2358,
      "step": 72500
    },
    {
      "epoch": 7.808632009471531,
      "grad_norm": 0.8382486701011658,
      "learning_rate": 0.00013181122723844411,
      "loss": 3.2543,
      "step": 72550
    },
    {
      "epoch": 7.81401356151114,
      "grad_norm": 0.8157098889350891,
      "learning_rate": 0.0001314879862083827,
      "loss": 3.2466,
      "step": 72600
    },
    {
      "epoch": 7.819395113550748,
      "grad_norm": 0.82196444272995,
      "learning_rate": 0.00013116474517832128,
      "loss": 3.2652,
      "step": 72650
    },
    {
      "epoch": 7.824776665590356,
      "grad_norm": 0.806985080242157,
      "learning_rate": 0.00013084150414825987,
      "loss": 3.2338,
      "step": 72700
    },
    {
      "epoch": 7.830158217629965,
      "grad_norm": 0.8326713442802429,
      "learning_rate": 0.00013051826311819844,
      "loss": 3.2533,
      "step": 72750
    },
    {
      "epoch": 7.835539769669572,
      "grad_norm": 0.8663233518600464,
      "learning_rate": 0.00013019502208813703,
      "loss": 3.247,
      "step": 72800
    },
    {
      "epoch": 7.840921321709181,
      "grad_norm": 0.8916839957237244,
      "learning_rate": 0.00012987178105807563,
      "loss": 3.2296,
      "step": 72850
    },
    {
      "epoch": 7.846302873748789,
      "grad_norm": 0.8581554889678955,
      "learning_rate": 0.00012954854002801422,
      "loss": 3.2234,
      "step": 72900
    },
    {
      "epoch": 7.851684425788397,
      "grad_norm": 0.907456636428833,
      "learning_rate": 0.0001292252989979528,
      "loss": 3.2416,
      "step": 72950
    },
    {
      "epoch": 7.857065977828006,
      "grad_norm": 0.7775632739067078,
      "learning_rate": 0.00012890205796789139,
      "loss": 3.2504,
      "step": 73000
    },
    {
      "epoch": 7.857065977828006,
      "eval_accuracy": 0.38898211910290525,
      "eval_loss": 3.3408749103546143,
      "eval_runtime": 184.2762,
      "eval_samples_per_second": 97.739,
      "eval_steps_per_second": 6.11,
      "step": 73000
    },
    {
      "epoch": 7.862447529867614,
      "grad_norm": 0.8112790584564209,
      "learning_rate": 0.00012857881693782998,
      "loss": 3.2435,
      "step": 73050
    },
    {
      "epoch": 7.867829081907222,
      "grad_norm": 0.8398995399475098,
      "learning_rate": 0.00012825557590776855,
      "loss": 3.2441,
      "step": 73100
    },
    {
      "epoch": 7.87321063394683,
      "grad_norm": 0.8528559803962708,
      "learning_rate": 0.00012793233487770714,
      "loss": 3.2495,
      "step": 73150
    },
    {
      "epoch": 7.878592185986438,
      "grad_norm": 0.8596890568733215,
      "learning_rate": 0.0001276090938476457,
      "loss": 3.2315,
      "step": 73200
    },
    {
      "epoch": 7.883973738026047,
      "grad_norm": 0.8342053890228271,
      "learning_rate": 0.0001272858528175843,
      "loss": 3.251,
      "step": 73250
    },
    {
      "epoch": 7.889355290065655,
      "grad_norm": 0.8437954783439636,
      "learning_rate": 0.00012696261178752287,
      "loss": 3.2371,
      "step": 73300
    },
    {
      "epoch": 7.894736842105263,
      "grad_norm": 0.8976452350616455,
      "learning_rate": 0.00012663937075746147,
      "loss": 3.2481,
      "step": 73350
    },
    {
      "epoch": 7.900118394144871,
      "grad_norm": 0.8226127028465271,
      "learning_rate": 0.00012631612972740006,
      "loss": 3.2483,
      "step": 73400
    },
    {
      "epoch": 7.90549994618448,
      "grad_norm": 0.8637513518333435,
      "learning_rate": 0.00012599288869733863,
      "loss": 3.2425,
      "step": 73450
    },
    {
      "epoch": 7.910881498224088,
      "grad_norm": 0.8683826923370361,
      "learning_rate": 0.00012566964766727722,
      "loss": 3.2518,
      "step": 73500
    },
    {
      "epoch": 7.916263050263696,
      "grad_norm": 0.8567214012145996,
      "learning_rate": 0.00012534640663721582,
      "loss": 3.2448,
      "step": 73550
    },
    {
      "epoch": 7.921644602303305,
      "grad_norm": 0.8188719749450684,
      "learning_rate": 0.00012502316560715439,
      "loss": 3.2381,
      "step": 73600
    },
    {
      "epoch": 7.927026154342912,
      "grad_norm": 0.8441392779350281,
      "learning_rate": 0.00012469992457709298,
      "loss": 3.2352,
      "step": 73650
    },
    {
      "epoch": 7.932407706382521,
      "grad_norm": 0.8125587701797485,
      "learning_rate": 0.00012437668354703158,
      "loss": 3.253,
      "step": 73700
    },
    {
      "epoch": 7.937789258422129,
      "grad_norm": 0.8730355501174927,
      "learning_rate": 0.00012405344251697014,
      "loss": 3.2302,
      "step": 73750
    },
    {
      "epoch": 7.943170810461737,
      "grad_norm": 0.8560552000999451,
      "learning_rate": 0.00012373020148690874,
      "loss": 3.2306,
      "step": 73800
    },
    {
      "epoch": 7.948552362501346,
      "grad_norm": 0.837441623210907,
      "learning_rate": 0.0001234069604568473,
      "loss": 3.2605,
      "step": 73850
    },
    {
      "epoch": 7.953933914540953,
      "grad_norm": 0.8452093005180359,
      "learning_rate": 0.0001230837194267859,
      "loss": 3.2341,
      "step": 73900
    },
    {
      "epoch": 7.959315466580562,
      "grad_norm": 0.8212348818778992,
      "learning_rate": 0.0001227669432173257,
      "loss": 3.241,
      "step": 73950
    },
    {
      "epoch": 7.96469701862017,
      "grad_norm": 0.8337867259979248,
      "learning_rate": 0.0001224437021872643,
      "loss": 3.237,
      "step": 74000
    },
    {
      "epoch": 7.96469701862017,
      "eval_accuracy": 0.3894580187515327,
      "eval_loss": 3.335864782333374,
      "eval_runtime": 184.7285,
      "eval_samples_per_second": 97.5,
      "eval_steps_per_second": 6.095,
      "step": 74000
    },
    {
      "epoch": 7.970078570659778,
      "grad_norm": 0.811427116394043,
      "learning_rate": 0.00012212046115720287,
      "loss": 3.2509,
      "step": 74050
    },
    {
      "epoch": 7.975460122699387,
      "grad_norm": 0.8365855813026428,
      "learning_rate": 0.00012179722012714146,
      "loss": 3.2346,
      "step": 74100
    },
    {
      "epoch": 7.980841674738995,
      "grad_norm": 0.8553631901741028,
      "learning_rate": 0.00012147397909708004,
      "loss": 3.2328,
      "step": 74150
    },
    {
      "epoch": 7.986223226778603,
      "grad_norm": 0.8902938365936279,
      "learning_rate": 0.00012115073806701864,
      "loss": 3.2461,
      "step": 74200
    },
    {
      "epoch": 7.991604778818211,
      "grad_norm": 0.8877424001693726,
      "learning_rate": 0.0001208274970369572,
      "loss": 3.248,
      "step": 74250
    },
    {
      "epoch": 7.996986330857819,
      "grad_norm": 0.8594383597373962,
      "learning_rate": 0.0001205042560068958,
      "loss": 3.2284,
      "step": 74300
    },
    {
      "epoch": 8.002367882897428,
      "grad_norm": 0.839432954788208,
      "learning_rate": 0.0001201810149768344,
      "loss": 3.2064,
      "step": 74350
    },
    {
      "epoch": 8.007749434937036,
      "grad_norm": 0.8433920741081238,
      "learning_rate": 0.00011985777394677296,
      "loss": 3.1804,
      "step": 74400
    },
    {
      "epoch": 8.013130986976645,
      "grad_norm": 0.8507816791534424,
      "learning_rate": 0.00011953453291671154,
      "loss": 3.1888,
      "step": 74450
    },
    {
      "epoch": 8.018512539016251,
      "grad_norm": 0.8435181379318237,
      "learning_rate": 0.00011921129188665014,
      "loss": 3.1712,
      "step": 74500
    },
    {
      "epoch": 8.02389409105586,
      "grad_norm": 0.8020918369293213,
      "learning_rate": 0.0001188880508565887,
      "loss": 3.1512,
      "step": 74550
    },
    {
      "epoch": 8.029275643095469,
      "grad_norm": 0.8797098994255066,
      "learning_rate": 0.0001185648098265273,
      "loss": 3.1775,
      "step": 74600
    },
    {
      "epoch": 8.034657195135077,
      "grad_norm": 0.8259567022323608,
      "learning_rate": 0.0001182415687964659,
      "loss": 3.1821,
      "step": 74650
    },
    {
      "epoch": 8.040038747174686,
      "grad_norm": 0.856727659702301,
      "learning_rate": 0.00011791832776640448,
      "loss": 3.1783,
      "step": 74700
    },
    {
      "epoch": 8.045420299214294,
      "grad_norm": 0.8482377529144287,
      "learning_rate": 0.00011759508673634306,
      "loss": 3.1699,
      "step": 74750
    },
    {
      "epoch": 8.050801851253901,
      "grad_norm": 0.8650091886520386,
      "learning_rate": 0.00011727184570628164,
      "loss": 3.1604,
      "step": 74800
    },
    {
      "epoch": 8.05618340329351,
      "grad_norm": 0.8241996169090271,
      "learning_rate": 0.00011694860467622023,
      "loss": 3.1704,
      "step": 74850
    },
    {
      "epoch": 8.061564955333118,
      "grad_norm": 0.865490198135376,
      "learning_rate": 0.0001166253636461588,
      "loss": 3.1623,
      "step": 74900
    },
    {
      "epoch": 8.066946507372727,
      "grad_norm": 0.8584668040275574,
      "learning_rate": 0.0001163021226160974,
      "loss": 3.175,
      "step": 74950
    },
    {
      "epoch": 8.072328059412335,
      "grad_norm": 0.8915995955467224,
      "learning_rate": 0.00011597888158603598,
      "loss": 3.1875,
      "step": 75000
    },
    {
      "epoch": 8.072328059412335,
      "eval_accuracy": 0.3893782675318769,
      "eval_loss": 3.341822862625122,
      "eval_runtime": 184.3064,
      "eval_samples_per_second": 97.723,
      "eval_steps_per_second": 6.109,
      "step": 75000
    },
    {
      "epoch": 8.077709611451942,
      "grad_norm": 0.8351331949234009,
      "learning_rate": 0.00011565564055597456,
      "loss": 3.1679,
      "step": 75050
    },
    {
      "epoch": 8.08309116349155,
      "grad_norm": 0.8961378931999207,
      "learning_rate": 0.00011533239952591314,
      "loss": 3.1748,
      "step": 75100
    },
    {
      "epoch": 8.088472715531159,
      "grad_norm": 0.8230721950531006,
      "learning_rate": 0.00011500915849585173,
      "loss": 3.1775,
      "step": 75150
    },
    {
      "epoch": 8.093854267570768,
      "grad_norm": 0.8151953816413879,
      "learning_rate": 0.00011468591746579033,
      "loss": 3.186,
      "step": 75200
    },
    {
      "epoch": 8.099235819610376,
      "grad_norm": 0.8393318057060242,
      "learning_rate": 0.0001143626764357289,
      "loss": 3.1826,
      "step": 75250
    },
    {
      "epoch": 8.104617371649983,
      "grad_norm": 0.8211478590965271,
      "learning_rate": 0.00011403943540566749,
      "loss": 3.1692,
      "step": 75300
    },
    {
      "epoch": 8.109998923689592,
      "grad_norm": 0.8727355003356934,
      "learning_rate": 0.00011371619437560607,
      "loss": 3.1792,
      "step": 75350
    },
    {
      "epoch": 8.1153804757292,
      "grad_norm": 0.8237702250480652,
      "learning_rate": 0.00011339295334554464,
      "loss": 3.175,
      "step": 75400
    },
    {
      "epoch": 8.120762027768809,
      "grad_norm": 0.8641880750656128,
      "learning_rate": 0.00011306971231548323,
      "loss": 3.1698,
      "step": 75450
    },
    {
      "epoch": 8.126143579808417,
      "grad_norm": 0.9081944227218628,
      "learning_rate": 0.00011274647128542183,
      "loss": 3.1758,
      "step": 75500
    },
    {
      "epoch": 8.131525131848026,
      "grad_norm": 0.904127299785614,
      "learning_rate": 0.0001124232302553604,
      "loss": 3.1685,
      "step": 75550
    },
    {
      "epoch": 8.136906683887632,
      "grad_norm": 0.9014459252357483,
      "learning_rate": 0.00011209998922529899,
      "loss": 3.1795,
      "step": 75600
    },
    {
      "epoch": 8.142288235927241,
      "grad_norm": 0.9156743288040161,
      "learning_rate": 0.00011177674819523757,
      "loss": 3.1707,
      "step": 75650
    },
    {
      "epoch": 8.14766978796685,
      "grad_norm": 0.7989534139633179,
      "learning_rate": 0.00011145350716517617,
      "loss": 3.1888,
      "step": 75700
    },
    {
      "epoch": 8.153051340006458,
      "grad_norm": 0.8750432729721069,
      "learning_rate": 0.00011113026613511473,
      "loss": 3.1768,
      "step": 75750
    },
    {
      "epoch": 8.158432892046067,
      "grad_norm": 0.8472692966461182,
      "learning_rate": 0.00011080702510505333,
      "loss": 3.1748,
      "step": 75800
    },
    {
      "epoch": 8.163814444085673,
      "grad_norm": 0.8728067278862,
      "learning_rate": 0.00011048378407499191,
      "loss": 3.1696,
      "step": 75850
    },
    {
      "epoch": 8.169195996125282,
      "grad_norm": 0.8552107214927673,
      "learning_rate": 0.00011016054304493049,
      "loss": 3.1829,
      "step": 75900
    },
    {
      "epoch": 8.17457754816489,
      "grad_norm": 0.9356092214584351,
      "learning_rate": 0.00010983730201486907,
      "loss": 3.1788,
      "step": 75950
    },
    {
      "epoch": 8.1799591002045,
      "grad_norm": 0.9444727897644043,
      "learning_rate": 0.00010951406098480767,
      "loss": 3.1911,
      "step": 76000
    },
    {
      "epoch": 8.1799591002045,
      "eval_accuracy": 0.38965935255265294,
      "eval_loss": 3.3401012420654297,
      "eval_runtime": 184.4345,
      "eval_samples_per_second": 97.655,
      "eval_steps_per_second": 6.105,
      "step": 76000
    },
    {
      "epoch": 8.185340652244108,
      "grad_norm": 0.8912250995635986,
      "learning_rate": 0.00010919081995474623,
      "loss": 3.1792,
      "step": 76050
    },
    {
      "epoch": 8.190722204283716,
      "grad_norm": 0.8142228722572327,
      "learning_rate": 0.00010886757892468483,
      "loss": 3.1765,
      "step": 76100
    },
    {
      "epoch": 8.196103756323323,
      "grad_norm": 0.8855531811714172,
      "learning_rate": 0.00010854433789462342,
      "loss": 3.1935,
      "step": 76150
    },
    {
      "epoch": 8.201485308362932,
      "grad_norm": 0.85954350233078,
      "learning_rate": 0.000108221096864562,
      "loss": 3.1694,
      "step": 76200
    },
    {
      "epoch": 8.20686686040254,
      "grad_norm": 0.8709927201271057,
      "learning_rate": 0.00010789785583450059,
      "loss": 3.1714,
      "step": 76250
    },
    {
      "epoch": 8.212248412442149,
      "grad_norm": 0.858425498008728,
      "learning_rate": 0.00010757461480443917,
      "loss": 3.1887,
      "step": 76300
    },
    {
      "epoch": 8.217629964481757,
      "grad_norm": 0.8919404745101929,
      "learning_rate": 0.00010725137377437776,
      "loss": 3.19,
      "step": 76350
    },
    {
      "epoch": 8.223011516521364,
      "grad_norm": 0.8458712100982666,
      "learning_rate": 0.00010692813274431633,
      "loss": 3.1963,
      "step": 76400
    },
    {
      "epoch": 8.228393068560973,
      "grad_norm": 0.8558886051177979,
      "learning_rate": 0.00010660489171425492,
      "loss": 3.2005,
      "step": 76450
    },
    {
      "epoch": 8.233774620600581,
      "grad_norm": 0.882189929485321,
      "learning_rate": 0.0001062816506841935,
      "loss": 3.2036,
      "step": 76500
    },
    {
      "epoch": 8.23915617264019,
      "grad_norm": 0.8683764338493347,
      "learning_rate": 0.00010595840965413209,
      "loss": 3.1813,
      "step": 76550
    },
    {
      "epoch": 8.244537724679798,
      "grad_norm": 0.8498860001564026,
      "learning_rate": 0.00010563516862407067,
      "loss": 3.1974,
      "step": 76600
    },
    {
      "epoch": 8.249919276719407,
      "grad_norm": 0.8949213027954102,
      "learning_rate": 0.00010531192759400926,
      "loss": 3.1948,
      "step": 76650
    },
    {
      "epoch": 8.255300828759013,
      "grad_norm": 0.8786093592643738,
      "learning_rate": 0.00010498868656394784,
      "loss": 3.1876,
      "step": 76700
    },
    {
      "epoch": 8.260682380798622,
      "grad_norm": 0.8458141684532166,
      "learning_rate": 0.00010466544553388642,
      "loss": 3.1877,
      "step": 76750
    },
    {
      "epoch": 8.26606393283823,
      "grad_norm": 0.9205921292304993,
      "learning_rate": 0.000104342204503825,
      "loss": 3.1627,
      "step": 76800
    },
    {
      "epoch": 8.27144548487784,
      "grad_norm": 0.8844447731971741,
      "learning_rate": 0.0001040189634737636,
      "loss": 3.1977,
      "step": 76850
    },
    {
      "epoch": 8.276827036917448,
      "grad_norm": 0.9396103024482727,
      "learning_rate": 0.00010369572244370217,
      "loss": 3.1662,
      "step": 76900
    },
    {
      "epoch": 8.282208588957054,
      "grad_norm": 0.849521279335022,
      "learning_rate": 0.00010337248141364076,
      "loss": 3.1707,
      "step": 76950
    },
    {
      "epoch": 8.287590140996663,
      "grad_norm": 0.8881746530532837,
      "learning_rate": 0.00010304924038357936,
      "loss": 3.1832,
      "step": 77000
    },
    {
      "epoch": 8.287590140996663,
      "eval_accuracy": 0.3898635113288563,
      "eval_loss": 3.3389201164245605,
      "eval_runtime": 184.5993,
      "eval_samples_per_second": 97.568,
      "eval_steps_per_second": 6.1,
      "step": 77000
    },
    {
      "epoch": 8.292971693036272,
      "grad_norm": 0.9691070318222046,
      "learning_rate": 0.00010272599935351792,
      "loss": 3.171,
      "step": 77050
    },
    {
      "epoch": 8.29835324507588,
      "grad_norm": 0.8988675475120544,
      "learning_rate": 0.00010240275832345652,
      "loss": 3.1843,
      "step": 77100
    },
    {
      "epoch": 8.303734797115489,
      "grad_norm": 0.8782178163528442,
      "learning_rate": 0.0001020795172933951,
      "loss": 3.1643,
      "step": 77150
    },
    {
      "epoch": 8.309116349155097,
      "grad_norm": 0.9221593141555786,
      "learning_rate": 0.0001017562762633337,
      "loss": 3.1716,
      "step": 77200
    },
    {
      "epoch": 8.314497901194704,
      "grad_norm": 0.8310253024101257,
      "learning_rate": 0.00010143303523327226,
      "loss": 3.1961,
      "step": 77250
    },
    {
      "epoch": 8.319879453234313,
      "grad_norm": 0.9973636269569397,
      "learning_rate": 0.00010110979420321086,
      "loss": 3.175,
      "step": 77300
    },
    {
      "epoch": 8.325261005273921,
      "grad_norm": 0.8973341584205627,
      "learning_rate": 0.00010078655317314944,
      "loss": 3.1962,
      "step": 77350
    },
    {
      "epoch": 8.33064255731353,
      "grad_norm": 0.8728778958320618,
      "learning_rate": 0.00010046331214308802,
      "loss": 3.1857,
      "step": 77400
    },
    {
      "epoch": 8.336024109353138,
      "grad_norm": 0.8861868977546692,
      "learning_rate": 0.0001001400711130266,
      "loss": 3.1937,
      "step": 77450
    },
    {
      "epoch": 8.341405661392745,
      "grad_norm": 0.8422714471817017,
      "learning_rate": 9.98168300829652e-05,
      "loss": 3.1863,
      "step": 77500
    },
    {
      "epoch": 8.346787213432354,
      "grad_norm": 0.8660024404525757,
      "learning_rate": 9.949358905290376e-05,
      "loss": 3.1804,
      "step": 77550
    },
    {
      "epoch": 8.352168765471962,
      "grad_norm": 0.9151667952537537,
      "learning_rate": 9.917034802284236e-05,
      "loss": 3.1984,
      "step": 77600
    },
    {
      "epoch": 8.35755031751157,
      "grad_norm": 0.8923110365867615,
      "learning_rate": 9.884710699278094e-05,
      "loss": 3.1928,
      "step": 77650
    },
    {
      "epoch": 8.36293186955118,
      "grad_norm": 0.8670412302017212,
      "learning_rate": 9.852386596271953e-05,
      "loss": 3.1825,
      "step": 77700
    },
    {
      "epoch": 8.368313421590786,
      "grad_norm": 0.8677191138267517,
      "learning_rate": 9.82006249326581e-05,
      "loss": 3.1833,
      "step": 77750
    },
    {
      "epoch": 8.373694973630395,
      "grad_norm": 0.9246715903282166,
      "learning_rate": 9.78773839025967e-05,
      "loss": 3.1757,
      "step": 77800
    },
    {
      "epoch": 8.379076525670003,
      "grad_norm": 0.8516037464141846,
      "learning_rate": 9.755414287253529e-05,
      "loss": 3.1849,
      "step": 77850
    },
    {
      "epoch": 8.384458077709612,
      "grad_norm": 0.8575651049613953,
      "learning_rate": 9.723736666307508e-05,
      "loss": 3.2093,
      "step": 77900
    },
    {
      "epoch": 8.38983962974922,
      "grad_norm": 0.8391822576522827,
      "learning_rate": 9.691412563301368e-05,
      "loss": 3.1936,
      "step": 77950
    },
    {
      "epoch": 8.395221181788829,
      "grad_norm": 0.9174009561538696,
      "learning_rate": 9.659088460295227e-05,
      "loss": 3.2091,
      "step": 78000
    },
    {
      "epoch": 8.395221181788829,
      "eval_accuracy": 0.3902286936848008,
      "eval_loss": 3.334301471710205,
      "eval_runtime": 184.4878,
      "eval_samples_per_second": 97.627,
      "eval_steps_per_second": 6.103,
      "step": 78000
    },
    {
      "epoch": 8.400602733828435,
      "grad_norm": 0.8629387021064758,
      "learning_rate": 9.626764357289084e-05,
      "loss": 3.2102,
      "step": 78050
    },
    {
      "epoch": 8.405984285868044,
      "grad_norm": 0.9055091738700867,
      "learning_rate": 9.594440254282943e-05,
      "loss": 3.1893,
      "step": 78100
    },
    {
      "epoch": 8.411365837907653,
      "grad_norm": 0.8033608794212341,
      "learning_rate": 9.562116151276802e-05,
      "loss": 3.1836,
      "step": 78150
    },
    {
      "epoch": 8.416747389947261,
      "grad_norm": 0.8721442818641663,
      "learning_rate": 9.52979204827066e-05,
      "loss": 3.2035,
      "step": 78200
    },
    {
      "epoch": 8.42212894198687,
      "grad_norm": 0.9385592341423035,
      "learning_rate": 9.497467945264518e-05,
      "loss": 3.1803,
      "step": 78250
    },
    {
      "epoch": 8.427510494026476,
      "grad_norm": 0.9042940139770508,
      "learning_rate": 9.465143842258377e-05,
      "loss": 3.1995,
      "step": 78300
    },
    {
      "epoch": 8.432892046066085,
      "grad_norm": 0.8394501805305481,
      "learning_rate": 9.432819739252234e-05,
      "loss": 3.1991,
      "step": 78350
    },
    {
      "epoch": 8.438273598105694,
      "grad_norm": 0.9093876481056213,
      "learning_rate": 9.400495636246093e-05,
      "loss": 3.1846,
      "step": 78400
    },
    {
      "epoch": 8.443655150145302,
      "grad_norm": 0.9105165004730225,
      "learning_rate": 9.368171533239952e-05,
      "loss": 3.2071,
      "step": 78450
    },
    {
      "epoch": 8.44903670218491,
      "grad_norm": 0.9067189693450928,
      "learning_rate": 9.335847430233811e-05,
      "loss": 3.2039,
      "step": 78500
    },
    {
      "epoch": 8.45441825422452,
      "grad_norm": 0.921536386013031,
      "learning_rate": 9.303523327227668e-05,
      "loss": 3.1901,
      "step": 78550
    },
    {
      "epoch": 8.459799806264126,
      "grad_norm": 0.9320562481880188,
      "learning_rate": 9.271199224221527e-05,
      "loss": 3.2056,
      "step": 78600
    },
    {
      "epoch": 8.465181358303735,
      "grad_norm": 0.8535770177841187,
      "learning_rate": 9.238875121215387e-05,
      "loss": 3.2026,
      "step": 78650
    },
    {
      "epoch": 8.470562910343343,
      "grad_norm": 0.8834720849990845,
      "learning_rate": 9.206551018209243e-05,
      "loss": 3.1858,
      "step": 78700
    },
    {
      "epoch": 8.475944462382952,
      "grad_norm": 0.8686609268188477,
      "learning_rate": 9.174226915203103e-05,
      "loss": 3.2154,
      "step": 78750
    },
    {
      "epoch": 8.48132601442256,
      "grad_norm": 0.8723481893539429,
      "learning_rate": 9.141902812196961e-05,
      "loss": 3.2003,
      "step": 78800
    },
    {
      "epoch": 8.486707566462167,
      "grad_norm": 0.9125947952270508,
      "learning_rate": 9.109578709190818e-05,
      "loss": 3.1998,
      "step": 78850
    },
    {
      "epoch": 8.492089118501776,
      "grad_norm": 0.918624758720398,
      "learning_rate": 9.077254606184677e-05,
      "loss": 3.1853,
      "step": 78900
    },
    {
      "epoch": 8.497470670541384,
      "grad_norm": 0.8740938901901245,
      "learning_rate": 9.044930503178537e-05,
      "loss": 3.1999,
      "step": 78950
    },
    {
      "epoch": 8.502852222580993,
      "grad_norm": 0.9546645879745483,
      "learning_rate": 9.012606400172395e-05,
      "loss": 3.1961,
      "step": 79000
    },
    {
      "epoch": 8.502852222580993,
      "eval_accuracy": 0.3906681946160104,
      "eval_loss": 3.329730749130249,
      "eval_runtime": 184.4619,
      "eval_samples_per_second": 97.641,
      "eval_steps_per_second": 6.104,
      "step": 79000
    },
    {
      "epoch": 8.508233774620601,
      "grad_norm": 0.9086668491363525,
      "learning_rate": 8.980282297166253e-05,
      "loss": 3.2004,
      "step": 79050
    },
    {
      "epoch": 8.513615326660208,
      "grad_norm": 0.917205274105072,
      "learning_rate": 8.947958194160111e-05,
      "loss": 3.1852,
      "step": 79100
    },
    {
      "epoch": 8.518996878699816,
      "grad_norm": 0.8701863884925842,
      "learning_rate": 8.91563409115397e-05,
      "loss": 3.1789,
      "step": 79150
    },
    {
      "epoch": 8.524378430739425,
      "grad_norm": 0.883725643157959,
      "learning_rate": 8.883309988147827e-05,
      "loss": 3.2018,
      "step": 79200
    },
    {
      "epoch": 8.529759982779034,
      "grad_norm": 0.8599157929420471,
      "learning_rate": 8.850985885141687e-05,
      "loss": 3.1994,
      "step": 79250
    },
    {
      "epoch": 8.535141534818642,
      "grad_norm": 0.9229496121406555,
      "learning_rate": 8.818661782135545e-05,
      "loss": 3.1917,
      "step": 79300
    },
    {
      "epoch": 8.54052308685825,
      "grad_norm": 0.876970112323761,
      "learning_rate": 8.786337679129403e-05,
      "loss": 3.1996,
      "step": 79350
    },
    {
      "epoch": 8.545904638897857,
      "grad_norm": 0.9309645295143127,
      "learning_rate": 8.754013576123261e-05,
      "loss": 3.1917,
      "step": 79400
    },
    {
      "epoch": 8.551286190937466,
      "grad_norm": 0.9548830986022949,
      "learning_rate": 8.72168947311712e-05,
      "loss": 3.1928,
      "step": 79450
    },
    {
      "epoch": 8.556667742977075,
      "grad_norm": 0.8856016993522644,
      "learning_rate": 8.68936537011098e-05,
      "loss": 3.1995,
      "step": 79500
    },
    {
      "epoch": 8.562049295016683,
      "grad_norm": 0.8640626072883606,
      "learning_rate": 8.657041267104837e-05,
      "loss": 3.1849,
      "step": 79550
    },
    {
      "epoch": 8.567430847056292,
      "grad_norm": 0.870431661605835,
      "learning_rate": 8.624717164098696e-05,
      "loss": 3.2112,
      "step": 79600
    },
    {
      "epoch": 8.572812399095898,
      "grad_norm": 0.9181054830551147,
      "learning_rate": 8.592393061092554e-05,
      "loss": 3.1814,
      "step": 79650
    },
    {
      "epoch": 8.578193951135507,
      "grad_norm": 0.8909810781478882,
      "learning_rate": 8.560068958086412e-05,
      "loss": 3.1987,
      "step": 79700
    },
    {
      "epoch": 8.583575503175116,
      "grad_norm": 0.937873125076294,
      "learning_rate": 8.52774485508027e-05,
      "loss": 3.2079,
      "step": 79750
    },
    {
      "epoch": 8.588957055214724,
      "grad_norm": 0.876768946647644,
      "learning_rate": 8.49542075207413e-05,
      "loss": 3.1886,
      "step": 79800
    },
    {
      "epoch": 8.594338607254333,
      "grad_norm": 0.9097497463226318,
      "learning_rate": 8.463096649067987e-05,
      "loss": 3.1766,
      "step": 79850
    },
    {
      "epoch": 8.599720159293941,
      "grad_norm": 0.8743965029716492,
      "learning_rate": 8.430772546061846e-05,
      "loss": 3.1857,
      "step": 79900
    },
    {
      "epoch": 8.605101711333548,
      "grad_norm": 0.8830518126487732,
      "learning_rate": 8.398448443055704e-05,
      "loss": 3.1977,
      "step": 79950
    },
    {
      "epoch": 8.610483263373157,
      "grad_norm": 0.9307997822761536,
      "learning_rate": 8.366770822109685e-05,
      "loss": 3.1893,
      "step": 80000
    },
    {
      "epoch": 8.610483263373157,
      "eval_accuracy": 0.39094504217416165,
      "eval_loss": 3.328190803527832,
      "eval_runtime": 184.582,
      "eval_samples_per_second": 97.577,
      "eval_steps_per_second": 6.1,
      "step": 80000
    },
    {
      "epoch": 8.615864815412765,
      "grad_norm": 0.923064649105072,
      "learning_rate": 8.334446719103544e-05,
      "loss": 3.2013,
      "step": 80050
    },
    {
      "epoch": 8.621246367452374,
      "grad_norm": 0.8655891418457031,
      "learning_rate": 8.302122616097403e-05,
      "loss": 3.2007,
      "step": 80100
    },
    {
      "epoch": 8.626627919491982,
      "grad_norm": 0.9024025797843933,
      "learning_rate": 8.269798513091261e-05,
      "loss": 3.2096,
      "step": 80150
    },
    {
      "epoch": 8.632009471531589,
      "grad_norm": 0.9122940897941589,
      "learning_rate": 8.237474410085119e-05,
      "loss": 3.1968,
      "step": 80200
    },
    {
      "epoch": 8.637391023571197,
      "grad_norm": 0.8803995251655579,
      "learning_rate": 8.205150307078978e-05,
      "loss": 3.2116,
      "step": 80250
    },
    {
      "epoch": 8.642772575610806,
      "grad_norm": 0.8858354687690735,
      "learning_rate": 8.172826204072838e-05,
      "loss": 3.2,
      "step": 80300
    },
    {
      "epoch": 8.648154127650415,
      "grad_norm": 0.8597419261932373,
      "learning_rate": 8.140502101066694e-05,
      "loss": 3.2082,
      "step": 80350
    },
    {
      "epoch": 8.653535679690023,
      "grad_norm": 0.8953153491020203,
      "learning_rate": 8.108177998060553e-05,
      "loss": 3.1982,
      "step": 80400
    },
    {
      "epoch": 8.658917231729632,
      "grad_norm": 0.8973867297172546,
      "learning_rate": 8.075853895054412e-05,
      "loss": 3.2055,
      "step": 80450
    },
    {
      "epoch": 8.664298783769238,
      "grad_norm": 0.878873884677887,
      "learning_rate": 8.043529792048269e-05,
      "loss": 3.192,
      "step": 80500
    },
    {
      "epoch": 8.669680335808847,
      "grad_norm": 0.9598588347434998,
      "learning_rate": 8.011205689042128e-05,
      "loss": 3.2015,
      "step": 80550
    },
    {
      "epoch": 8.675061887848456,
      "grad_norm": 1.002301573753357,
      "learning_rate": 7.978881586035988e-05,
      "loss": 3.1908,
      "step": 80600
    },
    {
      "epoch": 8.680443439888064,
      "grad_norm": 0.8532751202583313,
      "learning_rate": 7.946557483029845e-05,
      "loss": 3.2011,
      "step": 80650
    },
    {
      "epoch": 8.685824991927673,
      "grad_norm": 0.9273635149002075,
      "learning_rate": 7.914233380023704e-05,
      "loss": 3.2184,
      "step": 80700
    },
    {
      "epoch": 8.69120654396728,
      "grad_norm": 0.9215158224105835,
      "learning_rate": 7.881909277017562e-05,
      "loss": 3.2089,
      "step": 80750
    },
    {
      "epoch": 8.696588096006888,
      "grad_norm": 0.9349377155303955,
      "learning_rate": 7.849585174011422e-05,
      "loss": 3.182,
      "step": 80800
    },
    {
      "epoch": 8.701969648046497,
      "grad_norm": 0.8928416967391968,
      "learning_rate": 7.817261071005278e-05,
      "loss": 3.2162,
      "step": 80850
    },
    {
      "epoch": 8.707351200086105,
      "grad_norm": 0.9204657673835754,
      "learning_rate": 7.784936967999138e-05,
      "loss": 3.1906,
      "step": 80900
    },
    {
      "epoch": 8.712732752125714,
      "grad_norm": 0.945881724357605,
      "learning_rate": 7.752612864992996e-05,
      "loss": 3.1924,
      "step": 80950
    },
    {
      "epoch": 8.718114304165322,
      "grad_norm": 0.9191507697105408,
      "learning_rate": 7.720288761986854e-05,
      "loss": 3.1861,
      "step": 81000
    },
    {
      "epoch": 8.718114304165322,
      "eval_accuracy": 0.39130761686079857,
      "eval_loss": 3.3245849609375,
      "eval_runtime": 184.4423,
      "eval_samples_per_second": 97.651,
      "eval_steps_per_second": 6.105,
      "step": 81000
    },
    {
      "epoch": 8.723495856204929,
      "grad_norm": 0.8935908675193787,
      "learning_rate": 7.687964658980712e-05,
      "loss": 3.1751,
      "step": 81050
    },
    {
      "epoch": 8.728877408244538,
      "grad_norm": 0.9061697721481323,
      "learning_rate": 7.655640555974572e-05,
      "loss": 3.2033,
      "step": 81100
    },
    {
      "epoch": 8.734258960284146,
      "grad_norm": 0.8966718316078186,
      "learning_rate": 7.623316452968428e-05,
      "loss": 3.188,
      "step": 81150
    },
    {
      "epoch": 8.739640512323755,
      "grad_norm": 0.9048172831535339,
      "learning_rate": 7.590992349962288e-05,
      "loss": 3.1807,
      "step": 81200
    },
    {
      "epoch": 8.745022064363363,
      "grad_norm": 0.9137296080589294,
      "learning_rate": 7.558668246956147e-05,
      "loss": 3.2081,
      "step": 81250
    },
    {
      "epoch": 8.75040361640297,
      "grad_norm": 0.9856868386268616,
      "learning_rate": 7.526344143950005e-05,
      "loss": 3.1971,
      "step": 81300
    },
    {
      "epoch": 8.755785168442578,
      "grad_norm": 0.9088264107704163,
      "learning_rate": 7.494020040943862e-05,
      "loss": 3.1922,
      "step": 81350
    },
    {
      "epoch": 8.761166720482187,
      "grad_norm": 0.8949452638626099,
      "learning_rate": 7.462342419997844e-05,
      "loss": 3.1829,
      "step": 81400
    },
    {
      "epoch": 8.766548272521796,
      "grad_norm": 0.9669175744056702,
      "learning_rate": 7.430018316991704e-05,
      "loss": 3.1961,
      "step": 81450
    },
    {
      "epoch": 8.771929824561404,
      "grad_norm": 0.8602747917175293,
      "learning_rate": 7.397694213985562e-05,
      "loss": 3.1887,
      "step": 81500
    },
    {
      "epoch": 8.777311376601011,
      "grad_norm": 0.9349292516708374,
      "learning_rate": 7.36537011097942e-05,
      "loss": 3.196,
      "step": 81550
    },
    {
      "epoch": 8.78269292864062,
      "grad_norm": 0.8863019943237305,
      "learning_rate": 7.333046007973278e-05,
      "loss": 3.1857,
      "step": 81600
    },
    {
      "epoch": 8.788074480680228,
      "grad_norm": 0.904812753200531,
      "learning_rate": 7.300721904967136e-05,
      "loss": 3.2011,
      "step": 81650
    },
    {
      "epoch": 8.793456032719837,
      "grad_norm": 0.997829020023346,
      "learning_rate": 7.268397801960996e-05,
      "loss": 3.1943,
      "step": 81700
    },
    {
      "epoch": 8.798837584759445,
      "grad_norm": 0.8679459095001221,
      "learning_rate": 7.236073698954854e-05,
      "loss": 3.1768,
      "step": 81750
    },
    {
      "epoch": 8.804219136799054,
      "grad_norm": 0.8680683374404907,
      "learning_rate": 7.203749595948712e-05,
      "loss": 3.1979,
      "step": 81800
    },
    {
      "epoch": 8.80960068883866,
      "grad_norm": 0.8595784306526184,
      "learning_rate": 7.17142549294257e-05,
      "loss": 3.1817,
      "step": 81850
    },
    {
      "epoch": 8.814982240878269,
      "grad_norm": 0.9053434133529663,
      "learning_rate": 7.139101389936428e-05,
      "loss": 3.1952,
      "step": 81900
    },
    {
      "epoch": 8.820363792917878,
      "grad_norm": 0.9190144538879395,
      "learning_rate": 7.106777286930287e-05,
      "loss": 3.1968,
      "step": 81950
    },
    {
      "epoch": 8.825745344957486,
      "grad_norm": 0.9198845028877258,
      "learning_rate": 7.074453183924146e-05,
      "loss": 3.1985,
      "step": 82000
    },
    {
      "epoch": 8.825745344957486,
      "eval_accuracy": 0.3918313237800644,
      "eval_loss": 3.3202006816864014,
      "eval_runtime": 184.5351,
      "eval_samples_per_second": 97.602,
      "eval_steps_per_second": 6.102,
      "step": 82000
    },
    {
      "epoch": 8.831126896997095,
      "grad_norm": 0.904979407787323,
      "learning_rate": 7.042129080918004e-05,
      "loss": 3.2149,
      "step": 82050
    },
    {
      "epoch": 8.836508449036701,
      "grad_norm": 0.905988335609436,
      "learning_rate": 7.009804977911862e-05,
      "loss": 3.1666,
      "step": 82100
    },
    {
      "epoch": 8.84189000107631,
      "grad_norm": 0.8676588535308838,
      "learning_rate": 6.97748087490572e-05,
      "loss": 3.1978,
      "step": 82150
    },
    {
      "epoch": 8.847271553115919,
      "grad_norm": 0.880484938621521,
      "learning_rate": 6.945156771899579e-05,
      "loss": 3.1812,
      "step": 82200
    },
    {
      "epoch": 8.852653105155527,
      "grad_norm": 0.9128332734107971,
      "learning_rate": 6.912832668893437e-05,
      "loss": 3.2036,
      "step": 82250
    },
    {
      "epoch": 8.858034657195136,
      "grad_norm": 0.8518000841140747,
      "learning_rate": 6.880508565887297e-05,
      "loss": 3.1843,
      "step": 82300
    },
    {
      "epoch": 8.863416209234742,
      "grad_norm": 0.8791953921318054,
      "learning_rate": 6.848184462881155e-05,
      "loss": 3.2008,
      "step": 82350
    },
    {
      "epoch": 8.868797761274351,
      "grad_norm": 0.9222840070724487,
      "learning_rate": 6.815860359875013e-05,
      "loss": 3.1771,
      "step": 82400
    },
    {
      "epoch": 8.87417931331396,
      "grad_norm": 0.9001487493515015,
      "learning_rate": 6.783536256868871e-05,
      "loss": 3.2046,
      "step": 82450
    },
    {
      "epoch": 8.879560865353568,
      "grad_norm": 0.9023095369338989,
      "learning_rate": 6.75121215386273e-05,
      "loss": 3.191,
      "step": 82500
    },
    {
      "epoch": 8.884942417393177,
      "grad_norm": 0.9238621592521667,
      "learning_rate": 6.718888050856589e-05,
      "loss": 3.2048,
      "step": 82550
    },
    {
      "epoch": 8.890323969432785,
      "grad_norm": 0.9063026905059814,
      "learning_rate": 6.686563947850447e-05,
      "loss": 3.2089,
      "step": 82600
    },
    {
      "epoch": 8.895705521472392,
      "grad_norm": 0.9377871751785278,
      "learning_rate": 6.654239844844305e-05,
      "loss": 3.1831,
      "step": 82650
    },
    {
      "epoch": 8.901087073512,
      "grad_norm": 0.9342346787452698,
      "learning_rate": 6.621915741838163e-05,
      "loss": 3.2082,
      "step": 82700
    },
    {
      "epoch": 8.906468625551609,
      "grad_norm": 0.9092092514038086,
      "learning_rate": 6.589591638832021e-05,
      "loss": 3.2056,
      "step": 82750
    },
    {
      "epoch": 8.911850177591218,
      "grad_norm": 0.8764158487319946,
      "learning_rate": 6.557267535825881e-05,
      "loss": 3.183,
      "step": 82800
    },
    {
      "epoch": 8.917231729630826,
      "grad_norm": 0.8801301717758179,
      "learning_rate": 6.524943432819739e-05,
      "loss": 3.1944,
      "step": 82850
    },
    {
      "epoch": 8.922613281670433,
      "grad_norm": 0.8484103083610535,
      "learning_rate": 6.492619329813597e-05,
      "loss": 3.1984,
      "step": 82900
    },
    {
      "epoch": 8.927994833710041,
      "grad_norm": 0.8952755928039551,
      "learning_rate": 6.460295226807455e-05,
      "loss": 3.2031,
      "step": 82950
    },
    {
      "epoch": 8.93337638574965,
      "grad_norm": 0.8987452387809753,
      "learning_rate": 6.427971123801313e-05,
      "loss": 3.1953,
      "step": 83000
    },
    {
      "epoch": 8.93337638574965,
      "eval_accuracy": 0.39194823428735276,
      "eval_loss": 3.3176770210266113,
      "eval_runtime": 184.4706,
      "eval_samples_per_second": 97.636,
      "eval_steps_per_second": 6.104,
      "step": 83000
    },
    {
      "epoch": 8.938757937789259,
      "grad_norm": 0.9518873691558838,
      "learning_rate": 6.395647020795173e-05,
      "loss": 3.2081,
      "step": 83050
    },
    {
      "epoch": 8.944139489828867,
      "grad_norm": 0.9211958646774292,
      "learning_rate": 6.363322917789031e-05,
      "loss": 3.1907,
      "step": 83100
    },
    {
      "epoch": 8.949521041868476,
      "grad_norm": 0.964308500289917,
      "learning_rate": 6.330998814782889e-05,
      "loss": 3.2036,
      "step": 83150
    },
    {
      "epoch": 8.954902593908082,
      "grad_norm": 0.9281810522079468,
      "learning_rate": 6.298674711776748e-05,
      "loss": 3.1917,
      "step": 83200
    },
    {
      "epoch": 8.960284145947691,
      "grad_norm": 0.898323655128479,
      "learning_rate": 6.266350608770606e-05,
      "loss": 3.173,
      "step": 83250
    },
    {
      "epoch": 8.9656656979873,
      "grad_norm": 0.9448835849761963,
      "learning_rate": 6.234026505764465e-05,
      "loss": 3.2009,
      "step": 83300
    },
    {
      "epoch": 8.971047250026908,
      "grad_norm": 0.8978157043457031,
      "learning_rate": 6.201702402758323e-05,
      "loss": 3.1917,
      "step": 83350
    },
    {
      "epoch": 8.976428802066517,
      "grad_norm": 0.8895567655563354,
      "learning_rate": 6.169378299752181e-05,
      "loss": 3.1955,
      "step": 83400
    },
    {
      "epoch": 8.981810354106123,
      "grad_norm": 0.9051655530929565,
      "learning_rate": 6.13705419674604e-05,
      "loss": 3.195,
      "step": 83450
    },
    {
      "epoch": 8.987191906145732,
      "grad_norm": 0.8990358710289001,
      "learning_rate": 6.104730093739898e-05,
      "loss": 3.1953,
      "step": 83500
    },
    {
      "epoch": 8.99257345818534,
      "grad_norm": 0.8926938772201538,
      "learning_rate": 6.072405990733757e-05,
      "loss": 3.1902,
      "step": 83550
    },
    {
      "epoch": 8.997955010224949,
      "grad_norm": 0.8884627819061279,
      "learning_rate": 6.040081887727615e-05,
      "loss": 3.1765,
      "step": 83600
    },
    {
      "epoch": 9.003336562264558,
      "grad_norm": 0.9148643016815186,
      "learning_rate": 6.007757784721473e-05,
      "loss": 3.1392,
      "step": 83650
    },
    {
      "epoch": 9.008718114304166,
      "grad_norm": 0.893187940120697,
      "learning_rate": 5.975433681715332e-05,
      "loss": 3.1153,
      "step": 83700
    },
    {
      "epoch": 9.014099666343773,
      "grad_norm": 0.944554328918457,
      "learning_rate": 5.94310957870919e-05,
      "loss": 3.1489,
      "step": 83750
    },
    {
      "epoch": 9.019481218383381,
      "grad_norm": 0.9141628742218018,
      "learning_rate": 5.910785475703049e-05,
      "loss": 3.1336,
      "step": 83800
    },
    {
      "epoch": 9.02486277042299,
      "grad_norm": 0.9180608987808228,
      "learning_rate": 5.878461372696907e-05,
      "loss": 3.1426,
      "step": 83850
    },
    {
      "epoch": 9.030244322462599,
      "grad_norm": 0.8893204927444458,
      "learning_rate": 5.846137269690765e-05,
      "loss": 3.1251,
      "step": 83900
    },
    {
      "epoch": 9.035625874502207,
      "grad_norm": 0.9035342335700989,
      "learning_rate": 5.813813166684624e-05,
      "loss": 3.1153,
      "step": 83950
    },
    {
      "epoch": 9.041007426541814,
      "grad_norm": 0.939935564994812,
      "learning_rate": 5.781489063678482e-05,
      "loss": 3.1484,
      "step": 84000
    },
    {
      "epoch": 9.041007426541814,
      "eval_accuracy": 0.3919102057766177,
      "eval_loss": 3.320542573928833,
      "eval_runtime": 184.4613,
      "eval_samples_per_second": 97.641,
      "eval_steps_per_second": 6.104,
      "step": 84000
    },
    {
      "epoch": 9.046388978581422,
      "grad_norm": 0.9235522747039795,
      "learning_rate": 5.749164960672341e-05,
      "loss": 3.137,
      "step": 84050
    },
    {
      "epoch": 9.051770530621031,
      "grad_norm": 0.9166929721832275,
      "learning_rate": 5.716840857666199e-05,
      "loss": 3.1365,
      "step": 84100
    },
    {
      "epoch": 9.05715208266064,
      "grad_norm": 0.9141976237297058,
      "learning_rate": 5.684516754660057e-05,
      "loss": 3.1199,
      "step": 84150
    },
    {
      "epoch": 9.062533634700248,
      "grad_norm": 0.9272394776344299,
      "learning_rate": 5.652192651653916e-05,
      "loss": 3.1371,
      "step": 84200
    },
    {
      "epoch": 9.067915186739857,
      "grad_norm": 0.9308253526687622,
      "learning_rate": 5.619868548647774e-05,
      "loss": 3.1444,
      "step": 84250
    },
    {
      "epoch": 9.073296738779463,
      "grad_norm": 0.9347401857376099,
      "learning_rate": 5.5875444456416335e-05,
      "loss": 3.1211,
      "step": 84300
    },
    {
      "epoch": 9.078678290819072,
      "grad_norm": 0.9167459607124329,
      "learning_rate": 5.5552203426354916e-05,
      "loss": 3.1339,
      "step": 84350
    },
    {
      "epoch": 9.08405984285868,
      "grad_norm": 0.9475894570350647,
      "learning_rate": 5.52289623962935e-05,
      "loss": 3.1491,
      "step": 84400
    },
    {
      "epoch": 9.089441394898289,
      "grad_norm": 0.8984004259109497,
      "learning_rate": 5.4905721366232085e-05,
      "loss": 3.145,
      "step": 84450
    },
    {
      "epoch": 9.094822946937898,
      "grad_norm": 0.8978100419044495,
      "learning_rate": 5.4582480336170666e-05,
      "loss": 3.1325,
      "step": 84500
    },
    {
      "epoch": 9.100204498977504,
      "grad_norm": 0.9271031022071838,
      "learning_rate": 5.4259239306109254e-05,
      "loss": 3.143,
      "step": 84550
    },
    {
      "epoch": 9.105586051017113,
      "grad_norm": 1.0150744915008545,
      "learning_rate": 5.3935998276047835e-05,
      "loss": 3.144,
      "step": 84600
    },
    {
      "epoch": 9.110967603056721,
      "grad_norm": 0.9482972621917725,
      "learning_rate": 5.3612757245986416e-05,
      "loss": 3.1343,
      "step": 84650
    },
    {
      "epoch": 9.11634915509633,
      "grad_norm": 0.9138357639312744,
      "learning_rate": 5.3289516215925004e-05,
      "loss": 3.1334,
      "step": 84700
    },
    {
      "epoch": 9.121730707135939,
      "grad_norm": 0.9229753017425537,
      "learning_rate": 5.2966275185863585e-05,
      "loss": 3.1331,
      "step": 84750
    },
    {
      "epoch": 9.127112259175545,
      "grad_norm": 0.9260730147361755,
      "learning_rate": 5.264303415580217e-05,
      "loss": 3.1471,
      "step": 84800
    },
    {
      "epoch": 9.132493811215154,
      "grad_norm": 0.9400712847709656,
      "learning_rate": 5.2319793125740754e-05,
      "loss": 3.143,
      "step": 84850
    },
    {
      "epoch": 9.137875363254762,
      "grad_norm": 0.8856228590011597,
      "learning_rate": 5.1996552095679336e-05,
      "loss": 3.1429,
      "step": 84900
    },
    {
      "epoch": 9.143256915294371,
      "grad_norm": 0.8993318676948547,
      "learning_rate": 5.1673311065617923e-05,
      "loss": 3.1323,
      "step": 84950
    },
    {
      "epoch": 9.14863846733398,
      "grad_norm": 0.9230866432189941,
      "learning_rate": 5.1350070035556505e-05,
      "loss": 3.1263,
      "step": 85000
    },
    {
      "epoch": 9.14863846733398,
      "eval_accuracy": 0.39233319146887946,
      "eval_loss": 3.318763494491577,
      "eval_runtime": 184.5764,
      "eval_samples_per_second": 97.58,
      "eval_steps_per_second": 6.1,
      "step": 85000
    },
    {
      "epoch": 9.154020019373588,
      "grad_norm": 0.9071688652038574,
      "learning_rate": 5.102682900549509e-05,
      "loss": 3.1419,
      "step": 85050
    },
    {
      "epoch": 9.159401571413195,
      "grad_norm": 0.9356689453125,
      "learning_rate": 5.0703587975433674e-05,
      "loss": 3.1345,
      "step": 85100
    },
    {
      "epoch": 9.164783123452803,
      "grad_norm": 0.9222356677055359,
      "learning_rate": 5.0380346945372255e-05,
      "loss": 3.1289,
      "step": 85150
    },
    {
      "epoch": 9.170164675492412,
      "grad_norm": 0.9154790639877319,
      "learning_rate": 5.005710591531085e-05,
      "loss": 3.149,
      "step": 85200
    },
    {
      "epoch": 9.17554622753202,
      "grad_norm": 0.960292398929596,
      "learning_rate": 4.973386488524943e-05,
      "loss": 3.1448,
      "step": 85250
    },
    {
      "epoch": 9.180927779571629,
      "grad_norm": 0.9049267768859863,
      "learning_rate": 4.941062385518802e-05,
      "loss": 3.151,
      "step": 85300
    },
    {
      "epoch": 9.186309331611236,
      "grad_norm": 0.9267685413360596,
      "learning_rate": 4.90873828251266e-05,
      "loss": 3.1571,
      "step": 85350
    },
    {
      "epoch": 9.191690883650844,
      "grad_norm": 0.944338858127594,
      "learning_rate": 4.876414179506518e-05,
      "loss": 3.1312,
      "step": 85400
    },
    {
      "epoch": 9.197072435690453,
      "grad_norm": 0.9315845966339111,
      "learning_rate": 4.844090076500377e-05,
      "loss": 3.1364,
      "step": 85450
    },
    {
      "epoch": 9.202453987730062,
      "grad_norm": 0.9142685532569885,
      "learning_rate": 4.811765973494235e-05,
      "loss": 3.1221,
      "step": 85500
    },
    {
      "epoch": 9.20783553976967,
      "grad_norm": 0.9636183977127075,
      "learning_rate": 4.779441870488094e-05,
      "loss": 3.1489,
      "step": 85550
    },
    {
      "epoch": 9.213217091809279,
      "grad_norm": 0.9077907204627991,
      "learning_rate": 4.747117767481952e-05,
      "loss": 3.1571,
      "step": 85600
    },
    {
      "epoch": 9.218598643848885,
      "grad_norm": 0.9051309823989868,
      "learning_rate": 4.71479366447581e-05,
      "loss": 3.1413,
      "step": 85650
    },
    {
      "epoch": 9.223980195888494,
      "grad_norm": 0.9093771576881409,
      "learning_rate": 4.682469561469669e-05,
      "loss": 3.1384,
      "step": 85700
    },
    {
      "epoch": 9.229361747928102,
      "grad_norm": 0.9680438041687012,
      "learning_rate": 4.650145458463527e-05,
      "loss": 3.1292,
      "step": 85750
    },
    {
      "epoch": 9.234743299967711,
      "grad_norm": 0.9250264763832092,
      "learning_rate": 4.6178213554573856e-05,
      "loss": 3.1349,
      "step": 85800
    },
    {
      "epoch": 9.24012485200732,
      "grad_norm": 0.9208278656005859,
      "learning_rate": 4.585497252451244e-05,
      "loss": 3.1461,
      "step": 85850
    },
    {
      "epoch": 9.245506404046926,
      "grad_norm": 0.9370331764221191,
      "learning_rate": 4.553173149445102e-05,
      "loss": 3.1365,
      "step": 85900
    },
    {
      "epoch": 9.250887956086535,
      "grad_norm": 0.9302545785903931,
      "learning_rate": 4.520849046438961e-05,
      "loss": 3.1415,
      "step": 85950
    },
    {
      "epoch": 9.256269508126143,
      "grad_norm": 0.9758551716804504,
      "learning_rate": 4.488524943432819e-05,
      "loss": 3.1477,
      "step": 86000
    },
    {
      "epoch": 9.256269508126143,
      "eval_accuracy": 0.3924128340356475,
      "eval_loss": 3.317904472351074,
      "eval_runtime": 184.5327,
      "eval_samples_per_second": 97.603,
      "eval_steps_per_second": 6.102,
      "step": 86000
    },
    {
      "epoch": 9.261651060165752,
      "grad_norm": 0.9121786952018738,
      "learning_rate": 4.456200840426678e-05,
      "loss": 3.1417,
      "step": 86050
    },
    {
      "epoch": 9.26703261220536,
      "grad_norm": 0.9119290709495544,
      "learning_rate": 4.4238767374205363e-05,
      "loss": 3.1393,
      "step": 86100
    },
    {
      "epoch": 9.272414164244967,
      "grad_norm": 0.9074670672416687,
      "learning_rate": 4.3915526344143945e-05,
      "loss": 3.1474,
      "step": 86150
    },
    {
      "epoch": 9.277795716284576,
      "grad_norm": 0.9760359525680542,
      "learning_rate": 4.359228531408253e-05,
      "loss": 3.1323,
      "step": 86200
    },
    {
      "epoch": 9.283177268324184,
      "grad_norm": 0.9685165882110596,
      "learning_rate": 4.3269044284021114e-05,
      "loss": 3.1524,
      "step": 86250
    },
    {
      "epoch": 9.288558820363793,
      "grad_norm": 0.9074246287345886,
      "learning_rate": 4.29458032539597e-05,
      "loss": 3.1301,
      "step": 86300
    },
    {
      "epoch": 9.293940372403402,
      "grad_norm": 0.926539957523346,
      "learning_rate": 4.262256222389828e-05,
      "loss": 3.136,
      "step": 86350
    },
    {
      "epoch": 9.29932192444301,
      "grad_norm": 0.8760659694671631,
      "learning_rate": 4.229932119383687e-05,
      "loss": 3.1329,
      "step": 86400
    },
    {
      "epoch": 9.304703476482617,
      "grad_norm": 0.9046364426612854,
      "learning_rate": 4.197608016377545e-05,
      "loss": 3.1383,
      "step": 86450
    },
    {
      "epoch": 9.310085028522225,
      "grad_norm": 0.9421766400337219,
      "learning_rate": 4.165283913371403e-05,
      "loss": 3.1468,
      "step": 86500
    },
    {
      "epoch": 9.315466580561834,
      "grad_norm": 0.9465281963348389,
      "learning_rate": 4.132959810365262e-05,
      "loss": 3.1496,
      "step": 86550
    },
    {
      "epoch": 9.320848132601443,
      "grad_norm": 0.9032894372940063,
      "learning_rate": 4.10063570735912e-05,
      "loss": 3.1509,
      "step": 86600
    },
    {
      "epoch": 9.326229684641051,
      "grad_norm": 0.937605082988739,
      "learning_rate": 4.068311604352979e-05,
      "loss": 3.1387,
      "step": 86650
    },
    {
      "epoch": 9.331611236680658,
      "grad_norm": 0.9383586049079895,
      "learning_rate": 4.035987501346837e-05,
      "loss": 3.1366,
      "step": 86700
    },
    {
      "epoch": 9.336992788720266,
      "grad_norm": 0.9010090827941895,
      "learning_rate": 4.003663398340695e-05,
      "loss": 3.1429,
      "step": 86750
    },
    {
      "epoch": 9.342374340759875,
      "grad_norm": 0.9194101095199585,
      "learning_rate": 3.9713392953345546e-05,
      "loss": 3.1345,
      "step": 86800
    },
    {
      "epoch": 9.347755892799483,
      "grad_norm": 0.9066176414489746,
      "learning_rate": 3.939015192328413e-05,
      "loss": 3.1471,
      "step": 86850
    },
    {
      "epoch": 9.353137444839092,
      "grad_norm": 0.96162348985672,
      "learning_rate": 3.9066910893222715e-05,
      "loss": 3.153,
      "step": 86900
    },
    {
      "epoch": 9.3585189968787,
      "grad_norm": 0.9190367460250854,
      "learning_rate": 3.8743669863161296e-05,
      "loss": 3.1421,
      "step": 86950
    },
    {
      "epoch": 9.363900548918307,
      "grad_norm": 0.963890552520752,
      "learning_rate": 3.842042883309988e-05,
      "loss": 3.1468,
      "step": 87000
    },
    {
      "epoch": 9.363900548918307,
      "eval_accuracy": 0.39295935806135435,
      "eval_loss": 3.314394950866699,
      "eval_runtime": 184.5488,
      "eval_samples_per_second": 97.595,
      "eval_steps_per_second": 6.101,
      "step": 87000
    },
    {
      "epoch": 9.369282100957916,
      "grad_norm": 0.9209648370742798,
      "learning_rate": 3.8097187803038465e-05,
      "loss": 3.1471,
      "step": 87050
    },
    {
      "epoch": 9.374663652997524,
      "grad_norm": 0.9377780556678772,
      "learning_rate": 3.7773946772977047e-05,
      "loss": 3.1486,
      "step": 87100
    },
    {
      "epoch": 9.380045205037133,
      "grad_norm": 0.9694614410400391,
      "learning_rate": 3.745070574291563e-05,
      "loss": 3.142,
      "step": 87150
    },
    {
      "epoch": 9.385426757076742,
      "grad_norm": 0.9046480655670166,
      "learning_rate": 3.7127464712854216e-05,
      "loss": 3.1398,
      "step": 87200
    },
    {
      "epoch": 9.390808309116348,
      "grad_norm": 0.9260925054550171,
      "learning_rate": 3.6804223682792803e-05,
      "loss": 3.1306,
      "step": 87250
    },
    {
      "epoch": 9.396189861155957,
      "grad_norm": 0.9350437521934509,
      "learning_rate": 3.6480982652731385e-05,
      "loss": 3.153,
      "step": 87300
    },
    {
      "epoch": 9.401571413195565,
      "grad_norm": 0.9208142161369324,
      "learning_rate": 3.6157741622669966e-05,
      "loss": 3.1422,
      "step": 87350
    },
    {
      "epoch": 9.406952965235174,
      "grad_norm": 0.9761648178100586,
      "learning_rate": 3.5847430233811004e-05,
      "loss": 3.147,
      "step": 87400
    },
    {
      "epoch": 9.412334517274783,
      "grad_norm": 0.9446584582328796,
      "learning_rate": 3.552418920374959e-05,
      "loss": 3.1529,
      "step": 87450
    },
    {
      "epoch": 9.417716069314391,
      "grad_norm": 0.9459024667739868,
      "learning_rate": 3.520094817368818e-05,
      "loss": 3.157,
      "step": 87500
    },
    {
      "epoch": 9.423097621353998,
      "grad_norm": 0.9242775440216064,
      "learning_rate": 3.487770714362676e-05,
      "loss": 3.144,
      "step": 87550
    },
    {
      "epoch": 9.428479173393606,
      "grad_norm": 0.9069784879684448,
      "learning_rate": 3.455446611356535e-05,
      "loss": 3.1446,
      "step": 87600
    },
    {
      "epoch": 9.433860725433215,
      "grad_norm": 0.9493717551231384,
      "learning_rate": 3.423122508350393e-05,
      "loss": 3.1438,
      "step": 87650
    },
    {
      "epoch": 9.439242277472824,
      "grad_norm": 0.9008575081825256,
      "learning_rate": 3.390798405344251e-05,
      "loss": 3.1351,
      "step": 87700
    },
    {
      "epoch": 9.444623829512432,
      "grad_norm": 0.955755352973938,
      "learning_rate": 3.35847430233811e-05,
      "loss": 3.1579,
      "step": 87750
    },
    {
      "epoch": 9.450005381552039,
      "grad_norm": 0.9249664545059204,
      "learning_rate": 3.326150199331968e-05,
      "loss": 3.153,
      "step": 87800
    },
    {
      "epoch": 9.455386933591647,
      "grad_norm": 0.8965539336204529,
      "learning_rate": 3.293826096325827e-05,
      "loss": 3.1373,
      "step": 87850
    },
    {
      "epoch": 9.460768485631256,
      "grad_norm": 0.9269651770591736,
      "learning_rate": 3.261501993319685e-05,
      "loss": 3.1488,
      "step": 87900
    },
    {
      "epoch": 9.466150037670864,
      "grad_norm": 0.956444501876831,
      "learning_rate": 3.229177890313544e-05,
      "loss": 3.1615,
      "step": 87950
    },
    {
      "epoch": 9.471531589710473,
      "grad_norm": 0.9170980453491211,
      "learning_rate": 3.196853787307402e-05,
      "loss": 3.1468,
      "step": 88000
    },
    {
      "epoch": 9.471531589710473,
      "eval_accuracy": 0.39307724644463304,
      "eval_loss": 3.3124024868011475,
      "eval_runtime": 184.3687,
      "eval_samples_per_second": 97.69,
      "eval_steps_per_second": 6.107,
      "step": 88000
    },
    {
      "epoch": 9.476913141750082,
      "grad_norm": 0.9016711115837097,
      "learning_rate": 3.1645296843012606e-05,
      "loss": 3.1389,
      "step": 88050
    },
    {
      "epoch": 9.482294693789688,
      "grad_norm": 0.9824627041816711,
      "learning_rate": 3.132852063355242e-05,
      "loss": 3.1424,
      "step": 88100
    },
    {
      "epoch": 9.487676245829297,
      "grad_norm": 0.909274697303772,
      "learning_rate": 3.1005279603491e-05,
      "loss": 3.1389,
      "step": 88150
    },
    {
      "epoch": 9.493057797868905,
      "grad_norm": 0.9372968077659607,
      "learning_rate": 3.068203857342958e-05,
      "loss": 3.1531,
      "step": 88200
    },
    {
      "epoch": 9.498439349908514,
      "grad_norm": 0.9175796508789062,
      "learning_rate": 3.035879754336817e-05,
      "loss": 3.1525,
      "step": 88250
    },
    {
      "epoch": 9.503820901948123,
      "grad_norm": 0.9364547729492188,
      "learning_rate": 3.0035556513306754e-05,
      "loss": 3.1401,
      "step": 88300
    },
    {
      "epoch": 9.50920245398773,
      "grad_norm": 0.9171380400657654,
      "learning_rate": 2.971231548324534e-05,
      "loss": 3.134,
      "step": 88350
    },
    {
      "epoch": 9.514584006027338,
      "grad_norm": 0.9604135155677795,
      "learning_rate": 2.938907445318392e-05,
      "loss": 3.1486,
      "step": 88400
    },
    {
      "epoch": 9.519965558066946,
      "grad_norm": 0.9229456782341003,
      "learning_rate": 2.9065833423122504e-05,
      "loss": 3.1414,
      "step": 88450
    },
    {
      "epoch": 9.525347110106555,
      "grad_norm": 0.8929281830787659,
      "learning_rate": 2.874259239306109e-05,
      "loss": 3.1414,
      "step": 88500
    },
    {
      "epoch": 9.530728662146164,
      "grad_norm": 0.9471603631973267,
      "learning_rate": 2.8419351362999673e-05,
      "loss": 3.1446,
      "step": 88550
    },
    {
      "epoch": 9.536110214185772,
      "grad_norm": 0.9738565683364868,
      "learning_rate": 2.809611033293826e-05,
      "loss": 3.1476,
      "step": 88600
    },
    {
      "epoch": 9.541491766225379,
      "grad_norm": 0.9245936870574951,
      "learning_rate": 2.7772869302876845e-05,
      "loss": 3.1497,
      "step": 88650
    },
    {
      "epoch": 9.546873318264987,
      "grad_norm": 0.9588368535041809,
      "learning_rate": 2.7449628272815427e-05,
      "loss": 3.1421,
      "step": 88700
    },
    {
      "epoch": 9.552254870304596,
      "grad_norm": 0.9462993741035461,
      "learning_rate": 2.712638724275401e-05,
      "loss": 3.1708,
      "step": 88750
    },
    {
      "epoch": 9.557636422344205,
      "grad_norm": 0.9054816961288452,
      "learning_rate": 2.6803146212692596e-05,
      "loss": 3.1294,
      "step": 88800
    },
    {
      "epoch": 9.563017974383813,
      "grad_norm": 0.9444681406021118,
      "learning_rate": 2.647990518263118e-05,
      "loss": 3.1341,
      "step": 88850
    },
    {
      "epoch": 9.56839952642342,
      "grad_norm": 0.9182373881340027,
      "learning_rate": 2.6156664152569765e-05,
      "loss": 3.1468,
      "step": 88900
    },
    {
      "epoch": 9.573781078463028,
      "grad_norm": 0.9208268523216248,
      "learning_rate": 2.5833423122508346e-05,
      "loss": 3.1222,
      "step": 88950
    },
    {
      "epoch": 9.579162630502637,
      "grad_norm": 0.9134917259216309,
      "learning_rate": 2.5510182092446934e-05,
      "loss": 3.1367,
      "step": 89000
    },
    {
      "epoch": 9.579162630502637,
      "eval_accuracy": 0.3933979897694614,
      "eval_loss": 3.3100452423095703,
      "eval_runtime": 184.628,
      "eval_samples_per_second": 97.553,
      "eval_steps_per_second": 6.099,
      "step": 89000
    },
    {
      "epoch": 9.584544182542245,
      "grad_norm": 0.9227930307388306,
      "learning_rate": 2.5186941062385518e-05,
      "loss": 3.1234,
      "step": 89050
    },
    {
      "epoch": 9.589925734581854,
      "grad_norm": 0.9176948666572571,
      "learning_rate": 2.4863700032324103e-05,
      "loss": 3.1215,
      "step": 89100
    },
    {
      "epoch": 9.59530728662146,
      "grad_norm": 0.9101543426513672,
      "learning_rate": 2.4540459002262687e-05,
      "loss": 3.1434,
      "step": 89150
    },
    {
      "epoch": 9.60068883866107,
      "grad_norm": 0.9386678338050842,
      "learning_rate": 2.4217217972201268e-05,
      "loss": 3.1492,
      "step": 89200
    },
    {
      "epoch": 9.606070390700678,
      "grad_norm": 0.9272826910018921,
      "learning_rate": 2.3893976942139853e-05,
      "loss": 3.1449,
      "step": 89250
    },
    {
      "epoch": 9.611451942740286,
      "grad_norm": 0.9256073832511902,
      "learning_rate": 2.3570735912078437e-05,
      "loss": 3.1472,
      "step": 89300
    },
    {
      "epoch": 9.616833494779895,
      "grad_norm": 0.8588914275169373,
      "learning_rate": 2.324749488201702e-05,
      "loss": 3.1264,
      "step": 89350
    },
    {
      "epoch": 9.622215046819504,
      "grad_norm": 0.9573777914047241,
      "learning_rate": 2.292425385195561e-05,
      "loss": 3.1268,
      "step": 89400
    },
    {
      "epoch": 9.62759659885911,
      "grad_norm": 0.9397759437561035,
      "learning_rate": 2.260101282189419e-05,
      "loss": 3.1447,
      "step": 89450
    },
    {
      "epoch": 9.632978150898719,
      "grad_norm": 0.940040647983551,
      "learning_rate": 2.2277771791832775e-05,
      "loss": 3.1573,
      "step": 89500
    },
    {
      "epoch": 9.638359702938327,
      "grad_norm": 0.9524548649787903,
      "learning_rate": 2.195453076177136e-05,
      "loss": 3.1436,
      "step": 89550
    },
    {
      "epoch": 9.643741254977936,
      "grad_norm": 0.9198241233825684,
      "learning_rate": 2.1631289731709944e-05,
      "loss": 3.1404,
      "step": 89600
    },
    {
      "epoch": 9.649122807017545,
      "grad_norm": 0.9470341801643372,
      "learning_rate": 2.130804870164853e-05,
      "loss": 3.1588,
      "step": 89650
    },
    {
      "epoch": 9.654504359057151,
      "grad_norm": 0.9101405143737793,
      "learning_rate": 2.098480767158711e-05,
      "loss": 3.1406,
      "step": 89700
    },
    {
      "epoch": 9.65988591109676,
      "grad_norm": 0.929602861404419,
      "learning_rate": 2.0661566641525694e-05,
      "loss": 3.1384,
      "step": 89750
    },
    {
      "epoch": 9.665267463136368,
      "grad_norm": 0.9006006121635437,
      "learning_rate": 2.0338325611464282e-05,
      "loss": 3.1493,
      "step": 89800
    },
    {
      "epoch": 9.670649015175977,
      "grad_norm": 0.9310491681098938,
      "learning_rate": 2.0015084581402867e-05,
      "loss": 3.1326,
      "step": 89850
    },
    {
      "epoch": 9.676030567215586,
      "grad_norm": 0.9572315812110901,
      "learning_rate": 1.969184355134145e-05,
      "loss": 3.1229,
      "step": 89900
    },
    {
      "epoch": 9.681412119255192,
      "grad_norm": 0.9354777336120605,
      "learning_rate": 1.9368602521280032e-05,
      "loss": 3.133,
      "step": 89950
    },
    {
      "epoch": 9.6867936712948,
      "grad_norm": 0.9717013239860535,
      "learning_rate": 1.9045361491218617e-05,
      "loss": 3.1556,
      "step": 90000
    },
    {
      "epoch": 9.6867936712948,
      "eval_accuracy": 0.393602039892777,
      "eval_loss": 3.308566093444824,
      "eval_runtime": 184.5423,
      "eval_samples_per_second": 97.598,
      "eval_steps_per_second": 6.102,
      "step": 90000
    },
    {
      "epoch": 9.69217522333441,
      "grad_norm": 0.8886333703994751,
      "learning_rate": 1.87221204611572e-05,
      "loss": 3.1189,
      "step": 90050
    },
    {
      "epoch": 9.697556775374018,
      "grad_norm": 0.9751870036125183,
      "learning_rate": 1.8398879431095786e-05,
      "loss": 3.1628,
      "step": 90100
    },
    {
      "epoch": 9.702938327413626,
      "grad_norm": 0.9070645570755005,
      "learning_rate": 1.807563840103437e-05,
      "loss": 3.1287,
      "step": 90150
    },
    {
      "epoch": 9.708319879453235,
      "grad_norm": 0.951580822467804,
      "learning_rate": 1.7752397370972955e-05,
      "loss": 3.1543,
      "step": 90200
    },
    {
      "epoch": 9.713701431492842,
      "grad_norm": 0.9143083095550537,
      "learning_rate": 1.742915634091154e-05,
      "loss": 3.1521,
      "step": 90250
    },
    {
      "epoch": 9.71908298353245,
      "grad_norm": 0.9132610559463501,
      "learning_rate": 1.7105915310850124e-05,
      "loss": 3.1346,
      "step": 90300
    },
    {
      "epoch": 9.724464535572059,
      "grad_norm": 0.9471712112426758,
      "learning_rate": 1.6782674280788708e-05,
      "loss": 3.1427,
      "step": 90350
    },
    {
      "epoch": 9.729846087611667,
      "grad_norm": 0.9312596917152405,
      "learning_rate": 1.645943325072729e-05,
      "loss": 3.1341,
      "step": 90400
    },
    {
      "epoch": 9.735227639651276,
      "grad_norm": 0.9075988531112671,
      "learning_rate": 1.6136192220665877e-05,
      "loss": 3.1347,
      "step": 90450
    },
    {
      "epoch": 9.740609191690883,
      "grad_norm": 0.9479326009750366,
      "learning_rate": 1.5812951190604458e-05,
      "loss": 3.1556,
      "step": 90500
    },
    {
      "epoch": 9.745990743730491,
      "grad_norm": 0.9205537438392639,
      "learning_rate": 1.5489710160543043e-05,
      "loss": 3.1557,
      "step": 90550
    },
    {
      "epoch": 9.7513722957701,
      "grad_norm": 0.9379699230194092,
      "learning_rate": 1.5166469130481629e-05,
      "loss": 3.1403,
      "step": 90600
    },
    {
      "epoch": 9.756753847809708,
      "grad_norm": 0.9495012164115906,
      "learning_rate": 1.4843228100420212e-05,
      "loss": 3.1463,
      "step": 90650
    },
    {
      "epoch": 9.762135399849317,
      "grad_norm": 0.9626738429069519,
      "learning_rate": 1.4519987070358796e-05,
      "loss": 3.1378,
      "step": 90700
    },
    {
      "epoch": 9.767516951888926,
      "grad_norm": 0.9672257304191589,
      "learning_rate": 1.419674604029738e-05,
      "loss": 3.1392,
      "step": 90750
    },
    {
      "epoch": 9.772898503928532,
      "grad_norm": 0.9604204297065735,
      "learning_rate": 1.3873505010235965e-05,
      "loss": 3.1242,
      "step": 90800
    },
    {
      "epoch": 9.77828005596814,
      "grad_norm": 0.9456198811531067,
      "learning_rate": 1.355026398017455e-05,
      "loss": 3.1465,
      "step": 90850
    },
    {
      "epoch": 9.78366160800775,
      "grad_norm": 0.8965287208557129,
      "learning_rate": 1.3227022950113132e-05,
      "loss": 3.1389,
      "step": 90900
    },
    {
      "epoch": 9.789043160047358,
      "grad_norm": 0.9233707189559937,
      "learning_rate": 1.2903781920051719e-05,
      "loss": 3.1295,
      "step": 90950
    },
    {
      "epoch": 9.794424712086967,
      "grad_norm": 0.9338093400001526,
      "learning_rate": 1.2580540889990301e-05,
      "loss": 3.137,
      "step": 91000
    },
    {
      "epoch": 9.794424712086967,
      "eval_accuracy": 0.3938130438009127,
      "eval_loss": 3.306241035461426,
      "eval_runtime": 184.6512,
      "eval_samples_per_second": 97.541,
      "eval_steps_per_second": 6.098,
      "step": 91000
    },
    {
      "epoch": 9.799806264126573,
      "grad_norm": 0.9804956316947937,
      "learning_rate": 1.2257299859928886e-05,
      "loss": 3.1196,
      "step": 91050
    },
    {
      "epoch": 9.805187816166182,
      "grad_norm": 0.9253054261207581,
      "learning_rate": 1.193405882986747e-05,
      "loss": 3.1388,
      "step": 91100
    },
    {
      "epoch": 9.81056936820579,
      "grad_norm": 0.9409234523773193,
      "learning_rate": 1.1610817799806053e-05,
      "loss": 3.1378,
      "step": 91150
    },
    {
      "epoch": 9.815950920245399,
      "grad_norm": 0.9319542050361633,
      "learning_rate": 1.128757676974464e-05,
      "loss": 3.1361,
      "step": 91200
    },
    {
      "epoch": 9.821332472285007,
      "grad_norm": 0.9237921833992004,
      "learning_rate": 1.0964335739683222e-05,
      "loss": 3.1315,
      "step": 91250
    },
    {
      "epoch": 9.826714024324616,
      "grad_norm": 0.9368856549263,
      "learning_rate": 1.0641094709621807e-05,
      "loss": 3.145,
      "step": 91300
    },
    {
      "epoch": 9.832095576364223,
      "grad_norm": 0.9425742626190186,
      "learning_rate": 1.0317853679560393e-05,
      "loss": 3.1422,
      "step": 91350
    },
    {
      "epoch": 9.837477128403831,
      "grad_norm": 0.9450698494911194,
      "learning_rate": 9.994612649498976e-06,
      "loss": 3.1466,
      "step": 91400
    },
    {
      "epoch": 9.84285868044344,
      "grad_norm": 0.9163123369216919,
      "learning_rate": 9.67137161943756e-06,
      "loss": 3.1353,
      "step": 91450
    },
    {
      "epoch": 9.848240232483048,
      "grad_norm": 0.948509693145752,
      "learning_rate": 9.348130589376145e-06,
      "loss": 3.1439,
      "step": 91500
    },
    {
      "epoch": 9.853621784522657,
      "grad_norm": 0.9037002325057983,
      "learning_rate": 9.024889559314727e-06,
      "loss": 3.1362,
      "step": 91550
    },
    {
      "epoch": 9.859003336562264,
      "grad_norm": 0.9139311909675598,
      "learning_rate": 8.701648529253312e-06,
      "loss": 3.1425,
      "step": 91600
    },
    {
      "epoch": 9.864384888601872,
      "grad_norm": 1.0113749504089355,
      "learning_rate": 8.378407499191896e-06,
      "loss": 3.1521,
      "step": 91650
    },
    {
      "epoch": 9.869766440641481,
      "grad_norm": 0.9534078240394592,
      "learning_rate": 8.055166469130481e-06,
      "loss": 3.156,
      "step": 91700
    },
    {
      "epoch": 9.87514799268109,
      "grad_norm": 0.9743395447731018,
      "learning_rate": 7.731925439069065e-06,
      "loss": 3.1423,
      "step": 91750
    },
    {
      "epoch": 9.880529544720698,
      "grad_norm": 0.9121930599212646,
      "learning_rate": 7.408684409007649e-06,
      "loss": 3.1307,
      "step": 91800
    },
    {
      "epoch": 9.885911096760307,
      "grad_norm": 0.9581356048583984,
      "learning_rate": 7.085443378946234e-06,
      "loss": 3.1433,
      "step": 91850
    },
    {
      "epoch": 9.891292648799913,
      "grad_norm": 0.9198094010353088,
      "learning_rate": 6.762202348884817e-06,
      "loss": 3.1451,
      "step": 91900
    },
    {
      "epoch": 9.896674200839522,
      "grad_norm": 0.906086802482605,
      "learning_rate": 6.438961318823403e-06,
      "loss": 3.1467,
      "step": 91950
    },
    {
      "epoch": 9.90205575287913,
      "grad_norm": 0.9215152859687805,
      "learning_rate": 6.115720288761986e-06,
      "loss": 3.1342,
      "step": 92000
    },
    {
      "epoch": 9.90205575287913,
      "eval_accuracy": 0.3940013392554952,
      "eval_loss": 3.304426431655884,
      "eval_runtime": 184.5438,
      "eval_samples_per_second": 97.597,
      "eval_steps_per_second": 6.102,
      "step": 92000
    },
    {
      "epoch": 9.907437304918739,
      "grad_norm": 0.8989583849906921,
      "learning_rate": 5.792479258700571e-06,
      "loss": 3.151,
      "step": 92050
    },
    {
      "epoch": 9.912818856958348,
      "grad_norm": 0.8769867420196533,
      "learning_rate": 5.469238228639154e-06,
      "loss": 3.1253,
      "step": 92100
    },
    {
      "epoch": 9.918200408997954,
      "grad_norm": 0.9274564981460571,
      "learning_rate": 5.145997198577739e-06,
      "loss": 3.1504,
      "step": 92150
    },
    {
      "epoch": 9.923581961037563,
      "grad_norm": 0.9256285429000854,
      "learning_rate": 4.822756168516323e-06,
      "loss": 3.1404,
      "step": 92200
    },
    {
      "epoch": 9.928963513077171,
      "grad_norm": 0.8948553204536438,
      "learning_rate": 4.499515138454908e-06,
      "loss": 3.1175,
      "step": 92250
    },
    {
      "epoch": 9.93434506511678,
      "grad_norm": 0.912861704826355,
      "learning_rate": 4.1762741083934915e-06,
      "loss": 3.142,
      "step": 92300
    },
    {
      "epoch": 9.939726617156388,
      "grad_norm": 0.9803128838539124,
      "learning_rate": 3.859497898933305e-06,
      "loss": 3.15,
      "step": 92350
    },
    {
      "epoch": 9.945108169195997,
      "grad_norm": 0.9601973295211792,
      "learning_rate": 3.5362568688718884e-06,
      "loss": 3.141,
      "step": 92400
    },
    {
      "epoch": 9.950489721235604,
      "grad_norm": 0.9316169023513794,
      "learning_rate": 3.2130158388104725e-06,
      "loss": 3.15,
      "step": 92450
    },
    {
      "epoch": 9.955871273275212,
      "grad_norm": 0.9462074637413025,
      "learning_rate": 2.889774808749057e-06,
      "loss": 3.1426,
      "step": 92500
    },
    {
      "epoch": 9.961252825314821,
      "grad_norm": 0.9219121932983398,
      "learning_rate": 2.5729985992888694e-06,
      "loss": 3.1694,
      "step": 92550
    },
    {
      "epoch": 9.96663437735443,
      "grad_norm": 0.9301074147224426,
      "learning_rate": 2.249757569227454e-06,
      "loss": 3.1399,
      "step": 92600
    },
    {
      "epoch": 9.972015929394038,
      "grad_norm": 0.9258626103401184,
      "learning_rate": 1.926516539166038e-06,
      "loss": 3.1402,
      "step": 92650
    },
    {
      "epoch": 9.977397481433645,
      "grad_norm": 0.905373752117157,
      "learning_rate": 1.603275509104622e-06,
      "loss": 3.1532,
      "step": 92700
    },
    {
      "epoch": 9.982779033473253,
      "grad_norm": 0.9258732199668884,
      "learning_rate": 1.2800344790432064e-06,
      "loss": 3.1299,
      "step": 92750
    },
    {
      "epoch": 9.988160585512862,
      "grad_norm": 0.9034093022346497,
      "learning_rate": 9.567934489817906e-07,
      "loss": 3.1317,
      "step": 92800
    },
    {
      "epoch": 9.99354213755247,
      "grad_norm": 1.0189810991287231,
      "learning_rate": 6.335524189203748e-07,
      "loss": 3.1476,
      "step": 92850
    },
    {
      "epoch": 9.998923689592079,
      "grad_norm": 0.9120912551879883,
      "learning_rate": 3.103113888589591e-07,
      "loss": 3.1305,
      "step": 92900
    },
    {
      "epoch": 10.0,
      "step": 92910,
      "total_flos": 7.7681598529536e+17,
      "train_loss": 3.4636538602758975,
      "train_runtime": 80560.6393,
      "train_samples_per_second": 36.904,
      "train_steps_per_second": 1.153
    }
  ],
  "logging_steps": 50,
  "max_steps": 92910,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 10000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.7681598529536e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}