{
  "best_global_step": 31000,
  "best_metric": 0.7226839661598206,
  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-31000",
  "epoch": 2.999769248519345,
  "eval_steps": 250,
  "global_step": 39000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.003845858010922237,
      "grad_norm": 11.687983512878418,
      "learning_rate": 1.84e-05,
      "loss": 8.7035,
      "step": 50
    },
    {
      "epoch": 0.007691716021844474,
      "grad_norm": 2.6404802799224854,
      "learning_rate": 3.8400000000000005e-05,
      "loss": 2.7335,
      "step": 100
    },
    {
      "epoch": 0.01153757403276671,
      "grad_norm": 2.068481683731079,
      "learning_rate": 5.8399999999999997e-05,
      "loss": 2.0457,
      "step": 150
    },
    {
      "epoch": 0.015383432043688947,
      "grad_norm": 2.60369610786438,
      "learning_rate": 7.840000000000001e-05,
      "loss": 1.8505,
      "step": 200
    },
    {
      "epoch": 0.019229290054611183,
      "grad_norm": 1.6515765190124512,
      "learning_rate": 9.84e-05,
      "loss": 1.8158,
      "step": 250
    },
    {
      "epoch": 0.019229290054611183,
      "eval_loss": 1.8082822561264038,
      "eval_runtime": 18.1351,
      "eval_samples_per_second": 55.142,
      "eval_steps_per_second": 13.785,
      "step": 250
    },
    {
      "epoch": 0.02307514806553342,
      "grad_norm": 2.1014769077301025,
      "learning_rate": 0.0001184,
      "loss": 1.65,
      "step": 300
    },
    {
      "epoch": 0.02692100607645566,
      "grad_norm": 1.5384572744369507,
      "learning_rate": 0.0001384,
      "loss": 1.5594,
      "step": 350
    },
    {
      "epoch": 0.030766864087377895,
      "grad_norm": 1.4865778684616089,
      "learning_rate": 0.00015840000000000003,
      "loss": 1.6723,
      "step": 400
    },
    {
      "epoch": 0.03461272209830013,
      "grad_norm": 1.0966717004776,
      "learning_rate": 0.0001784,
      "loss": 1.602,
      "step": 450
    },
    {
      "epoch": 0.038458580109222366,
      "grad_norm": 1.6298224925994873,
      "learning_rate": 0.0001984,
      "loss": 1.5659,
      "step": 500
    },
    {
      "epoch": 0.038458580109222366,
      "eval_loss": 1.6280817985534668,
      "eval_runtime": 18.1181,
      "eval_samples_per_second": 55.194,
      "eval_steps_per_second": 13.798,
      "step": 500
    },
    {
      "epoch": 0.0423044381201446,
      "grad_norm": 1.2955571413040161,
      "learning_rate": 0.00019976105757992885,
      "loss": 1.6388,
      "step": 550
    },
    {
      "epoch": 0.04615029613106684,
      "grad_norm": 1.2672921419143677,
      "learning_rate": 0.00019950133755811236,
      "loss": 1.6029,
      "step": 600
    },
    {
      "epoch": 0.04999615414198908,
      "grad_norm": 1.6246057748794556,
      "learning_rate": 0.00019924161753629587,
      "loss": 1.4845,
      "step": 650
    },
    {
      "epoch": 0.05384201215291132,
      "grad_norm": 1.0235854387283325,
      "learning_rate": 0.0001989818975144794,
      "loss": 1.5184,
      "step": 700
    },
    {
      "epoch": 0.05768787016383355,
      "grad_norm": 1.5333527326583862,
      "learning_rate": 0.0001987221774926629,
      "loss": 1.6136,
      "step": 750
    },
    {
      "epoch": 0.05768787016383355,
      "eval_loss": 1.5622245073318481,
      "eval_runtime": 18.2497,
      "eval_samples_per_second": 54.795,
      "eval_steps_per_second": 13.699,
      "step": 750
    },
    {
      "epoch": 0.06153372817475579,
      "grad_norm": 1.1980785131454468,
      "learning_rate": 0.00019846245747084644,
      "loss": 1.571,
      "step": 800
    },
    {
      "epoch": 0.06537958618567802,
      "grad_norm": 1.6081124544143677,
      "learning_rate": 0.00019820273744902995,
      "loss": 1.5542,
      "step": 850
    },
    {
      "epoch": 0.06922544419660026,
      "grad_norm": 1.2620774507522583,
      "learning_rate": 0.0001979430174272135,
      "loss": 1.5513,
      "step": 900
    },
    {
      "epoch": 0.0730713022075225,
      "grad_norm": 0.8911245465278625,
      "learning_rate": 0.000197683297405397,
      "loss": 1.4516,
      "step": 950
    },
    {
      "epoch": 0.07691716021844473,
      "grad_norm": 1.2976778745651245,
      "learning_rate": 0.0001974235773835805,
      "loss": 1.5757,
      "step": 1000
    },
    {
      "epoch": 0.07691716021844473,
      "eval_loss": 1.5385061502456665,
      "eval_runtime": 18.1057,
      "eval_samples_per_second": 55.231,
      "eval_steps_per_second": 13.808,
      "step": 1000
    },
    {
      "epoch": 0.08076301822936698,
      "grad_norm": 1.451479196548462,
      "learning_rate": 0.00019716385736176403,
      "loss": 1.4287,
      "step": 1050
    },
    {
      "epoch": 0.0846088762402892,
      "grad_norm": 1.0982426404953003,
      "learning_rate": 0.00019690413733994754,
      "loss": 1.4962,
      "step": 1100
    },
    {
      "epoch": 0.08845473425121145,
      "grad_norm": 0.8899670839309692,
      "learning_rate": 0.00019664441731813106,
      "loss": 1.4675,
      "step": 1150
    },
    {
      "epoch": 0.09230059226213368,
      "grad_norm": 1.5194021463394165,
      "learning_rate": 0.0001963846972963146,
      "loss": 1.4573,
      "step": 1200
    },
    {
      "epoch": 0.09614645027305592,
      "grad_norm": 1.716470718383789,
      "learning_rate": 0.0001961249772744981,
      "loss": 1.5143,
      "step": 1250
    },
    {
      "epoch": 0.09614645027305592,
      "eval_loss": 1.5197770595550537,
      "eval_runtime": 18.2416,
      "eval_samples_per_second": 54.82,
      "eval_steps_per_second": 13.705,
      "step": 1250
    },
    {
      "epoch": 0.09999230828397816,
      "grad_norm": 1.986771583557129,
      "learning_rate": 0.00019586525725268162,
      "loss": 1.5265,
      "step": 1300
    },
    {
      "epoch": 0.10383816629490039,
      "grad_norm": 0.86269211769104,
      "learning_rate": 0.00019560553723086514,
      "loss": 1.4842,
      "step": 1350
    },
    {
      "epoch": 0.10768402430582263,
      "grad_norm": 1.187501072883606,
      "learning_rate": 0.00019534581720904865,
      "loss": 1.4054,
      "step": 1400
    },
    {
      "epoch": 0.11152988231674486,
      "grad_norm": 1.5051347017288208,
      "learning_rate": 0.00019508609718723216,
      "loss": 1.4141,
      "step": 1450
    },
    {
      "epoch": 0.1153757403276671,
      "grad_norm": 1.99917471408844,
      "learning_rate": 0.0001948263771654157,
      "loss": 1.4767,
      "step": 1500
    },
    {
      "epoch": 0.1153757403276671,
      "eval_loss": 1.4949736595153809,
      "eval_runtime": 18.0402,
      "eval_samples_per_second": 55.432,
      "eval_steps_per_second": 13.858,
      "step": 1500
    },
    {
      "epoch": 0.11922159833858934,
      "grad_norm": 1.6421241760253906,
      "learning_rate": 0.00019456665714359921,
      "loss": 1.5029,
      "step": 1550
    },
    {
      "epoch": 0.12306745634951158,
      "grad_norm": 1.8251460790634155,
      "learning_rate": 0.00019430693712178273,
      "loss": 1.4666,
      "step": 1600
    },
    {
      "epoch": 0.12691331436043382,
      "grad_norm": 1.2284319400787354,
      "learning_rate": 0.00019404721709996624,
      "loss": 1.4785,
      "step": 1650
    },
    {
      "epoch": 0.13075917237135604,
      "grad_norm": 3.7399282455444336,
      "learning_rate": 0.00019378749707814975,
      "loss": 1.5027,
      "step": 1700
    },
    {
      "epoch": 0.13460503038227828,
      "grad_norm": 1.2193188667297363,
      "learning_rate": 0.0001935277770563333,
      "loss": 1.4944,
      "step": 1750
    },
    {
      "epoch": 0.13460503038227828,
      "eval_loss": 1.4724599123001099,
      "eval_runtime": 18.0061,
      "eval_samples_per_second": 55.537,
      "eval_steps_per_second": 13.884,
      "step": 1750
    },
    {
      "epoch": 0.13845088839320052,
      "grad_norm": 0.5916198492050171,
      "learning_rate": 0.0001932680570345168,
      "loss": 1.5119,
      "step": 1800
    },
    {
      "epoch": 0.14229674640412276,
      "grad_norm": 1.4087570905685425,
      "learning_rate": 0.00019300833701270032,
      "loss": 1.3608,
      "step": 1850
    },
    {
      "epoch": 0.146142604415045,
      "grad_norm": 1.2559338808059692,
      "learning_rate": 0.00019274861699088386,
      "loss": 1.4772,
      "step": 1900
    },
    {
      "epoch": 0.14998846242596722,
      "grad_norm": 0.9022719860076904,
      "learning_rate": 0.00019248889696906734,
      "loss": 1.4339,
      "step": 1950
    },
    {
      "epoch": 0.15383432043688947,
      "grad_norm": 1.2900218963623047,
      "learning_rate": 0.00019222917694725086,
      "loss": 1.4612,
      "step": 2000
    },
    {
      "epoch": 0.15383432043688947,
      "eval_loss": 1.4609016180038452,
      "eval_runtime": 18.1146,
      "eval_samples_per_second": 55.204,
      "eval_steps_per_second": 13.801,
      "step": 2000
    },
    {
      "epoch": 0.1576801784478117,
      "grad_norm": 0.8418329358100891,
      "learning_rate": 0.0001919694569254344,
      "loss": 1.4944,
      "step": 2050
    },
    {
      "epoch": 0.16152603645873395,
      "grad_norm": 1.538751482963562,
      "learning_rate": 0.0001917097369036179,
      "loss": 1.4135,
      "step": 2100
    },
    {
      "epoch": 0.16537189446965617,
      "grad_norm": 1.3898651599884033,
      "learning_rate": 0.00019145001688180142,
      "loss": 1.3683,
      "step": 2150
    },
    {
      "epoch": 0.1692177524805784,
      "grad_norm": 0.7671115398406982,
      "learning_rate": 0.00019119029685998496,
      "loss": 1.365,
      "step": 2200
    },
    {
      "epoch": 0.17306361049150065,
      "grad_norm": 0.732802152633667,
      "learning_rate": 0.00019093057683816848,
      "loss": 1.3213,
      "step": 2250
    },
    {
      "epoch": 0.17306361049150065,
      "eval_loss": 1.4241567850112915,
      "eval_runtime": 17.9984,
      "eval_samples_per_second": 55.56,
      "eval_steps_per_second": 13.89,
      "step": 2250
    },
    {
      "epoch": 0.1769094685024229,
      "grad_norm": 1.6236932277679443,
      "learning_rate": 0.00019067085681635196,
      "loss": 1.4366,
      "step": 2300
    },
    {
      "epoch": 0.18075532651334514,
      "grad_norm": 1.3093007802963257,
      "learning_rate": 0.0001904111367945355,
      "loss": 1.3468,
      "step": 2350
    },
    {
      "epoch": 0.18460118452426735,
      "grad_norm": 1.409177303314209,
      "learning_rate": 0.00019015141677271901,
      "loss": 1.4066,
      "step": 2400
    },
    {
      "epoch": 0.1884470425351896,
      "grad_norm": 1.0054073333740234,
      "learning_rate": 0.00018989169675090253,
      "loss": 1.4608,
      "step": 2450
    },
    {
      "epoch": 0.19229290054611184,
      "grad_norm": 1.0325884819030762,
      "learning_rate": 0.00018963197672908607,
      "loss": 1.3857,
      "step": 2500
    },
    {
      "epoch": 0.19229290054611184,
      "eval_loss": 1.4193872213363647,
      "eval_runtime": 18.115,
      "eval_samples_per_second": 55.203,
      "eval_steps_per_second": 13.801,
      "step": 2500
    },
    {
      "epoch": 0.19613875855703408,
      "grad_norm": 0.7152838110923767,
      "learning_rate": 0.00018937225670726958,
      "loss": 1.3343,
      "step": 2550
    },
    {
      "epoch": 0.19998461656795632,
      "grad_norm": 0.9736573100090027,
      "learning_rate": 0.0001891125366854531,
      "loss": 1.3832,
      "step": 2600
    },
    {
      "epoch": 0.20383047457887854,
      "grad_norm": 0.9278397560119629,
      "learning_rate": 0.0001888528166636366,
      "loss": 1.3807,
      "step": 2650
    },
    {
      "epoch": 0.20767633258980078,
      "grad_norm": 1.8133916854858398,
      "learning_rate": 0.00018859309664182012,
      "loss": 1.3844,
      "step": 2700
    },
    {
      "epoch": 0.21152219060072303,
      "grad_norm": 1.1289211511611938,
      "learning_rate": 0.00018833337662000366,
      "loss": 1.2984,
      "step": 2750
    },
    {
      "epoch": 0.21152219060072303,
      "eval_loss": 1.4020246267318726,
      "eval_runtime": 18.086,
      "eval_samples_per_second": 55.291,
      "eval_steps_per_second": 13.823,
      "step": 2750
    },
    {
      "epoch": 0.21536804861164527,
      "grad_norm": 1.9358755350112915,
      "learning_rate": 0.00018807365659818717,
      "loss": 1.3231,
      "step": 2800
    },
    {
      "epoch": 0.21921390662256748,
      "grad_norm": 1.453515887260437,
      "learning_rate": 0.00018781393657637068,
      "loss": 1.3395,
      "step": 2850
    },
    {
      "epoch": 0.22305976463348973,
      "grad_norm": 1.423431396484375,
      "learning_rate": 0.0001875542165545542,
      "loss": 1.435,
      "step": 2900
    },
    {
      "epoch": 0.22690562264441197,
      "grad_norm": 0.9964897632598877,
      "learning_rate": 0.0001872944965327377,
      "loss": 1.3356,
      "step": 2950
    },
    {
      "epoch": 0.2307514806553342,
      "grad_norm": 1.5574508905410767,
      "learning_rate": 0.00018703477651092122,
      "loss": 1.4102,
      "step": 3000
    },
    {
      "epoch": 0.2307514806553342,
      "eval_loss": 1.3945672512054443,
      "eval_runtime": 18.0752,
      "eval_samples_per_second": 55.324,
      "eval_steps_per_second": 13.831,
      "step": 3000
    },
    {
      "epoch": 0.23459733866625646,
      "grad_norm": 1.7693700790405273,
      "learning_rate": 0.00018677505648910476,
      "loss": 1.4095,
      "step": 3050
    },
    {
      "epoch": 0.23844319667717867,
      "grad_norm": 1.0146111249923706,
      "learning_rate": 0.00018651533646728827,
      "loss": 1.3416,
      "step": 3100
    },
    {
      "epoch": 0.2422890546881009,
      "grad_norm": 1.228946566581726,
      "learning_rate": 0.0001862556164454718,
      "loss": 1.3326,
      "step": 3150
    },
    {
      "epoch": 0.24613491269902316,
      "grad_norm": 0.9278371930122375,
      "learning_rate": 0.0001859958964236553,
      "loss": 1.4023,
      "step": 3200
    },
    {
      "epoch": 0.2499807707099454,
      "grad_norm": 1.858821988105774,
      "learning_rate": 0.0001857361764018388,
      "loss": 1.3583,
      "step": 3250
    },
    {
      "epoch": 0.2499807707099454,
      "eval_loss": 1.3779631853103638,
      "eval_runtime": 18.0487,
      "eval_samples_per_second": 55.406,
      "eval_steps_per_second": 13.851,
      "step": 3250
    },
    {
      "epoch": 0.25382662872086764,
      "grad_norm": 1.0133246183395386,
      "learning_rate": 0.00018547645638002233,
      "loss": 1.3258,
      "step": 3300
    },
    {
      "epoch": 0.2576724867317899,
      "grad_norm": 1.142626166343689,
      "learning_rate": 0.00018521673635820587,
      "loss": 1.3315,
      "step": 3350
    },
    {
      "epoch": 0.26151834474271207,
      "grad_norm": 0.9573944211006165,
      "learning_rate": 0.00018495701633638938,
      "loss": 1.2949,
      "step": 3400
    },
    {
      "epoch": 0.2653642027536343,
      "grad_norm": 0.8417842984199524,
      "learning_rate": 0.00018469729631457292,
      "loss": 1.3291,
      "step": 3450
    },
    {
      "epoch": 0.26921006076455656,
      "grad_norm": 0.8505682945251465,
      "learning_rate": 0.00018443757629275643,
      "loss": 1.2611,
      "step": 3500
    },
    {
      "epoch": 0.26921006076455656,
      "eval_loss": 1.3726494312286377,
      "eval_runtime": 18.0195,
      "eval_samples_per_second": 55.495,
      "eval_steps_per_second": 13.874,
      "step": 3500
    },
    {
      "epoch": 0.2730559187754788,
      "grad_norm": 1.0631035566329956,
      "learning_rate": 0.00018417785627093992,
      "loss": 1.3384,
      "step": 3550
    },
    {
      "epoch": 0.27690177678640104,
      "grad_norm": 1.1145228147506714,
      "learning_rate": 0.00018391813624912346,
      "loss": 1.4159,
      "step": 3600
    },
    {
      "epoch": 0.2807476347973233,
      "grad_norm": 1.286778450012207,
      "learning_rate": 0.00018365841622730697,
      "loss": 1.3372,
      "step": 3650
    },
    {
      "epoch": 0.28459349280824553,
      "grad_norm": 1.1863288879394531,
      "learning_rate": 0.00018339869620549048,
      "loss": 1.2993,
      "step": 3700
    },
    {
      "epoch": 0.2884393508191678,
      "grad_norm": 1.6189292669296265,
      "learning_rate": 0.00018313897618367402,
      "loss": 1.3464,
      "step": 3750
    },
    {
      "epoch": 0.2884393508191678,
      "eval_loss": 1.3553545475006104,
      "eval_runtime": 18.0591,
      "eval_samples_per_second": 55.374,
      "eval_steps_per_second": 13.843,
      "step": 3750
    },
    {
      "epoch": 0.29228520883009,
      "grad_norm": 1.4823222160339355,
      "learning_rate": 0.00018287925616185754,
      "loss": 1.3392,
      "step": 3800
    },
    {
      "epoch": 0.2961310668410122,
      "grad_norm": 1.4085184335708618,
      "learning_rate": 0.00018261953614004105,
      "loss": 1.2989,
      "step": 3850
    },
    {
      "epoch": 0.29997692485193445,
      "grad_norm": 1.7249082326889038,
      "learning_rate": 0.00018235981611822456,
      "loss": 1.3866,
      "step": 3900
    },
    {
      "epoch": 0.3038227828628567,
      "grad_norm": 0.9753608107566833,
      "learning_rate": 0.00018210009609640807,
      "loss": 1.2916,
      "step": 3950
    },
    {
      "epoch": 0.30766864087377893,
      "grad_norm": 0.6619511246681213,
      "learning_rate": 0.0001818403760745916,
      "loss": 1.2768,
      "step": 4000
    },
    {
      "epoch": 0.30766864087377893,
      "eval_loss": 1.337461233139038,
      "eval_runtime": 18.006,
      "eval_samples_per_second": 55.537,
      "eval_steps_per_second": 13.884,
      "step": 4000
    },
    {
      "epoch": 0.3115144988847012,
      "grad_norm": 0.9473676085472107,
      "learning_rate": 0.00018158065605277513,
      "loss": 1.3116,
      "step": 4050
    },
    {
      "epoch": 0.3153603568956234,
      "grad_norm": 1.2772737741470337,
      "learning_rate": 0.00018132093603095864,
      "loss": 1.3058,
      "step": 4100
    },
    {
      "epoch": 0.31920621490654566,
      "grad_norm": 1.7045694589614868,
      "learning_rate": 0.0001810664104095785,
      "loss": 1.3398,
      "step": 4150
    },
    {
      "epoch": 0.3230520729174679,
      "grad_norm": 1.498179316520691,
      "learning_rate": 0.000180806690387762,
      "loss": 1.3434,
      "step": 4200
    },
    {
      "epoch": 0.32689793092839015,
      "grad_norm": 1.5777134895324707,
      "learning_rate": 0.00018054697036594552,
      "loss": 1.3437,
      "step": 4250
    },
    {
      "epoch": 0.32689793092839015,
      "eval_loss": 1.334365963935852,
      "eval_runtime": 18.026,
      "eval_samples_per_second": 55.475,
      "eval_steps_per_second": 13.869,
      "step": 4250
    },
    {
      "epoch": 0.33074378893931233,
      "grad_norm": 0.7399800419807434,
      "learning_rate": 0.00018028725034412903,
      "loss": 1.2932,
      "step": 4300
    },
    {
      "epoch": 0.3345896469502346,
      "grad_norm": 0.7411991953849792,
      "learning_rate": 0.00018002753032231257,
      "loss": 1.2928,
      "step": 4350
    },
    {
      "epoch": 0.3384355049611568,
      "grad_norm": 1.308003544807434,
      "learning_rate": 0.00017976781030049608,
      "loss": 1.3192,
      "step": 4400
    },
    {
      "epoch": 0.34228136297207906,
      "grad_norm": 1.1857889890670776,
      "learning_rate": 0.0001795080902786796,
      "loss": 1.2718,
      "step": 4450
    },
    {
      "epoch": 0.3461272209830013,
      "grad_norm": 0.5179012417793274,
      "learning_rate": 0.0001792483702568631,
      "loss": 1.2513,
      "step": 4500
    },
    {
      "epoch": 0.3461272209830013,
      "eval_loss": 1.346890926361084,
      "eval_runtime": 18.086,
      "eval_samples_per_second": 55.291,
      "eval_steps_per_second": 13.823,
      "step": 4500
    },
    {
      "epoch": 0.34997307899392355,
      "grad_norm": 1.2267632484436035,
      "learning_rate": 0.00017898865023504662,
      "loss": 1.349,
      "step": 4550
    },
    {
      "epoch": 0.3538189370048458,
      "grad_norm": 0.9660719037055969,
      "learning_rate": 0.00017872893021323013,
      "loss": 1.3194,
      "step": 4600
    },
    {
      "epoch": 0.35766479501576803,
      "grad_norm": 1.4557528495788574,
      "learning_rate": 0.00017846921019141367,
      "loss": 1.3132,
      "step": 4650
    },
    {
      "epoch": 0.3615106530266903,
      "grad_norm": 0.9239174723625183,
      "learning_rate": 0.0001782094901695972,
      "loss": 1.2598,
      "step": 4700
    },
    {
      "epoch": 0.36535651103761246,
      "grad_norm": 1.1237714290618896,
      "learning_rate": 0.0001779497701477807,
      "loss": 1.2506,
      "step": 4750
    },
    {
      "epoch": 0.36535651103761246,
      "eval_loss": 1.3218908309936523,
      "eval_runtime": 18.1211,
      "eval_samples_per_second": 55.184,
      "eval_steps_per_second": 13.796,
      "step": 4750
    },
    {
      "epoch": 0.3692023690485347,
      "grad_norm": 1.0127383470535278,
      "learning_rate": 0.0001776900501259642,
      "loss": 1.3277,
      "step": 4800
    },
    {
      "epoch": 0.37304822705945695,
      "grad_norm": 1.1309473514556885,
      "learning_rate": 0.00017743033010414773,
      "loss": 1.2991,
      "step": 4850
    },
    {
      "epoch": 0.3768940850703792,
      "grad_norm": 1.321747899055481,
      "learning_rate": 0.00017717061008233124,
      "loss": 1.3213,
      "step": 4900
    },
    {
      "epoch": 0.38073994308130144,
      "grad_norm": 1.1251367330551147,
      "learning_rate": 0.00017691089006051478,
      "loss": 1.2742,
      "step": 4950
    },
    {
      "epoch": 0.3845858010922237,
      "grad_norm": 1.1043410301208496,
      "learning_rate": 0.0001766511700386983,
      "loss": 1.2755,
      "step": 5000
    },
    {
      "epoch": 0.3845858010922237,
      "eval_loss": 1.3034751415252686,
      "eval_runtime": 18.0006,
      "eval_samples_per_second": 55.554,
      "eval_steps_per_second": 13.888,
      "step": 5000
    },
    {
      "epoch": 0.3884316591031459,
      "grad_norm": 0.8127657175064087,
      "learning_rate": 0.00017639145001688183,
      "loss": 1.3019,
      "step": 5050
    },
    {
      "epoch": 0.39227751711406816,
      "grad_norm": 0.56494140625,
      "learning_rate": 0.00017613172999506534,
      "loss": 1.3148,
      "step": 5100
    },
    {
      "epoch": 0.3961233751249904,
      "grad_norm": 2.181711435317993,
      "learning_rate": 0.00017587200997324883,
      "loss": 1.2761,
      "step": 5150
    },
    {
      "epoch": 0.39996923313591265,
      "grad_norm": 0.6779603362083435,
      "learning_rate": 0.00017561228995143237,
      "loss": 1.3149,
      "step": 5200
    },
    {
      "epoch": 0.40381509114683484,
      "grad_norm": 0.5844702124595642,
      "learning_rate": 0.00017535256992961588,
      "loss": 1.3437,
      "step": 5250
    },
    {
      "epoch": 0.40381509114683484,
      "eval_loss": 1.2923167943954468,
      "eval_runtime": 17.9871,
      "eval_samples_per_second": 55.595,
      "eval_steps_per_second": 13.899,
      "step": 5250
    },
    {
      "epoch": 0.4076609491577571,
      "grad_norm": 0.9879493117332458,
      "learning_rate": 0.0001750928499077994,
      "loss": 1.2764,
      "step": 5300
    },
    {
      "epoch": 0.4115068071686793,
      "grad_norm": 1.443860650062561,
      "learning_rate": 0.00017483312988598293,
      "loss": 1.3204,
      "step": 5350
    },
    {
      "epoch": 0.41535266517960157,
      "grad_norm": 0.8753446340560913,
      "learning_rate": 0.00017457340986416645,
      "loss": 1.2762,
      "step": 5400
    },
    {
      "epoch": 0.4191985231905238,
      "grad_norm": 1.2027819156646729,
      "learning_rate": 0.00017431368984234996,
      "loss": 1.2097,
      "step": 5450
    },
    {
      "epoch": 0.42304438120144605,
      "grad_norm": 1.1534991264343262,
      "learning_rate": 0.00017405396982053347,
      "loss": 1.2723,
      "step": 5500
    },
    {
      "epoch": 0.42304438120144605,
      "eval_loss": 1.2931731939315796,
      "eval_runtime": 17.9258,
      "eval_samples_per_second": 55.785,
      "eval_steps_per_second": 13.946,
      "step": 5500
    },
    {
      "epoch": 0.4268902392123683,
      "grad_norm": 1.0256164073944092,
      "learning_rate": 0.00017379424979871699,
      "loss": 1.2313,
      "step": 5550
    },
    {
      "epoch": 0.43073609722329054,
      "grad_norm": 1.276945948600769,
      "learning_rate": 0.0001735345297769005,
      "loss": 1.1982,
      "step": 5600
    },
    {
      "epoch": 0.4345819552342128,
      "grad_norm": 0.9002663493156433,
      "learning_rate": 0.00017327480975508404,
      "loss": 1.2726,
      "step": 5650
    },
    {
      "epoch": 0.43842781324513497,
      "grad_norm": 1.1424119472503662,
      "learning_rate": 0.00017301508973326755,
      "loss": 1.3473,
      "step": 5700
    },
    {
      "epoch": 0.4422736712560572,
      "grad_norm": 0.6811870336532593,
      "learning_rate": 0.00017275536971145106,
      "loss": 1.228,
      "step": 5750
    },
    {
      "epoch": 0.4422736712560572,
      "eval_loss": 1.2861703634262085,
      "eval_runtime": 18.0823,
      "eval_samples_per_second": 55.303,
      "eval_steps_per_second": 13.826,
      "step": 5750
    },
    {
      "epoch": 0.44611952926697945,
      "grad_norm": 1.0646696090698242,
      "learning_rate": 0.00017249564968963458,
      "loss": 1.2946,
      "step": 5800
    },
    {
      "epoch": 0.4499653872779017,
      "grad_norm": 1.436909556388855,
      "learning_rate": 0.0001722359296678181,
      "loss": 1.3121,
      "step": 5850
    },
    {
      "epoch": 0.45381124528882394,
      "grad_norm": 0.937135636806488,
      "learning_rate": 0.00017197620964600163,
      "loss": 1.2248,
      "step": 5900
    },
    {
      "epoch": 0.4576571032997462,
      "grad_norm": 0.908935010433197,
      "learning_rate": 0.00017171648962418514,
      "loss": 1.308,
      "step": 5950
    },
    {
      "epoch": 0.4615029613106684,
      "grad_norm": 1.3925087451934814,
      "learning_rate": 0.00017145676960236866,
      "loss": 1.3007,
      "step": 6000
    },
    {
      "epoch": 0.4615029613106684,
      "eval_loss": 1.27406644821167,
      "eval_runtime": 18.0211,
      "eval_samples_per_second": 55.49,
      "eval_steps_per_second": 13.873,
      "step": 6000
    },
    {
      "epoch": 0.46534881932159067,
      "grad_norm": 1.2292288541793823,
      "learning_rate": 0.00017119704958055217,
      "loss": 1.2707,
      "step": 6050
    },
    {
      "epoch": 0.4691946773325129,
      "grad_norm": 0.8948924541473389,
      "learning_rate": 0.00017093732955873568,
      "loss": 1.2481,
      "step": 6100
    },
    {
      "epoch": 0.4730405353434351,
      "grad_norm": 0.7155699133872986,
      "learning_rate": 0.0001706776095369192,
      "loss": 1.2663,
      "step": 6150
    },
    {
      "epoch": 0.47688639335435734,
      "grad_norm": 0.7100064158439636,
      "learning_rate": 0.00017041788951510273,
      "loss": 1.2976,
      "step": 6200
    },
    {
      "epoch": 0.4807322513652796,
      "grad_norm": 1.3250987529754639,
      "learning_rate": 0.00017015816949328625,
      "loss": 1.2368,
      "step": 6250
    },
    {
      "epoch": 0.4807322513652796,
      "eval_loss": 1.2630703449249268,
      "eval_runtime": 18.0054,
      "eval_samples_per_second": 55.539,
      "eval_steps_per_second": 13.885,
      "step": 6250
    },
    {
      "epoch": 0.4845781093762018,
      "grad_norm": 0.9060600996017456,
      "learning_rate": 0.00016989844947146976,
      "loss": 1.2344,
      "step": 6300
    },
    {
      "epoch": 0.48842396738712407,
      "grad_norm": 0.8371444940567017,
      "learning_rate": 0.0001696387294496533,
      "loss": 1.277,
      "step": 6350
    },
    {
      "epoch": 0.4922698253980463,
      "grad_norm": 1.2833727598190308,
      "learning_rate": 0.00016937900942783679,
      "loss": 1.2941,
      "step": 6400
    },
    {
      "epoch": 0.49611568340896856,
      "grad_norm": 1.5922775268554688,
      "learning_rate": 0.0001691192894060203,
      "loss": 1.2448,
      "step": 6450
    },
    {
      "epoch": 0.4999615414198908,
      "grad_norm": 0.9083874225616455,
      "learning_rate": 0.00016885956938420384,
      "loss": 1.2187,
      "step": 6500
    },
    {
      "epoch": 0.4999615414198908,
      "eval_loss": 1.2612597942352295,
      "eval_runtime": 17.9521,
      "eval_samples_per_second": 55.704,
      "eval_steps_per_second": 13.926,
      "step": 6500
    },
    {
      "epoch": 0.503807399430813,
      "grad_norm": 1.3177634477615356,
      "learning_rate": 0.00016859984936238735,
      "loss": 1.305,
      "step": 6550
    },
    {
      "epoch": 0.5076532574417353,
      "grad_norm": 1.8331613540649414,
      "learning_rate": 0.00016834012934057086,
      "loss": 1.2468,
      "step": 6600
    },
    {
      "epoch": 0.5114991154526575,
      "grad_norm": 0.8823532462120056,
      "learning_rate": 0.0001680804093187544,
      "loss": 1.265,
      "step": 6650
    },
    {
      "epoch": 0.5153449734635798,
      "grad_norm": 1.1489806175231934,
      "learning_rate": 0.00016782068929693792,
      "loss": 1.1942,
      "step": 6700
    },
    {
      "epoch": 0.519190831474502,
      "grad_norm": 4.0805816650390625,
      "learning_rate": 0.00016756096927512143,
      "loss": 1.2906,
      "step": 6750
    },
    {
      "epoch": 0.519190831474502,
      "eval_loss": 1.252502202987671,
      "eval_runtime": 17.9221,
      "eval_samples_per_second": 55.797,
      "eval_steps_per_second": 13.949,
      "step": 6750
    },
    {
      "epoch": 0.5230366894854241,
      "grad_norm": 0.9559470415115356,
      "learning_rate": 0.00016730124925330494,
      "loss": 1.1792,
      "step": 6800
    },
    {
      "epoch": 0.5268825474963464,
      "grad_norm": 2.268700361251831,
      "learning_rate": 0.00016704152923148846,
      "loss": 1.2541,
      "step": 6850
    },
    {
      "epoch": 0.5307284055072686,
      "grad_norm": 1.0873395204544067,
      "learning_rate": 0.000166781809209672,
      "loss": 1.2089,
      "step": 6900
    },
    {
      "epoch": 0.5345742635181909,
      "grad_norm": 0.877153217792511,
      "learning_rate": 0.0001665220891878555,
      "loss": 1.2655,
      "step": 6950
    },
    {
      "epoch": 0.5384201215291131,
      "grad_norm": 1.1317107677459717,
      "learning_rate": 0.00016626236916603902,
      "loss": 1.2115,
      "step": 7000
    },
    {
      "epoch": 0.5384201215291131,
      "eval_loss": 1.247037649154663,
      "eval_runtime": 17.9248,
      "eval_samples_per_second": 55.789,
      "eval_steps_per_second": 13.947,
      "step": 7000
    },
    {
      "epoch": 0.5422659795400354,
      "grad_norm": 1.1601048707962036,
      "learning_rate": 0.00016600264914422253,
      "loss": 1.2494,
      "step": 7050
    },
    {
      "epoch": 0.5461118375509576,
      "grad_norm": 0.7940592765808105,
      "learning_rate": 0.00016574292912240605,
      "loss": 1.2906,
      "step": 7100
    },
    {
      "epoch": 0.5499576955618799,
      "grad_norm": 0.6271395087242126,
      "learning_rate": 0.00016548320910058956,
      "loss": 1.2684,
      "step": 7150
    },
    {
      "epoch": 0.5538035535728021,
      "grad_norm": 1.3025091886520386,
      "learning_rate": 0.0001652234890787731,
      "loss": 1.2694,
      "step": 7200
    },
    {
      "epoch": 0.5576494115837243,
      "grad_norm": 1.3218464851379395,
      "learning_rate": 0.0001649637690569566,
      "loss": 1.3098,
      "step": 7250
    },
    {
      "epoch": 0.5576494115837243,
      "eval_loss": 1.2370234727859497,
      "eval_runtime": 18.0496,
      "eval_samples_per_second": 55.403,
      "eval_steps_per_second": 13.851,
      "step": 7250
    },
    {
      "epoch": 0.5614952695946466,
      "grad_norm": 1.1432136297225952,
      "learning_rate": 0.00016470404903514013,
      "loss": 1.2567,
      "step": 7300
    },
    {
      "epoch": 0.5653411276055688,
      "grad_norm": 0.9530320763587952,
      "learning_rate": 0.00016444432901332364,
      "loss": 1.1878,
      "step": 7350
    },
    {
      "epoch": 0.5691869856164911,
      "grad_norm": 1.1852946281433105,
      "learning_rate": 0.00016418460899150715,
      "loss": 1.2153,
      "step": 7400
    },
    {
      "epoch": 0.5730328436274132,
      "grad_norm": 0.7916271686553955,
      "learning_rate": 0.00016392488896969066,
      "loss": 1.2574,
      "step": 7450
    },
    {
      "epoch": 0.5768787016383355,
      "grad_norm": 0.8115867972373962,
      "learning_rate": 0.0001636651689478742,
      "loss": 1.2777,
      "step": 7500
    },
    {
      "epoch": 0.5768787016383355,
      "eval_loss": 1.2334003448486328,
      "eval_runtime": 18.0145,
      "eval_samples_per_second": 55.511,
      "eval_steps_per_second": 13.878,
      "step": 7500
    },
    {
      "epoch": 0.5807245596492577,
      "grad_norm": 0.9350728988647461,
      "learning_rate": 0.00016340544892605772,
      "loss": 1.2261,
      "step": 7550
    },
    {
      "epoch": 0.58457041766018,
      "grad_norm": 0.7061731815338135,
      "learning_rate": 0.00016314572890424126,
      "loss": 1.172,
      "step": 7600
    },
    {
      "epoch": 0.5884162756711022,
      "grad_norm": 1.091739296913147,
      "learning_rate": 0.00016288600888242477,
      "loss": 1.272,
      "step": 7650
    },
    {
      "epoch": 0.5922621336820244,
      "grad_norm": 0.8880358338356018,
      "learning_rate": 0.00016262628886060826,
      "loss": 1.1847,
      "step": 7700
    },
    {
      "epoch": 0.5961079916929467,
      "grad_norm": 2.7609329223632812,
      "learning_rate": 0.0001623665688387918,
      "loss": 1.2636,
      "step": 7750
    },
    {
      "epoch": 0.5961079916929467,
      "eval_loss": 1.2325551509857178,
      "eval_runtime": 17.8663,
      "eval_samples_per_second": 55.971,
      "eval_steps_per_second": 13.993,
      "step": 7750
    },
    {
      "epoch": 0.5999538497038689,
      "grad_norm": 0.8872610926628113,
      "learning_rate": 0.0001621068488169753,
      "loss": 1.245,
      "step": 7800
    },
    {
      "epoch": 0.6037997077147912,
      "grad_norm": 0.7548871040344238,
      "learning_rate": 0.00016184712879515882,
      "loss": 1.2349,
      "step": 7850
    },
    {
      "epoch": 0.6076455657257134,
      "grad_norm": 1.104351282119751,
      "learning_rate": 0.00016158740877334236,
      "loss": 1.2397,
      "step": 7900
    },
    {
      "epoch": 0.6114914237366357,
      "grad_norm": 0.8331647515296936,
      "learning_rate": 0.00016132768875152587,
      "loss": 1.2634,
      "step": 7950
    },
    {
      "epoch": 0.6153372817475579,
      "grad_norm": 1.0910013914108276,
      "learning_rate": 0.00016106796872970939,
      "loss": 1.2897,
      "step": 8000
    },
    {
      "epoch": 0.6153372817475579,
      "eval_loss": 1.2241965532302856,
      "eval_runtime": 18.04,
      "eval_samples_per_second": 55.432,
      "eval_steps_per_second": 13.858,
      "step": 8000
    },
    {
      "epoch": 0.6191831397584802,
      "grad_norm": 1.4128022193908691,
      "learning_rate": 0.0001608082487078929,
      "loss": 1.1912,
      "step": 8050
    },
    {
      "epoch": 0.6230289977694023,
      "grad_norm": 1.5363566875457764,
      "learning_rate": 0.0001605485286860764,
      "loss": 1.2622,
      "step": 8100
    },
    {
      "epoch": 0.6268748557803245,
      "grad_norm": 1.334889531135559,
      "learning_rate": 0.00016028880866425992,
      "loss": 1.2462,
      "step": 8150
    },
    {
      "epoch": 0.6307207137912468,
      "grad_norm": 1.63850998878479,
      "learning_rate": 0.00016002908864244346,
      "loss": 1.2099,
      "step": 8200
    },
    {
      "epoch": 0.634566571802169,
      "grad_norm": 1.2087870836257935,
      "learning_rate": 0.00015976936862062698,
      "loss": 1.1669,
      "step": 8250
    },
    {
      "epoch": 0.634566571802169,
      "eval_loss": 1.2158918380737305,
      "eval_runtime": 17.8293,
      "eval_samples_per_second": 56.087,
      "eval_steps_per_second": 14.022,
      "step": 8250
    },
    {
      "epoch": 0.6384124298130913,
      "grad_norm": 2.0238049030303955,
      "learning_rate": 0.0001595096485988105,
      "loss": 1.1893,
      "step": 8300
    },
    {
      "epoch": 0.6422582878240135,
      "grad_norm": 0.7206680178642273,
      "learning_rate": 0.000159249928576994,
      "loss": 1.2155,
      "step": 8350
    },
    {
      "epoch": 0.6461041458349358,
      "grad_norm": 1.0200512409210205,
      "learning_rate": 0.00015899020855517752,
      "loss": 1.2049,
      "step": 8400
    },
    {
      "epoch": 0.649950003845858,
      "grad_norm": 0.7880833745002747,
      "learning_rate": 0.00015873048853336106,
      "loss": 1.2564,
      "step": 8450
    },
    {
      "epoch": 0.6537958618567803,
      "grad_norm": 1.0986734628677368,
      "learning_rate": 0.00015847076851154457,
      "loss": 1.2989,
      "step": 8500
    },
    {
      "epoch": 0.6537958618567803,
      "eval_loss": 1.2085261344909668,
      "eval_runtime": 17.887,
      "eval_samples_per_second": 55.907,
      "eval_steps_per_second": 13.977,
      "step": 8500
    },
    {
      "epoch": 0.6576417198677025,
      "grad_norm": 0.5527728796005249,
      "learning_rate": 0.00015821104848972808,
      "loss": 1.1798,
      "step": 8550
    },
    {
      "epoch": 0.6614875778786247,
      "grad_norm": 1.0168190002441406,
      "learning_rate": 0.0001579513284679116,
      "loss": 1.211,
      "step": 8600
    },
    {
      "epoch": 0.665333435889547,
      "grad_norm": 0.8436816334724426,
      "learning_rate": 0.0001576916084460951,
      "loss": 1.2018,
      "step": 8650
    },
    {
      "epoch": 0.6691792939004692,
      "grad_norm": 0.967677891254425,
      "learning_rate": 0.00015743188842427862,
      "loss": 1.2134,
      "step": 8700
    },
    {
      "epoch": 0.6730251519113915,
      "grad_norm": 0.9716609120368958,
      "learning_rate": 0.00015717216840246216,
      "loss": 1.2206,
      "step": 8750
    },
    {
      "epoch": 0.6730251519113915,
      "eval_loss": 1.2069470882415771,
      "eval_runtime": 17.9386,
      "eval_samples_per_second": 55.746,
      "eval_steps_per_second": 13.936,
      "step": 8750
    },
    {
      "epoch": 0.6768710099223136,
      "grad_norm": 1.1798300743103027,
      "learning_rate": 0.00015691244838064567,
      "loss": 1.2258,
      "step": 8800
    },
    {
      "epoch": 0.6807168679332359,
      "grad_norm": 1.2064564228057861,
      "learning_rate": 0.00015665272835882919,
      "loss": 1.217,
      "step": 8850
    },
    {
      "epoch": 0.6845627259441581,
      "grad_norm": 1.2753881216049194,
      "learning_rate": 0.00015639300833701273,
      "loss": 1.1915,
      "step": 8900
    },
    {
      "epoch": 0.6884085839550804,
      "grad_norm": 1.2899794578552246,
      "learning_rate": 0.00015613848271563255,
      "loss": 1.1967,
      "step": 8950
    },
    {
      "epoch": 0.6922544419660026,
      "grad_norm": 0.5771601796150208,
      "learning_rate": 0.00015587876269381606,
      "loss": 1.1958,
      "step": 9000
    },
    {
      "epoch": 0.6922544419660026,
      "eval_loss": 1.19791841506958,
      "eval_runtime": 18.0075,
      "eval_samples_per_second": 55.532,
      "eval_steps_per_second": 13.883,
      "step": 9000
    },
    {
      "epoch": 0.6961002999769248,
      "grad_norm": 1.467191457748413,
      "learning_rate": 0.00015561904267199958,
      "loss": 1.1938,
      "step": 9050
    },
    {
      "epoch": 0.6999461579878471,
      "grad_norm": 0.7669786214828491,
      "learning_rate": 0.00015535932265018312,
      "loss": 1.2228,
      "step": 9100
    },
    {
      "epoch": 0.7037920159987693,
      "grad_norm": 0.843961238861084,
      "learning_rate": 0.00015509960262836663,
      "loss": 1.1543,
      "step": 9150
    },
    {
      "epoch": 0.7076378740096916,
      "grad_norm": 1.2265573740005493,
      "learning_rate": 0.00015483988260655017,
      "loss": 1.1518,
      "step": 9200
    },
    {
      "epoch": 0.7114837320206138,
      "grad_norm": 1.1644186973571777,
      "learning_rate": 0.00015458016258473365,
      "loss": 1.23,
      "step": 9250
    },
    {
      "epoch": 0.7114837320206138,
      "eval_loss": 1.1922409534454346,
      "eval_runtime": 17.9384,
      "eval_samples_per_second": 55.746,
      "eval_steps_per_second": 13.937,
      "step": 9250
    },
    {
      "epoch": 0.7153295900315361,
      "grad_norm": 1.3787301778793335,
      "learning_rate": 0.00015432044256291717,
      "loss": 1.2347,
      "step": 9300
    },
    {
      "epoch": 0.7191754480424583,
      "grad_norm": 1.4755727052688599,
      "learning_rate": 0.0001540607225411007,
      "loss": 1.1596,
      "step": 9350
    },
    {
      "epoch": 0.7230213060533806,
      "grad_norm": 1.031275749206543,
      "learning_rate": 0.00015380100251928422,
      "loss": 1.2045,
      "step": 9400
    },
    {
      "epoch": 0.7268671640643027,
      "grad_norm": 1.2802574634552002,
      "learning_rate": 0.00015354128249746773,
      "loss": 1.205,
      "step": 9450
    },
    {
      "epoch": 0.7307130220752249,
      "grad_norm": 0.5222998857498169,
      "learning_rate": 0.00015328156247565127,
      "loss": 1.1678,
      "step": 9500
    },
    {
      "epoch": 0.7307130220752249,
      "eval_loss": 1.1886347532272339,
      "eval_runtime": 17.9243,
      "eval_samples_per_second": 55.79,
      "eval_steps_per_second": 13.948,
      "step": 9500
    },
    {
      "epoch": 0.7345588800861472,
      "grad_norm": 0.8676270842552185,
      "learning_rate": 0.00015302184245383479,
      "loss": 1.1388,
      "step": 9550
    },
    {
      "epoch": 0.7384047380970694,
      "grad_norm": 1.198843240737915,
      "learning_rate": 0.00015276212243201827,
      "loss": 1.1995,
      "step": 9600
    },
    {
      "epoch": 0.7422505961079917,
      "grad_norm": 0.6684653162956238,
      "learning_rate": 0.0001525024024102018,
      "loss": 1.1567,
      "step": 9650
    },
    {
      "epoch": 0.7460964541189139,
      "grad_norm": 0.931119441986084,
      "learning_rate": 0.00015224268238838532,
      "loss": 1.1375,
      "step": 9700
    },
    {
      "epoch": 0.7499423121298362,
      "grad_norm": 0.7734692096710205,
      "learning_rate": 0.00015198296236656884,
      "loss": 1.1836,
      "step": 9750
    },
    {
      "epoch": 0.7499423121298362,
      "eval_loss": 1.1750942468643188,
      "eval_runtime": 18.0078,
      "eval_samples_per_second": 55.532,
      "eval_steps_per_second": 13.883,
      "step": 9750
    },
    {
      "epoch": 0.7537881701407584,
      "grad_norm": 1.0753988027572632,
      "learning_rate": 0.00015172324234475238,
      "loss": 1.1666,
      "step": 9800
    },
    {
      "epoch": 0.7576340281516807,
      "grad_norm": 0.45949143171310425,
      "learning_rate": 0.0001514635223229359,
      "loss": 1.1498,
      "step": 9850
    },
    {
      "epoch": 0.7614798861626029,
      "grad_norm": 1.0716335773468018,
      "learning_rate": 0.0001512038023011194,
      "loss": 1.1669,
      "step": 9900
    },
    {
      "epoch": 0.7653257441735252,
      "grad_norm": 1.043646216392517,
      "learning_rate": 0.00015094408227930292,
      "loss": 1.1689,
      "step": 9950
    },
    {
      "epoch": 0.7691716021844474,
      "grad_norm": 1.0494813919067383,
      "learning_rate": 0.00015068436225748643,
      "loss": 1.2343,
      "step": 10000
    },
    {
      "epoch": 0.7691716021844474,
      "eval_loss": 1.1786631345748901,
      "eval_runtime": 17.9206,
      "eval_samples_per_second": 55.802,
      "eval_steps_per_second": 13.95,
      "step": 10000
    },
    {
      "epoch": 0.7730174601953695,
      "grad_norm": 0.7689708471298218,
      "learning_rate": 0.00015042983663610628,
      "loss": 1.2341,
      "step": 10050
    },
    {
      "epoch": 0.7768633182062918,
      "grad_norm": 1.0559266805648804,
      "learning_rate": 0.00015017011661428982,
      "loss": 1.0868,
      "step": 10100
    },
    {
      "epoch": 0.780709176217214,
      "grad_norm": 1.01194429397583,
      "learning_rate": 0.00014991039659247333,
      "loss": 1.2187,
      "step": 10150
    },
    {
      "epoch": 0.7845550342281363,
      "grad_norm": 0.9095450043678284,
      "learning_rate": 0.00014965067657065684,
      "loss": 1.1432,
      "step": 10200
    },
    {
      "epoch": 0.7884008922390585,
      "grad_norm": 1.1280279159545898,
      "learning_rate": 0.00014939095654884036,
      "loss": 1.1631,
      "step": 10250
    },
    {
      "epoch": 0.7884008922390585,
      "eval_loss": 1.1772924661636353,
      "eval_runtime": 18.038,
      "eval_samples_per_second": 55.438,
      "eval_steps_per_second": 13.86,
      "step": 10250
    },
    {
      "epoch": 0.7922467502499808,
      "grad_norm": 1.1410025358200073,
      "learning_rate": 0.00014913123652702387,
      "loss": 1.2058,
      "step": 10300
    },
    {
      "epoch": 0.796092608260903,
      "grad_norm": 0.7516416311264038,
      "learning_rate": 0.00014887151650520738,
      "loss": 1.2702,
      "step": 10350
    },
    {
      "epoch": 0.7999384662718253,
      "grad_norm": 1.9470982551574707,
      "learning_rate": 0.00014861179648339092,
      "loss": 1.1322,
      "step": 10400
    },
    {
      "epoch": 0.8037843242827475,
      "grad_norm": 1.1969455480575562,
      "learning_rate": 0.00014835207646157444,
      "loss": 1.1882,
      "step": 10450
    },
    {
      "epoch": 0.8076301822936697,
      "grad_norm": 1.2365367412567139,
      "learning_rate": 0.00014809235643975795,
      "loss": 1.1929,
      "step": 10500
    },
    {
      "epoch": 0.8076301822936697,
      "eval_loss": 1.168535590171814,
      "eval_runtime": 17.931,
      "eval_samples_per_second": 55.769,
      "eval_steps_per_second": 13.942,
      "step": 10500
    },
    {
      "epoch": 0.811476040304592,
      "grad_norm": 1.2798963785171509,
      "learning_rate": 0.00014783263641794146,
      "loss": 1.2693,
      "step": 10550
    },
    {
      "epoch": 0.8153218983155142,
      "grad_norm": 0.8195398449897766,
      "learning_rate": 0.00014757291639612497,
      "loss": 1.1933,
      "step": 10600
    },
    {
      "epoch": 0.8191677563264365,
      "grad_norm": 1.039031744003296,
      "learning_rate": 0.0001473131963743085,
      "loss": 1.1257,
      "step": 10650
    },
    {
      "epoch": 0.8230136143373586,
      "grad_norm": 1.0875959396362305,
      "learning_rate": 0.00014705347635249203,
      "loss": 1.2578,
      "step": 10700
    },
    {
      "epoch": 0.8268594723482809,
      "grad_norm": 1.5674726963043213,
      "learning_rate": 0.00014679375633067554,
      "loss": 1.1895,
      "step": 10750
    },
    {
      "epoch": 0.8268594723482809,
      "eval_loss": 1.1696391105651855,
      "eval_runtime": 17.8658,
      "eval_samples_per_second": 55.973,
      "eval_steps_per_second": 13.993,
      "step": 10750
    },
    {
      "epoch": 0.8307053303592031,
      "grad_norm": 0.7315701842308044,
      "learning_rate": 0.00014653403630885905,
      "loss": 1.1379,
      "step": 10800
    },
    {
      "epoch": 0.8345511883701254,
      "grad_norm": 1.0033215284347534,
      "learning_rate": 0.00014627431628704257,
      "loss": 1.198,
      "step": 10850
    },
    {
      "epoch": 0.8383970463810476,
      "grad_norm": 1.0194263458251953,
      "learning_rate": 0.00014601459626522608,
      "loss": 1.2186,
      "step": 10900
    },
    {
      "epoch": 0.8422429043919698,
      "grad_norm": 0.9829340577125549,
      "learning_rate": 0.00014575487624340962,
      "loss": 1.1458,
      "step": 10950
    },
    {
      "epoch": 0.8460887624028921,
      "grad_norm": 1.3082759380340576,
      "learning_rate": 0.00014549515622159313,
      "loss": 1.1545,
      "step": 11000
    },
    {
      "epoch": 0.8460887624028921,
      "eval_loss": 1.1625275611877441,
      "eval_runtime": 17.8346,
      "eval_samples_per_second": 56.071,
      "eval_steps_per_second": 14.018,
      "step": 11000
    },
    {
      "epoch": 0.8499346204138143,
      "grad_norm": 1.2006999254226685,
      "learning_rate": 0.00014523543619977664,
      "loss": 1.1187,
      "step": 11050
    },
    {
      "epoch": 0.8537804784247366,
      "grad_norm": 1.0500705242156982,
      "learning_rate": 0.00014497571617796018,
      "loss": 1.1911,
      "step": 11100
    },
    {
      "epoch": 0.8576263364356588,
      "grad_norm": 0.8597742915153503,
      "learning_rate": 0.0001447159961561437,
      "loss": 1.1176,
      "step": 11150
    },
    {
      "epoch": 0.8614721944465811,
      "grad_norm": 1.338990569114685,
      "learning_rate": 0.00014445627613432718,
      "loss": 1.1699,
      "step": 11200
    },
    {
      "epoch": 0.8653180524575033,
      "grad_norm": 0.903128445148468,
      "learning_rate": 0.00014419655611251072,
      "loss": 1.1504,
      "step": 11250
    },
    {
      "epoch": 0.8653180524575033,
      "eval_loss": 1.161923885345459,
      "eval_runtime": 17.9304,
      "eval_samples_per_second": 55.771,
      "eval_steps_per_second": 13.943,
      "step": 11250
    },
    {
      "epoch": 0.8691639104684256,
      "grad_norm": 0.8747849464416504,
      "learning_rate": 0.00014393683609069424,
      "loss": 1.1305,
      "step": 11300
    },
    {
      "epoch": 0.8730097684793477,
      "grad_norm": 1.1505181789398193,
      "learning_rate": 0.00014367711606887775,
      "loss": 1.1515,
      "step": 11350
    },
    {
      "epoch": 0.8768556264902699,
      "grad_norm": 0.6178755164146423,
      "learning_rate": 0.0001434173960470613,
      "loss": 1.1785,
      "step": 11400
    },
    {
      "epoch": 0.8807014845011922,
      "grad_norm": 0.7437123656272888,
      "learning_rate": 0.0001431576760252448,
      "loss": 1.1567,
      "step": 11450
    },
    {
      "epoch": 0.8845473425121144,
      "grad_norm": 1.574104905128479,
      "learning_rate": 0.00014289795600342831,
      "loss": 1.1371,
      "step": 11500
    },
    {
      "epoch": 0.8845473425121144,
      "eval_loss": 1.154138445854187,
      "eval_runtime": 17.9772,
      "eval_samples_per_second": 55.626,
      "eval_steps_per_second": 13.906,
      "step": 11500
    },
    {
      "epoch": 0.8883932005230367,
      "grad_norm": 2.203948497772217,
      "learning_rate": 0.00014264343038204814,
      "loss": 1.2118,
      "step": 11550
    },
    {
      "epoch": 0.8922390585339589,
      "grad_norm": 1.1410473585128784,
      "learning_rate": 0.00014238371036023168,
      "loss": 1.2239,
      "step": 11600
    },
    {
      "epoch": 0.8960849165448812,
      "grad_norm": 1.0226402282714844,
      "learning_rate": 0.0001421239903384152,
      "loss": 1.1686,
      "step": 11650
    },
    {
      "epoch": 0.8999307745558034,
      "grad_norm": 1.0350555181503296,
      "learning_rate": 0.00014186427031659873,
      "loss": 1.1584,
      "step": 11700
    },
    {
      "epoch": 0.9037766325667257,
      "grad_norm": 1.6758803129196167,
      "learning_rate": 0.00014160455029478224,
      "loss": 1.1638,
      "step": 11750
    },
    {
      "epoch": 0.9037766325667257,
      "eval_loss": 1.1433300971984863,
      "eval_runtime": 17.8796,
      "eval_samples_per_second": 55.93,
      "eval_steps_per_second": 13.982,
      "step": 11750
    },
    {
      "epoch": 0.9076224905776479,
      "grad_norm": 1.2579885721206665,
      "learning_rate": 0.00014134483027296576,
      "loss": 1.2198,
      "step": 11800
    },
    {
      "epoch": 0.9114683485885701,
      "grad_norm": 0.886987030506134,
      "learning_rate": 0.00014108511025114927,
      "loss": 1.1477,
      "step": 11850
    },
    {
      "epoch": 0.9153142065994924,
      "grad_norm": 0.9828807711601257,
      "learning_rate": 0.00014082539022933278,
      "loss": 1.1115,
      "step": 11900
    },
    {
      "epoch": 0.9191600646104146,
      "grad_norm": 1.2362446784973145,
      "learning_rate": 0.0001405656702075163,
      "loss": 1.161,
      "step": 11950
    },
    {
      "epoch": 0.9230059226213368,
      "grad_norm": 1.1353052854537964,
      "learning_rate": 0.00014030595018569984,
      "loss": 1.1369,
      "step": 12000
    },
    {
      "epoch": 0.9230059226213368,
      "eval_loss": 1.144094467163086,
      "eval_runtime": 17.8499,
      "eval_samples_per_second": 56.023,
      "eval_steps_per_second": 14.006,
      "step": 12000
    },
    {
      "epoch": 0.926851780632259,
      "grad_norm": 1.0390766859054565,
      "learning_rate": 0.00014004623016388335,
      "loss": 1.1763,
      "step": 12050
    },
    {
      "epoch": 0.9306976386431813,
      "grad_norm": 1.1437292098999023,
      "learning_rate": 0.00013978651014206686,
      "loss": 1.153,
      "step": 12100
    },
    {
      "epoch": 0.9345434966541035,
      "grad_norm": 0.7012118697166443,
      "learning_rate": 0.00013952679012025037,
      "loss": 1.1209,
      "step": 12150
    },
    {
      "epoch": 0.9383893546650258,
      "grad_norm": 0.558203399181366,
      "learning_rate": 0.0001392670700984339,
      "loss": 1.0998,
      "step": 12200
    },
    {
      "epoch": 0.942235212675948,
      "grad_norm": 1.031898021697998,
      "learning_rate": 0.0001390073500766174,
      "loss": 1.1308,
      "step": 12250
    },
    {
      "epoch": 0.942235212675948,
      "eval_loss": 1.1427006721496582,
      "eval_runtime": 17.9677,
      "eval_samples_per_second": 55.656,
      "eval_steps_per_second": 13.914,
      "step": 12250
    },
    {
      "epoch": 0.9460810706868702,
      "grad_norm": 0.9146320223808289,
      "learning_rate": 0.00013874763005480094,
      "loss": 1.1441,
      "step": 12300
    },
    {
      "epoch": 0.9499269286977925,
      "grad_norm": 1.7698357105255127,
      "learning_rate": 0.00013848791003298445,
      "loss": 1.1278,
      "step": 12350
    },
    {
      "epoch": 0.9537727867087147,
      "grad_norm": 1.7621064186096191,
      "learning_rate": 0.00013822819001116797,
      "loss": 1.2191,
      "step": 12400
    },
    {
      "epoch": 0.957618644719637,
      "grad_norm": 1.2093744277954102,
      "learning_rate": 0.00013796846998935148,
      "loss": 1.1541,
      "step": 12450
    },
    {
      "epoch": 0.9614645027305592,
      "grad_norm": 1.0879639387130737,
      "learning_rate": 0.000137708749967535,
      "loss": 1.1278,
      "step": 12500
    },
    {
      "epoch": 0.9614645027305592,
      "eval_loss": 1.1357394456863403,
      "eval_runtime": 17.9695,
      "eval_samples_per_second": 55.65,
      "eval_steps_per_second": 13.912,
      "step": 12500
    },
    {
      "epoch": 0.9653103607414815,
      "grad_norm": 1.415139079093933,
      "learning_rate": 0.00013744902994571853,
      "loss": 1.1986,
      "step": 12550
    },
    {
      "epoch": 0.9691562187524037,
      "grad_norm": 1.0320454835891724,
      "learning_rate": 0.00013718930992390204,
      "loss": 1.153,
      "step": 12600
    },
    {
      "epoch": 0.973002076763326,
      "grad_norm": 1.0736747980117798,
      "learning_rate": 0.00013692958990208556,
      "loss": 1.1931,
      "step": 12650
    },
    {
      "epoch": 0.9768479347742481,
      "grad_norm": 0.8954864740371704,
      "learning_rate": 0.0001366698698802691,
      "loss": 1.1908,
      "step": 12700
    },
    {
      "epoch": 0.9806937927851703,
      "grad_norm": 1.3287911415100098,
      "learning_rate": 0.0001364101498584526,
      "loss": 1.1502,
      "step": 12750
    },
    {
      "epoch": 0.9806937927851703,
      "eval_loss": 1.133840799331665,
      "eval_runtime": 18.0959,
      "eval_samples_per_second": 55.261,
      "eval_steps_per_second": 13.815,
      "step": 12750
    },
    {
      "epoch": 0.9845396507960926,
      "grad_norm": 1.000588059425354,
      "learning_rate": 0.0001361504298366361,
      "loss": 1.1001,
      "step": 12800
    },
    {
      "epoch": 0.9883855088070148,
      "grad_norm": 0.9359833598136902,
      "learning_rate": 0.00013589070981481963,
      "loss": 1.2289,
      "step": 12850
    },
    {
      "epoch": 0.9922313668179371,
      "grad_norm": 1.5241230726242065,
      "learning_rate": 0.00013563098979300315,
      "loss": 1.1726,
      "step": 12900
    },
    {
      "epoch": 0.9960772248288593,
      "grad_norm": 1.0804429054260254,
      "learning_rate": 0.00013537126977118666,
      "loss": 1.1755,
      "step": 12950
    },
    {
      "epoch": 0.9999230828397816,
      "grad_norm": 0.8376865983009338,
      "learning_rate": 0.0001351115497493702,
      "loss": 1.158,
      "step": 13000
    },
    {
      "epoch": 0.9999230828397816,
      "eval_loss": 1.1301764249801636,
      "eval_runtime": 18.0468,
      "eval_samples_per_second": 55.411,
      "eval_steps_per_second": 13.853,
      "step": 13000
    },
    {
      "epoch": 1.0037689408507038,
      "grad_norm": 0.8493902087211609,
      "learning_rate": 0.0001348518297275537,
      "loss": 1.0935,
      "step": 13050
    },
    {
      "epoch": 1.007614798861626,
      "grad_norm": 0.9508585929870605,
      "learning_rate": 0.00013459210970573723,
      "loss": 1.0773,
      "step": 13100
    },
    {
      "epoch": 1.0114606568725482,
      "grad_norm": 1.047767162322998,
      "learning_rate": 0.00013433238968392074,
      "loss": 1.1722,
      "step": 13150
    },
    {
      "epoch": 1.0153065148834706,
      "grad_norm": 1.0310213565826416,
      "learning_rate": 0.00013407266966210425,
      "loss": 1.059,
      "step": 13200
    },
    {
      "epoch": 1.0191523728943928,
      "grad_norm": 0.7563040852546692,
      "learning_rate": 0.00013381294964028776,
      "loss": 1.0438,
      "step": 13250
    },
    {
      "epoch": 1.0191523728943928,
      "eval_loss": 1.121547818183899,
      "eval_runtime": 17.9914,
      "eval_samples_per_second": 55.582,
      "eval_steps_per_second": 13.895,
      "step": 13250
    },
    {
      "epoch": 1.022998230905315,
      "grad_norm": 0.8817610144615173,
      "learning_rate": 0.0001335532296184713,
      "loss": 1.1176,
      "step": 13300
    },
    {
      "epoch": 1.0268440889162371,
      "grad_norm": 0.8703081011772156,
      "learning_rate": 0.00013329350959665482,
      "loss": 1.0568,
      "step": 13350
    },
    {
      "epoch": 1.0306899469271595,
      "grad_norm": 1.0551347732543945,
      "learning_rate": 0.00013303378957483833,
      "loss": 1.1746,
      "step": 13400
    },
    {
      "epoch": 1.0345358049380817,
      "grad_norm": 1.2630723714828491,
      "learning_rate": 0.00013277406955302184,
      "loss": 1.0232,
      "step": 13450
    },
    {
      "epoch": 1.038381662949004,
      "grad_norm": 1.2565157413482666,
      "learning_rate": 0.00013251434953120536,
      "loss": 1.0523,
      "step": 13500
    },
    {
      "epoch": 1.038381662949004,
      "eval_loss": 1.1202689409255981,
      "eval_runtime": 17.8981,
      "eval_samples_per_second": 55.872,
      "eval_steps_per_second": 13.968,
      "step": 13500
    },
    {
      "epoch": 1.042227520959926,
      "grad_norm": 1.9204115867614746,
      "learning_rate": 0.0001322546295093889,
      "loss": 1.1115,
      "step": 13550
    },
    {
      "epoch": 1.0460733789708483,
      "grad_norm": 1.1753497123718262,
      "learning_rate": 0.0001319949094875724,
      "loss": 1.0463,
      "step": 13600
    },
    {
      "epoch": 1.0499192369817707,
      "grad_norm": 1.0144574642181396,
      "learning_rate": 0.00013173518946575592,
      "loss": 1.0914,
      "step": 13650
    },
    {
      "epoch": 1.0537650949926929,
      "grad_norm": 2.0906662940979004,
      "learning_rate": 0.00013147546944393943,
      "loss": 1.1635,
      "step": 13700
    },
    {
      "epoch": 1.057610953003615,
      "grad_norm": 1.0127108097076416,
      "learning_rate": 0.00013121574942212295,
      "loss": 1.1266,
      "step": 13750
    },
    {
      "epoch": 1.057610953003615,
      "eval_loss": 1.115894079208374,
      "eval_runtime": 17.8891,
      "eval_samples_per_second": 55.9,
      "eval_steps_per_second": 13.975,
      "step": 13750
    },
    {
      "epoch": 1.0614568110145373,
      "grad_norm": 1.2559298276901245,
      "learning_rate": 0.00013095602940030646,
      "loss": 1.0924,
      "step": 13800
    },
    {
      "epoch": 1.0653026690254597,
      "grad_norm": 0.7502859234809875,
      "learning_rate": 0.00013069630937849,
      "loss": 1.0756,
      "step": 13850
    },
    {
      "epoch": 1.0691485270363819,
      "grad_norm": 0.6954963207244873,
      "learning_rate": 0.0001304365893566735,
      "loss": 1.1284,
      "step": 13900
    },
    {
      "epoch": 1.072994385047304,
      "grad_norm": 1.3833235502243042,
      "learning_rate": 0.00013017686933485703,
      "loss": 1.1175,
      "step": 13950
    },
    {
      "epoch": 1.0768402430582262,
      "grad_norm": 0.9848393201828003,
      "learning_rate": 0.00012991714931304057,
      "loss": 1.1295,
      "step": 14000
    },
    {
      "epoch": 1.0768402430582262,
      "eval_loss": 1.1086758375167847,
      "eval_runtime": 17.8007,
      "eval_samples_per_second": 56.177,
      "eval_steps_per_second": 14.044,
      "step": 14000
    },
    {
      "epoch": 1.0806861010691486,
      "grad_norm": 1.1354585886001587,
      "learning_rate": 0.00012965742929122405,
      "loss": 1.0351,
      "step": 14050
    },
    {
      "epoch": 1.0845319590800708,
      "grad_norm": 0.760317325592041,
      "learning_rate": 0.00012939770926940756,
      "loss": 1.0614,
      "step": 14100
    },
    {
      "epoch": 1.088377817090993,
      "grad_norm": 0.9277663230895996,
      "learning_rate": 0.0001291379892475911,
      "loss": 1.0689,
      "step": 14150
    },
    {
      "epoch": 1.0922236751019152,
      "grad_norm": 1.0846219062805176,
      "learning_rate": 0.00012887826922577462,
      "loss": 1.1679,
      "step": 14200
    },
    {
      "epoch": 1.0960695331128374,
      "grad_norm": 1.204969048500061,
      "learning_rate": 0.00012861854920395816,
      "loss": 1.1198,
      "step": 14250
    },
    {
      "epoch": 1.0960695331128374,
      "eval_loss": 1.1072660684585571,
      "eval_runtime": 17.9757,
      "eval_samples_per_second": 55.631,
      "eval_steps_per_second": 13.908,
      "step": 14250
    },
    {
      "epoch": 1.0999153911237598,
      "grad_norm": 1.556897759437561,
      "learning_rate": 0.00012835882918214167,
      "loss": 1.1511,
      "step": 14300
    },
    {
      "epoch": 1.103761249134682,
      "grad_norm": 1.4192557334899902,
      "learning_rate": 0.00012809910916032518,
      "loss": 1.1074,
      "step": 14350
    },
    {
      "epoch": 1.1076071071456042,
      "grad_norm": 0.5456421971321106,
      "learning_rate": 0.0001278393891385087,
      "loss": 1.0818,
      "step": 14400
    },
    {
      "epoch": 1.1114529651565264,
      "grad_norm": 1.282106876373291,
      "learning_rate": 0.0001275796691166922,
      "loss": 1.0431,
      "step": 14450
    },
    {
      "epoch": 1.1152988231674485,
      "grad_norm": 0.7997551560401917,
      "learning_rate": 0.00012731994909487572,
      "loss": 1.0882,
      "step": 14500
    },
    {
      "epoch": 1.1152988231674485,
      "eval_loss": 1.1057496070861816,
      "eval_runtime": 17.7977,
      "eval_samples_per_second": 56.187,
      "eval_steps_per_second": 14.047,
      "step": 14500
    },
    {
      "epoch": 1.119144681178371,
      "grad_norm": 1.2388238906860352,
      "learning_rate": 0.00012706022907305926,
      "loss": 1.1321,
      "step": 14550
    },
    {
      "epoch": 1.1229905391892931,
      "grad_norm": 0.5141006708145142,
      "learning_rate": 0.00012680050905124277,
      "loss": 1.0453,
      "step": 14600
    },
    {
      "epoch": 1.1268363972002153,
      "grad_norm": 1.1240845918655396,
      "learning_rate": 0.0001265407890294263,
      "loss": 1.0742,
      "step": 14650
    },
    {
      "epoch": 1.1306822552111375,
      "grad_norm": 1.433976650238037,
      "learning_rate": 0.0001262810690076098,
      "loss": 1.1468,
      "step": 14700
    },
    {
      "epoch": 1.13452811322206,
      "grad_norm": 1.077966332435608,
      "learning_rate": 0.0001260213489857933,
      "loss": 1.1143,
      "step": 14750
    },
    {
      "epoch": 1.13452811322206,
      "eval_loss": 1.1030505895614624,
      "eval_runtime": 17.8195,
      "eval_samples_per_second": 56.118,
      "eval_steps_per_second": 14.03,
      "step": 14750
    },
    {
      "epoch": 1.1383739712329821,
      "grad_norm": 1.1456421613693237,
      "learning_rate": 0.00012576162896397683,
      "loss": 1.1173,
      "step": 14800
    },
    {
      "epoch": 1.1422198292439043,
      "grad_norm": 0.9305130243301392,
      "learning_rate": 0.00012550190894216037,
      "loss": 1.1158,
      "step": 14850
    },
    {
      "epoch": 1.1460656872548265,
      "grad_norm": 1.3796011209487915,
      "learning_rate": 0.00012524218892034388,
      "loss": 1.0785,
      "step": 14900
    },
    {
      "epoch": 1.149911545265749,
      "grad_norm": 0.9901970028877258,
      "learning_rate": 0.0001249824688985274,
      "loss": 1.1383,
      "step": 14950
    },
    {
      "epoch": 1.153757403276671,
      "grad_norm": 0.8250207304954529,
      "learning_rate": 0.00012472794327714724,
      "loss": 1.1363,
      "step": 15000
    },
    {
      "epoch": 1.153757403276671,
      "eval_loss": 1.103262186050415,
      "eval_runtime": 17.8852,
      "eval_samples_per_second": 55.912,
      "eval_steps_per_second": 13.978,
      "step": 15000
    },
    {
      "epoch": 1.1576032612875933,
      "grad_norm": 0.7949115037918091,
      "learning_rate": 0.00012446822325533076,
      "loss": 1.0542,
      "step": 15050
    },
    {
      "epoch": 1.1614491192985155,
      "grad_norm": 0.8414244055747986,
      "learning_rate": 0.00012420850323351427,
      "loss": 1.133,
      "step": 15100
    },
    {
      "epoch": 1.1652949773094377,
      "grad_norm": 0.7031393647193909,
      "learning_rate": 0.0001239487832116978,
      "loss": 1.0804,
      "step": 15150
    },
    {
      "epoch": 1.16914083532036,
      "grad_norm": 0.8476413488388062,
      "learning_rate": 0.00012368906318988132,
      "loss": 1.1252,
      "step": 15200
    },
    {
      "epoch": 1.1729866933312822,
      "grad_norm": 1.7877459526062012,
      "learning_rate": 0.00012342934316806483,
      "loss": 1.1491,
      "step": 15250
    },
    {
      "epoch": 1.1729866933312822,
      "eval_loss": 1.0958014726638794,
      "eval_runtime": 17.9511,
      "eval_samples_per_second": 55.707,
      "eval_steps_per_second": 13.927,
      "step": 15250
    },
    {
      "epoch": 1.1768325513422044,
      "grad_norm": 0.8797541856765747,
      "learning_rate": 0.00012316962314624835,
      "loss": 1.0446,
      "step": 15300
    },
    {
      "epoch": 1.1806784093531266,
      "grad_norm": 0.9383549690246582,
      "learning_rate": 0.00012290990312443186,
      "loss": 1.0703,
      "step": 15350
    },
    {
      "epoch": 1.1845242673640488,
      "grad_norm": 1.0788028240203857,
      "learning_rate": 0.00012265018310261537,
      "loss": 1.0976,
      "step": 15400
    },
    {
      "epoch": 1.1883701253749712,
      "grad_norm": 0.8661052584648132,
      "learning_rate": 0.0001223904630807989,
      "loss": 1.0872,
      "step": 15450
    },
    {
      "epoch": 1.1922159833858934,
      "grad_norm": 0.9346690773963928,
      "learning_rate": 0.00012213074305898242,
      "loss": 1.1296,
      "step": 15500
    },
    {
      "epoch": 1.1922159833858934,
      "eval_loss": 1.089566946029663,
      "eval_runtime": 17.8688,
      "eval_samples_per_second": 55.963,
      "eval_steps_per_second": 13.991,
      "step": 15500
    },
    {
      "epoch": 1.1960618413968156,
      "grad_norm": 1.229148030281067,
      "learning_rate": 0.00012187102303716594,
      "loss": 1.049,
      "step": 15550
    },
    {
      "epoch": 1.1999076994077378,
      "grad_norm": 0.9896694421768188,
      "learning_rate": 0.00012161130301534946,
      "loss": 1.0519,
      "step": 15600
    },
    {
      "epoch": 1.2037535574186602,
      "grad_norm": 0.7709591388702393,
      "learning_rate": 0.00012135158299353298,
      "loss": 0.8004,
      "step": 15650
    },
    {
      "epoch": 1.2075994154295824,
      "grad_norm": 0.8033544421195984,
      "learning_rate": 0.00012109186297171649,
      "loss": 0.8073,
      "step": 15700
    },
    {
      "epoch": 1.2114452734405046,
      "grad_norm": 0.955243706703186,
      "learning_rate": 0.00012083214294990002,
      "loss": 0.8575,
      "step": 15750
    },
    {
      "epoch": 1.2114452734405046,
      "eval_loss": 0.8102548718452454,
      "eval_runtime": 18.0003,
      "eval_samples_per_second": 55.555,
      "eval_steps_per_second": 13.889,
      "step": 15750
    },
    {
      "epoch": 1.2152911314514268,
      "grad_norm": 1.1329740285873413,
      "learning_rate": 0.00012057242292808353,
      "loss": 0.8421,
      "step": 15800
    },
    {
      "epoch": 1.2191369894623492,
      "grad_norm": 0.6005277633666992,
      "learning_rate": 0.00012031270290626706,
      "loss": 0.8149,
      "step": 15850
    },
    {
      "epoch": 1.2229828474732714,
      "grad_norm": 0.6579030156135559,
      "learning_rate": 0.00012005298288445057,
      "loss": 0.7783,
      "step": 15900
    },
    {
      "epoch": 1.2268287054841935,
      "grad_norm": 1.1820577383041382,
      "learning_rate": 0.00011979326286263408,
      "loss": 0.8133,
      "step": 15950
    },
    {
      "epoch": 1.2306745634951157,
      "grad_norm": 0.7112457156181335,
      "learning_rate": 0.00011953354284081761,
      "loss": 0.8004,
      "step": 16000
    },
    {
      "epoch": 1.2306745634951157,
      "eval_loss": 0.8041366338729858,
      "eval_runtime": 17.8867,
      "eval_samples_per_second": 55.907,
      "eval_steps_per_second": 13.977,
      "step": 16000
    },
    {
      "epoch": 1.234520421506038,
      "grad_norm": 0.7006672024726868,
      "learning_rate": 0.00011927382281900112,
      "loss": 0.7673,
      "step": 16050
    },
    {
      "epoch": 1.2383662795169603,
      "grad_norm": 1.0331398248672485,
      "learning_rate": 0.00011901410279718463,
      "loss": 0.7992,
      "step": 16100
    },
    {
      "epoch": 1.2422121375278825,
      "grad_norm": 0.8857033848762512,
      "learning_rate": 0.00011875438277536817,
      "loss": 0.8189,
      "step": 16150
    },
    {
      "epoch": 1.2460579955388047,
      "grad_norm": 0.6674500107765198,
      "learning_rate": 0.00011849466275355167,
      "loss": 0.7849,
      "step": 16200
    },
    {
      "epoch": 1.2499038535497269,
      "grad_norm": 0.8113058805465698,
      "learning_rate": 0.00011823494273173519,
      "loss": 0.7628,
      "step": 16250
    },
    {
      "epoch": 1.2499038535497269,
      "eval_loss": 0.8041785955429077,
      "eval_runtime": 18.0307,
      "eval_samples_per_second": 55.461,
      "eval_steps_per_second": 13.865,
      "step": 16250
    },
    {
      "epoch": 1.253749711560649,
      "grad_norm": 1.1236894130706787,
      "learning_rate": 0.00011797522270991873,
      "loss": 0.8078,
      "step": 16300
    },
    {
      "epoch": 1.2575955695715715,
      "grad_norm": 0.8019891977310181,
      "learning_rate": 0.00011771550268810222,
      "loss": 0.7999,
      "step": 16350
    },
    {
      "epoch": 1.2614414275824937,
      "grad_norm": 0.6394712924957275,
      "learning_rate": 0.00011745578266628574,
      "loss": 0.8132,
      "step": 16400
    },
    {
      "epoch": 1.2652872855934159,
      "grad_norm": 0.8335860371589661,
      "learning_rate": 0.00011719606264446928,
      "loss": 0.7494,
      "step": 16450
    },
    {
      "epoch": 1.2691331436043383,
      "grad_norm": 0.5847200155258179,
      "learning_rate": 0.00011693634262265279,
      "loss": 0.7966,
      "step": 16500
    },
    {
      "epoch": 1.2691331436043383,
      "eval_loss": 0.8012632131576538,
      "eval_runtime": 17.8773,
      "eval_samples_per_second": 55.937,
      "eval_steps_per_second": 13.984,
      "step": 16500
    },
    {
      "epoch": 1.2729790016152602,
      "grad_norm": 0.7161915302276611,
      "learning_rate": 0.00011667662260083629,
      "loss": 0.7969,
      "step": 16550
    },
    {
      "epoch": 1.2768248596261826,
      "grad_norm": 0.7693639397621155,
      "learning_rate": 0.00011641690257901983,
      "loss": 0.7909,
      "step": 16600
    },
    {
      "epoch": 1.2806707176371048,
      "grad_norm": 0.6645983457565308,
      "learning_rate": 0.00011615718255720334,
      "loss": 0.7816,
      "step": 16650
    },
    {
      "epoch": 1.284516575648027,
      "grad_norm": 0.7736928462982178,
      "learning_rate": 0.00011589746253538687,
      "loss": 0.745,
      "step": 16700
    },
    {
      "epoch": 1.2883624336589494,
      "grad_norm": 0.6743229627609253,
      "learning_rate": 0.00011563774251357038,
      "loss": 0.7856,
      "step": 16750
    },
    {
      "epoch": 1.2883624336589494,
      "eval_loss": 0.8006194233894348,
      "eval_runtime": 17.9046,
      "eval_samples_per_second": 55.852,
      "eval_steps_per_second": 13.963,
      "step": 16750
    },
    {
      "epoch": 1.2922082916698716,
      "grad_norm": 0.8146863579750061,
      "learning_rate": 0.0001153780224917539,
      "loss": 0.7559,
      "step": 16800
    },
    {
      "epoch": 1.2960541496807938,
      "grad_norm": 0.9186645746231079,
      "learning_rate": 0.00011511830246993742,
      "loss": 0.7902,
      "step": 16850
    },
    {
      "epoch": 1.299900007691716,
      "grad_norm": 0.7634202241897583,
      "learning_rate": 0.00011485858244812093,
      "loss": 0.749,
      "step": 16900
    },
    {
      "epoch": 1.3037458657026382,
      "grad_norm": 0.890457272529602,
      "learning_rate": 0.00011459886242630445,
      "loss": 0.7539,
      "step": 16950
    },
    {
      "epoch": 1.3075917237135606,
      "grad_norm": 0.8466306328773499,
      "learning_rate": 0.00011433914240448797,
      "loss": 0.7997,
      "step": 17000
    },
    {
      "epoch": 1.3075917237135606,
      "eval_loss": 0.8002509474754333,
      "eval_runtime": 17.9496,
      "eval_samples_per_second": 55.712,
      "eval_steps_per_second": 13.928,
      "step": 17000
    },
    {
      "epoch": 1.3114375817244828,
      "grad_norm": 1.403998851776123,
      "learning_rate": 0.00011407942238267149,
      "loss": 0.8369,
      "step": 17050
    },
    {
      "epoch": 1.315283439735405,
      "grad_norm": 0.8802525401115417,
      "learning_rate": 0.000113819702360855,
      "loss": 0.7818,
      "step": 17100
    },
    {
      "epoch": 1.3191292977463271,
      "grad_norm": 0.7463178634643555,
      "learning_rate": 0.00011355998233903852,
      "loss": 0.7661,
      "step": 17150
    },
    {
      "epoch": 1.3229751557572493,
      "grad_norm": 0.47771725058555603,
      "learning_rate": 0.00011330026231722204,
      "loss": 0.8141,
      "step": 17200
    },
    {
      "epoch": 1.3268210137681717,
      "grad_norm": 0.8294114470481873,
      "learning_rate": 0.00011304054229540555,
      "loss": 0.7979,
      "step": 17250
    },
    {
      "epoch": 1.3268210137681717,
      "eval_loss": 0.7951701879501343,
      "eval_runtime": 17.958,
      "eval_samples_per_second": 55.686,
      "eval_steps_per_second": 13.921,
      "step": 17250
    },
    {
      "epoch": 1.330666871779094,
      "grad_norm": 1.0813144445419312,
      "learning_rate": 0.00011278082227358908,
      "loss": 0.824,
      "step": 17300
    },
    {
      "epoch": 1.3345127297900161,
      "grad_norm": 0.5647908449172974,
      "learning_rate": 0.00011252110225177259,
      "loss": 0.7718,
      "step": 17350
    },
    {
      "epoch": 1.3383585878009385,
      "grad_norm": 0.7901347279548645,
      "learning_rate": 0.0001122613822299561,
      "loss": 0.7883,
      "step": 17400
    },
    {
      "epoch": 1.3422044458118605,
      "grad_norm": 1.0769431591033936,
      "learning_rate": 0.00011200166220813963,
      "loss": 0.7625,
      "step": 17450
    },
    {
      "epoch": 1.346050303822783,
      "grad_norm": 0.7366082072257996,
      "learning_rate": 0.00011174194218632314,
      "loss": 0.8167,
      "step": 17500
    },
    {
      "epoch": 1.346050303822783,
      "eval_loss": 0.7991219758987427,
      "eval_runtime": 17.9932,
      "eval_samples_per_second": 55.577,
      "eval_steps_per_second": 13.894,
      "step": 17500
    },
    {
      "epoch": 1.349896161833705,
      "grad_norm": 0.8688609600067139,
      "learning_rate": 0.00011148222216450668,
      "loss": 0.7882,
      "step": 17550
    },
    {
      "epoch": 1.3537420198446273,
      "grad_norm": 0.7067499160766602,
      "learning_rate": 0.0001112225021426902,
      "loss": 0.8193,
      "step": 17600
    },
    {
      "epoch": 1.3575878778555497,
      "grad_norm": 0.8119627833366394,
      "learning_rate": 0.0001109627821208737,
      "loss": 0.7902,
      "step": 17650
    },
    {
      "epoch": 1.3614337358664719,
      "grad_norm": 0.4667348265647888,
      "learning_rate": 0.00011070306209905723,
      "loss": 0.7595,
      "step": 17700
    },
    {
      "epoch": 1.365279593877394,
      "grad_norm": 1.1219276189804077,
      "learning_rate": 0.00011044334207724075,
      "loss": 0.8049,
      "step": 17750
    },
    {
      "epoch": 1.365279593877394,
      "eval_loss": 0.7947649955749512,
      "eval_runtime": 17.9375,
      "eval_samples_per_second": 55.749,
      "eval_steps_per_second": 13.937,
      "step": 17750
    },
    {
      "epoch": 1.3691254518883162,
      "grad_norm": 0.8262600302696228,
      "learning_rate": 0.00011018362205542425,
      "loss": 0.794,
      "step": 17800
    },
    {
      "epoch": 1.3729713098992384,
      "grad_norm": 0.6738994121551514,
      "learning_rate": 0.00010992390203360779,
      "loss": 0.7749,
      "step": 17850
    },
    {
      "epoch": 1.3768171679101608,
      "grad_norm": 0.40902426838874817,
      "learning_rate": 0.0001096641820117913,
      "loss": 0.8111,
      "step": 17900
    },
    {
      "epoch": 1.380663025921083,
      "grad_norm": 0.7617566585540771,
      "learning_rate": 0.00010940446198997481,
      "loss": 0.7663,
      "step": 17950
    },
    {
      "epoch": 1.3845088839320052,
      "grad_norm": 0.5269647836685181,
      "learning_rate": 0.00010914474196815834,
      "loss": 0.7913,
      "step": 18000
    },
    {
      "epoch": 1.3845088839320052,
      "eval_loss": 0.7882509231567383,
      "eval_runtime": 17.938,
      "eval_samples_per_second": 55.748,
      "eval_steps_per_second": 13.937,
      "step": 18000
    },
    {
      "epoch": 1.3883547419429274,
      "grad_norm": 0.9016252160072327,
      "learning_rate": 0.00010888502194634185,
      "loss": 0.8222,
      "step": 18050
    },
    {
      "epoch": 1.3922005999538496,
      "grad_norm": 0.6058124899864197,
      "learning_rate": 0.00010862530192452536,
      "loss": 0.7657,
      "step": 18100
    },
    {
      "epoch": 1.396046457964772,
      "grad_norm": 0.8505234122276306,
      "learning_rate": 0.00010836558190270889,
      "loss": 0.7962,
      "step": 18150
    },
    {
      "epoch": 1.3998923159756942,
      "grad_norm": 0.7518420815467834,
      "learning_rate": 0.0001081058618808924,
      "loss": 0.8364,
      "step": 18200
    },
    {
      "epoch": 1.4037381739866164,
      "grad_norm": 0.7778449058532715,
      "learning_rate": 0.00010784614185907592,
      "loss": 0.7836,
      "step": 18250
    },
    {
      "epoch": 1.4037381739866164,
      "eval_loss": 0.7889594435691833,
      "eval_runtime": 17.8326,
      "eval_samples_per_second": 56.077,
      "eval_steps_per_second": 14.019,
      "step": 18250
    },
    {
      "epoch": 1.4075840319975388,
      "grad_norm": 1.2029508352279663,
      "learning_rate": 0.00010758642183725944,
      "loss": 0.7859,
      "step": 18300
    },
    {
      "epoch": 1.4114298900084608,
      "grad_norm": 0.7625166773796082,
      "learning_rate": 0.00010732670181544295,
      "loss": 0.8018,
      "step": 18350
    },
    {
      "epoch": 1.4152757480193832,
      "grad_norm": 0.6327937245368958,
      "learning_rate": 0.00010706698179362648,
      "loss": 0.808,
      "step": 18400
    },
    {
      "epoch": 1.4191216060303053,
      "grad_norm": 0.7097195386886597,
      "learning_rate": 0.00010680726177181,
      "loss": 0.7928,
      "step": 18450
    },
    {
      "epoch": 1.4229674640412275,
      "grad_norm": 0.5188928246498108,
      "learning_rate": 0.00010654754174999351,
      "loss": 0.7934,
      "step": 18500
    },
    {
      "epoch": 1.4229674640412275,
      "eval_loss": 0.7918493151664734,
      "eval_runtime": 17.9735,
      "eval_samples_per_second": 55.638,
      "eval_steps_per_second": 13.909,
      "step": 18500
    },
    {
      "epoch": 1.42681332205215,
      "grad_norm": 0.6486705541610718,
      "learning_rate": 0.00010628782172817703,
      "loss": 0.7317,
      "step": 18550
    },
    {
      "epoch": 1.4306591800630721,
      "grad_norm": 0.6570118069648743,
      "learning_rate": 0.00010602810170636055,
      "loss": 0.8149,
      "step": 18600
    },
    {
      "epoch": 1.4345050380739943,
      "grad_norm": 0.8024285435676575,
      "learning_rate": 0.00010576838168454406,
      "loss": 0.8392,
      "step": 18650
    },
    {
      "epoch": 1.4383508960849165,
      "grad_norm": 0.5735141038894653,
      "learning_rate": 0.0001055086616627276,
      "loss": 0.7932,
      "step": 18700
    },
    {
      "epoch": 1.4421967540958387,
      "grad_norm": 0.5087122917175293,
      "learning_rate": 0.0001052489416409111,
      "loss": 0.7508,
      "step": 18750
    },
    {
      "epoch": 1.4421967540958387,
      "eval_loss": 0.7895762920379639,
      "eval_runtime": 17.9534,
      "eval_samples_per_second": 55.7,
      "eval_steps_per_second": 13.925,
      "step": 18750
    },
    {
      "epoch": 1.446042612106761,
      "grad_norm": 0.7478468418121338,
      "learning_rate": 0.00010498922161909461,
      "loss": 0.8533,
      "step": 18800
    },
    {
      "epoch": 1.4498884701176833,
      "grad_norm": 1.2165393829345703,
      "learning_rate": 0.00010472950159727815,
      "loss": 0.8025,
      "step": 18850
    },
    {
      "epoch": 1.4537343281286055,
      "grad_norm": 0.8682180643081665,
      "learning_rate": 0.00010446978157546165,
      "loss": 0.8086,
      "step": 18900
    },
    {
      "epoch": 1.4575801861395277,
      "grad_norm": 0.9063705205917358,
      "learning_rate": 0.00010421006155364516,
      "loss": 0.7583,
      "step": 18950
    },
    {
      "epoch": 1.4614260441504499,
      "grad_norm": 0.7133361101150513,
      "learning_rate": 0.0001039503415318287,
      "loss": 0.7381,
      "step": 19000
    },
    {
      "epoch": 1.4614260441504499,
      "eval_loss": 0.7869898080825806,
      "eval_runtime": 17.9874,
      "eval_samples_per_second": 55.595,
      "eval_steps_per_second": 13.899,
      "step": 19000
    },
    {
      "epoch": 1.4652719021613723,
      "grad_norm": 0.6205143928527832,
      "learning_rate": 0.00010369062151001222,
      "loss": 0.8141,
      "step": 19050
    },
    {
      "epoch": 1.4691177601722945,
      "grad_norm": 1.1060974597930908,
      "learning_rate": 0.00010343090148819572,
      "loss": 0.7771,
      "step": 19100
    },
    {
      "epoch": 1.4729636181832166,
      "grad_norm": 0.7808921933174133,
      "learning_rate": 0.00010317118146637926,
      "loss": 0.8006,
      "step": 19150
    },
    {
      "epoch": 1.476809476194139,
      "grad_norm": 0.5509454011917114,
      "learning_rate": 0.00010291146144456277,
      "loss": 0.7922,
      "step": 19200
    },
    {
      "epoch": 1.480655334205061,
      "grad_norm": 1.2427464723587036,
      "learning_rate": 0.0001026517414227463,
      "loss": 0.7801,
      "step": 19250
    },
    {
      "epoch": 1.480655334205061,
      "eval_loss": 0.7850660085678101,
      "eval_runtime": 17.8458,
      "eval_samples_per_second": 56.036,
      "eval_steps_per_second": 14.009,
      "step": 19250
    },
    {
      "epoch": 1.4845011922159834,
      "grad_norm": 0.8217543959617615,
      "learning_rate": 0.00010239202140092981,
      "loss": 0.782,
      "step": 19300
    },
    {
      "epoch": 1.4883470502269056,
      "grad_norm": 0.6753976941108704,
      "learning_rate": 0.00010213230137911332,
      "loss": 0.7787,
      "step": 19350
    },
    {
      "epoch": 1.4921929082378278,
      "grad_norm": 1.0264923572540283,
      "learning_rate": 0.00010187258135729685,
      "loss": 0.7713,
      "step": 19400
    },
    {
      "epoch": 1.4960387662487502,
      "grad_norm": 0.7317076325416565,
      "learning_rate": 0.00010161286133548036,
      "loss": 0.7634,
      "step": 19450
    },
    {
      "epoch": 1.4998846242596724,
      "grad_norm": 0.6011013388633728,
      "learning_rate": 0.00010135314131366387,
      "loss": 0.7715,
      "step": 19500
    },
    {
      "epoch": 1.4998846242596724,
      "eval_loss": 0.7843549847602844,
      "eval_runtime": 17.919,
      "eval_samples_per_second": 55.807,
      "eval_steps_per_second": 13.952,
      "step": 19500
    },
    {
      "epoch": 1.5037304822705946,
      "grad_norm": 0.5136408805847168,
      "learning_rate": 0.0001010934212918474,
      "loss": 0.7533,
      "step": 19550
    },
    {
      "epoch": 1.5075763402815168,
      "grad_norm": 0.8944385647773743,
      "learning_rate": 0.00010083370127003091,
      "loss": 0.8483,
      "step": 19600
    },
    {
      "epoch": 1.511422198292439,
      "grad_norm": 0.9759209752082825,
      "learning_rate": 0.00010057398124821442,
      "loss": 0.8338,
      "step": 19650
    },
    {
      "epoch": 1.5152680563033614,
      "grad_norm": 0.7823068499565125,
      "learning_rate": 0.00010031426122639795,
      "loss": 0.7739,
      "step": 19700
    },
    {
      "epoch": 1.5191139143142836,
      "grad_norm": 0.9423583745956421,
      "learning_rate": 0.00010005454120458146,
      "loss": 0.8092,
      "step": 19750
    },
    {
      "epoch": 1.5191139143142836,
      "eval_loss": 0.782052218914032,
      "eval_runtime": 17.8818,
      "eval_samples_per_second": 55.923,
      "eval_steps_per_second": 13.981,
      "step": 19750
    },
    {
      "epoch": 1.5229597723252057,
      "grad_norm": 1.067384123802185,
      "learning_rate": 9.979482118276499e-05,
      "loss": 0.7581,
      "step": 19800
    },
    {
      "epoch": 1.5268056303361282,
      "grad_norm": 1.0136195421218872,
      "learning_rate": 9.95351011609485e-05,
      "loss": 0.7923,
      "step": 19850
    },
    {
      "epoch": 1.5306514883470501,
      "grad_norm": 0.8290442824363708,
      "learning_rate": 9.927538113913202e-05,
      "loss": 0.7991,
      "step": 19900
    },
    {
      "epoch": 1.5344973463579725,
      "grad_norm": 0.8200196623802185,
      "learning_rate": 9.901566111731554e-05,
      "loss": 0.7382,
      "step": 19950
    },
    {
      "epoch": 1.5383432043688947,
      "grad_norm": 0.7905208468437195,
      "learning_rate": 9.875594109549905e-05,
      "loss": 0.7615,
      "step": 20000
    },
    {
      "epoch": 1.5383432043688947,
      "eval_loss": 0.7793118953704834,
      "eval_runtime": 17.8391,
      "eval_samples_per_second": 56.057,
      "eval_steps_per_second": 14.014,
      "step": 20000
    },
    {
      "epoch": 1.542189062379817,
      "grad_norm": 0.8522188067436218,
      "learning_rate": 9.849622107368257e-05,
      "loss": 0.8041,
      "step": 20050
    },
    {
      "epoch": 1.5460349203907393,
      "grad_norm": 1.0029702186584473,
      "learning_rate": 9.82365010518661e-05,
      "loss": 0.7873,
      "step": 20100
    },
    {
      "epoch": 1.5498807784016613,
      "grad_norm": 1.007730484008789,
      "learning_rate": 9.797678103004962e-05,
      "loss": 0.7685,
      "step": 20150
    },
    {
      "epoch": 1.5537266364125837,
      "grad_norm": 0.6971302032470703,
      "learning_rate": 9.771706100823312e-05,
      "loss": 0.805,
      "step": 20200
    },
    {
      "epoch": 1.5575724944235059,
      "grad_norm": 0.9766409993171692,
      "learning_rate": 9.745734098641665e-05,
      "loss": 0.765,
      "step": 20250
    },
    {
      "epoch": 1.5575724944235059,
      "eval_loss": 0.7780515551567078,
      "eval_runtime": 17.9362,
      "eval_samples_per_second": 55.753,
      "eval_steps_per_second": 13.938,
      "step": 20250
    },
    {
      "epoch": 1.561418352434428,
      "grad_norm": 0.8611739873886108,
      "learning_rate": 9.719762096460017e-05,
      "loss": 0.7768,
      "step": 20300
    },
    {
      "epoch": 1.5652642104453505,
      "grad_norm": 0.9571949243545532,
      "learning_rate": 9.693790094278369e-05,
      "loss": 0.8106,
      "step": 20350
    },
    {
      "epoch": 1.5691100684562724,
      "grad_norm": 0.9941520094871521,
      "learning_rate": 9.66781809209672e-05,
      "loss": 0.7952,
      "step": 20400
    },
    {
      "epoch": 1.5729559264671948,
      "grad_norm": 0.7494879364967346,
      "learning_rate": 9.641846089915072e-05,
      "loss": 0.7882,
      "step": 20450
    },
    {
      "epoch": 1.576801784478117,
      "grad_norm": 0.25608131289482117,
      "learning_rate": 9.615874087733424e-05,
      "loss": 0.7424,
      "step": 20500
    },
    {
      "epoch": 1.576801784478117,
      "eval_loss": 0.7784542441368103,
      "eval_runtime": 17.9827,
      "eval_samples_per_second": 55.609,
      "eval_steps_per_second": 13.902,
      "step": 20500
    },
    {
      "epoch": 1.5806476424890392,
      "grad_norm": 0.781563401222229,
      "learning_rate": 9.589902085551775e-05,
      "loss": 0.7251,
      "step": 20550
    },
    {
      "epoch": 1.5844935004999616,
      "grad_norm": 0.8129003047943115,
      "learning_rate": 9.563930083370128e-05,
      "loss": 0.7783,
      "step": 20600
    },
    {
      "epoch": 1.5883393585108838,
      "grad_norm": 0.7955138087272644,
      "learning_rate": 9.537958081188479e-05,
      "loss": 0.7692,
      "step": 20650
    },
    {
      "epoch": 1.592185216521806,
      "grad_norm": 0.7752518653869629,
      "learning_rate": 9.511986079006832e-05,
      "loss": 0.7578,
      "step": 20700
    },
    {
      "epoch": 1.5960310745327284,
      "grad_norm": 0.7433210611343384,
      "learning_rate": 9.486014076825183e-05,
      "loss": 0.7885,
      "step": 20750
    },
    {
      "epoch": 1.5960310745327284,
      "eval_loss": 0.7757794260978699,
      "eval_runtime": 17.9593,
      "eval_samples_per_second": 55.682,
      "eval_steps_per_second": 13.92,
      "step": 20750
    },
    {
      "epoch": 1.5998769325436504,
      "grad_norm": 0.7218450903892517,
      "learning_rate": 9.460042074643534e-05,
      "loss": 0.8179,
      "step": 20800
    },
    {
      "epoch": 1.6037227905545728,
      "grad_norm": 0.8611409664154053,
      "learning_rate": 9.434070072461887e-05,
      "loss": 0.8166,
      "step": 20850
    },
    {
      "epoch": 1.607568648565495,
      "grad_norm": 0.8060470223426819,
      "learning_rate": 9.408098070280238e-05,
      "loss": 0.7831,
      "step": 20900
    },
    {
      "epoch": 1.6114145065764172,
      "grad_norm": 0.9832955002784729,
      "learning_rate": 9.382126068098591e-05,
      "loss": 0.7679,
      "step": 20950
    },
    {
      "epoch": 1.6152603645873396,
      "grad_norm": 0.7749195098876953,
      "learning_rate": 9.356154065916942e-05,
      "loss": 0.8001,
      "step": 21000
    },
    {
      "epoch": 1.6152603645873396,
      "eval_loss": 0.7739425301551819,
      "eval_runtime": 17.9752,
      "eval_samples_per_second": 55.632,
      "eval_steps_per_second": 13.908,
      "step": 21000
    },
    {
      "epoch": 1.6191062225982615,
      "grad_norm": 0.8098833560943604,
      "learning_rate": 9.330182063735295e-05,
      "loss": 0.7673,
      "step": 21050
    },
    {
      "epoch": 1.622952080609184,
      "grad_norm": 0.7867510318756104,
      "learning_rate": 9.304210061553646e-05,
      "loss": 0.7619,
      "step": 21100
    },
    {
      "epoch": 1.6267979386201061,
      "grad_norm": 1.2399013042449951,
      "learning_rate": 9.278238059371997e-05,
      "loss": 0.8664,
      "step": 21150
    },
    {
      "epoch": 1.6306437966310283,
      "grad_norm": 0.6171821355819702,
      "learning_rate": 9.25226605719035e-05,
      "loss": 0.8543,
      "step": 21200
    },
    {
      "epoch": 1.6344896546419507,
      "grad_norm": 0.7197456955909729,
      "learning_rate": 9.226294055008701e-05,
      "loss": 0.7989,
      "step": 21250
    },
    {
      "epoch": 1.6344896546419507,
      "eval_loss": 0.7736220955848694,
      "eval_runtime": 18.0274,
      "eval_samples_per_second": 55.471,
      "eval_steps_per_second": 13.868,
      "step": 21250
    },
    {
      "epoch": 1.6383355126528727,
      "grad_norm": 0.6752798557281494,
      "learning_rate": 9.200322052827052e-05,
      "loss": 0.7925,
      "step": 21300
    },
    {
      "epoch": 1.642181370663795,
      "grad_norm": 0.7389090061187744,
      "learning_rate": 9.174350050645405e-05,
      "loss": 0.762,
      "step": 21350
    },
    {
      "epoch": 1.6460272286747173,
      "grad_norm": 0.7688984870910645,
      "learning_rate": 9.148378048463756e-05,
      "loss": 0.771,
      "step": 21400
    },
    {
      "epoch": 1.6498730866856395,
      "grad_norm": 0.7231914401054382,
      "learning_rate": 9.122406046282108e-05,
      "loss": 0.8244,
      "step": 21450
    },
    {
      "epoch": 1.6537189446965619,
      "grad_norm": 0.786527156829834,
      "learning_rate": 9.09643404410046e-05,
      "loss": 0.7738,
      "step": 21500
    },
    {
      "epoch": 1.6537189446965619,
      "eval_loss": 0.77164626121521,
      "eval_runtime": 17.9322,
      "eval_samples_per_second": 55.765,
      "eval_steps_per_second": 13.941,
      "step": 21500
    },
    {
      "epoch": 1.657564802707484,
      "grad_norm": 0.8676180243492126,
      "learning_rate": 9.070462041918813e-05,
      "loss": 0.8113,
      "step": 21550
    },
    {
      "epoch": 1.6614106607184063,
      "grad_norm": 0.6232962012290955,
      "learning_rate": 9.044490039737163e-05,
      "loss": 0.7847,
      "step": 21600
    },
    {
      "epoch": 1.6652565187293287,
      "grad_norm": 1.151957631111145,
      "learning_rate": 9.018518037555515e-05,
      "loss": 0.7924,
      "step": 21650
    },
    {
      "epoch": 1.6691023767402506,
      "grad_norm": 0.8149247169494629,
      "learning_rate": 8.992546035373868e-05,
      "loss": 0.7395,
      "step": 21700
    },
    {
      "epoch": 1.672948234751173,
      "grad_norm": 1.0983916521072388,
      "learning_rate": 8.96657403319222e-05,
      "loss": 0.7715,
      "step": 21750
    },
    {
      "epoch": 1.672948234751173,
      "eval_loss": 0.7681905031204224,
      "eval_runtime": 17.9279,
      "eval_samples_per_second": 55.779,
      "eval_steps_per_second": 13.945,
      "step": 21750
    },
    {
      "epoch": 1.6767940927620952,
      "grad_norm": 0.7059574723243713,
      "learning_rate": 8.94060203101057e-05,
      "loss": 0.784,
      "step": 21800
    },
    {
      "epoch": 1.6806399507730174,
      "grad_norm": 1.0587022304534912,
      "learning_rate": 8.914630028828923e-05,
      "loss": 0.7608,
      "step": 21850
    },
    {
      "epoch": 1.6844858087839398,
      "grad_norm": 0.7638582587242126,
      "learning_rate": 8.888658026647275e-05,
      "loss": 0.8159,
      "step": 21900
    },
    {
      "epoch": 1.6883316667948618,
      "grad_norm": 0.5783549547195435,
      "learning_rate": 8.862686024465626e-05,
      "loss": 0.7959,
      "step": 21950
    },
    {
      "epoch": 1.6921775248057842,
      "grad_norm": 1.2192896604537964,
      "learning_rate": 8.836714022283978e-05,
      "loss": 0.8272,
      "step": 22000
    },
    {
      "epoch": 1.6921775248057842,
      "eval_loss": 0.7701475024223328,
      "eval_runtime": 17.8145,
      "eval_samples_per_second": 56.134,
      "eval_steps_per_second": 14.034,
      "step": 22000
    },
    {
      "epoch": 1.6960233828167064,
      "grad_norm": 0.5081881880760193,
      "learning_rate": 8.810742020102331e-05,
      "loss": 0.7883,
      "step": 22050
    },
    {
      "epoch": 1.6998692408276286,
      "grad_norm": 0.5658268332481384,
      "learning_rate": 8.784770017920681e-05,
      "loss": 0.77,
      "step": 22100
    },
    {
      "epoch": 1.703715098838551,
      "grad_norm": 0.6888287663459778,
      "learning_rate": 8.758798015739034e-05,
      "loss": 0.8013,
      "step": 22150
    },
    {
      "epoch": 1.707560956849473,
      "grad_norm": 1.0599181652069092,
      "learning_rate": 8.732826013557386e-05,
      "loss": 0.7649,
      "step": 22200
    },
    {
      "epoch": 1.7114068148603954,
      "grad_norm": 0.7679368257522583,
      "learning_rate": 8.706854011375736e-05,
      "loss": 0.7338,
      "step": 22250
    },
    {
      "epoch": 1.7114068148603954,
      "eval_loss": 0.7665286660194397,
      "eval_runtime": 17.8822,
      "eval_samples_per_second": 55.921,
      "eval_steps_per_second": 13.98,
      "step": 22250
    },
    {
      "epoch": 1.7152526728713176,
      "grad_norm": 0.6636808514595032,
      "learning_rate": 8.680882009194089e-05,
      "loss": 0.7949,
      "step": 22300
    },
    {
      "epoch": 1.7190985308822397,
      "grad_norm": 0.7327821254730225,
      "learning_rate": 8.654910007012442e-05,
      "loss": 0.7484,
      "step": 22350
    },
    {
      "epoch": 1.7229443888931621,
      "grad_norm": 0.8187472224235535,
      "learning_rate": 8.628938004830793e-05,
      "loss": 0.7482,
      "step": 22400
    },
    {
      "epoch": 1.7267902469040843,
      "grad_norm": 0.4527030289173126,
      "learning_rate": 8.602966002649144e-05,
      "loss": 0.7902,
      "step": 22450
    },
    {
      "epoch": 1.7306361049150065,
      "grad_norm": 0.6475220918655396,
      "learning_rate": 8.576994000467497e-05,
      "loss": 0.7998,
      "step": 22500
    },
    {
      "epoch": 1.7306361049150065,
      "eval_loss": 0.7646552920341492,
      "eval_runtime": 17.8108,
      "eval_samples_per_second": 56.146,
      "eval_steps_per_second": 14.036,
      "step": 22500
    },
    {
      "epoch": 1.734481962925929,
      "grad_norm": 0.538769543170929,
      "learning_rate": 8.551021998285848e-05,
      "loss": 0.7658,
      "step": 22550
    },
    {
      "epoch": 1.738327820936851,
      "grad_norm": 0.629510223865509,
      "learning_rate": 8.5250499961042e-05,
      "loss": 0.7885,
      "step": 22600
    },
    {
      "epoch": 1.7421736789477733,
      "grad_norm": 0.6914022564888,
      "learning_rate": 8.499077993922552e-05,
      "loss": 0.7648,
      "step": 22650
    },
    {
      "epoch": 1.7460195369586955,
      "grad_norm": 0.5563036799430847,
      "learning_rate": 8.473105991740903e-05,
      "loss": 0.7558,
      "step": 22700
    },
    {
      "epoch": 1.7498653949696177,
      "grad_norm": 0.7851826548576355,
      "learning_rate": 8.447133989559256e-05,
      "loss": 0.7961,
      "step": 22750
    },
    {
      "epoch": 1.7498653949696177,
      "eval_loss": 0.7664644718170166,
      "eval_runtime": 17.9733,
      "eval_samples_per_second": 55.638,
      "eval_steps_per_second": 13.91,
      "step": 22750
    },
    {
      "epoch": 1.75371125298054,
      "grad_norm": 0.48695698380470276,
      "learning_rate": 8.421161987377607e-05,
      "loss": 0.8264,
      "step": 22800
    },
    {
      "epoch": 1.757557110991462,
      "grad_norm": 0.8053486347198486,
      "learning_rate": 8.39518998519596e-05,
      "loss": 0.8084,
      "step": 22850
    },
    {
      "epoch": 1.7614029690023845,
      "grad_norm": 1.1373741626739502,
      "learning_rate": 8.369217983014311e-05,
      "loss": 0.7842,
      "step": 22900
    },
    {
      "epoch": 1.7652488270133067,
      "grad_norm": 1.1318634748458862,
      "learning_rate": 8.343245980832662e-05,
      "loss": 0.7727,
      "step": 22950
    },
    {
      "epoch": 1.7690946850242288,
      "grad_norm": 0.8140521049499512,
      "learning_rate": 8.317273978651015e-05,
      "loss": 0.7945,
      "step": 23000
    },
    {
      "epoch": 1.7690946850242288,
      "eval_loss": 0.7635026574134827,
      "eval_runtime": 17.8969,
      "eval_samples_per_second": 55.875,
      "eval_steps_per_second": 13.969,
      "step": 23000
    },
    {
      "epoch": 1.7729405430351513,
      "grad_norm": 0.7365099787712097,
      "learning_rate": 8.291301976469366e-05,
      "loss": 0.7915,
      "step": 23050
    },
    {
      "epoch": 1.7767864010460732,
      "grad_norm": 0.7158268690109253,
      "learning_rate": 8.265329974287718e-05,
      "loss": 0.8089,
      "step": 23100
    },
    {
      "epoch": 1.7806322590569956,
      "grad_norm": 0.7917172312736511,
      "learning_rate": 8.23935797210607e-05,
      "loss": 0.7881,
      "step": 23150
    },
    {
      "epoch": 1.7844781170679178,
      "grad_norm": 0.9002280831336975,
      "learning_rate": 8.213385969924422e-05,
      "loss": 0.7861,
      "step": 23200
    },
    {
      "epoch": 1.78832397507884,
      "grad_norm": 1.118498682975769,
      "learning_rate": 8.187413967742774e-05,
      "loss": 0.7787,
      "step": 23250
    },
    {
      "epoch": 1.78832397507884,
      "eval_loss": 0.7597912549972534,
      "eval_runtime": 18.0126,
      "eval_samples_per_second": 55.517,
      "eval_steps_per_second": 13.879,
      "step": 23250
    },
    {
      "epoch": 1.7921698330897624,
      "grad_norm": 0.555014967918396,
      "learning_rate": 8.161441965561125e-05,
      "loss": 0.7698,
      "step": 23300
    },
    {
      "epoch": 1.7960156911006846,
      "grad_norm": 0.7749983072280884,
      "learning_rate": 8.135469963379477e-05,
      "loss": 0.7976,
      "step": 23350
    },
    {
      "epoch": 1.7998615491116068,
      "grad_norm": 0.8833787441253662,
      "learning_rate": 8.10949796119783e-05,
      "loss": 0.7901,
      "step": 23400
    },
    {
      "epoch": 1.8037074071225292,
      "grad_norm": 1.099992036819458,
      "learning_rate": 8.08352595901618e-05,
      "loss": 0.7843,
      "step": 23450
    },
    {
      "epoch": 1.8075532651334512,
      "grad_norm": 0.7118529677391052,
      "learning_rate": 8.057553956834533e-05,
      "loss": 0.7818,
      "step": 23500
    },
    {
      "epoch": 1.8075532651334512,
      "eval_loss": 0.7594859600067139,
      "eval_runtime": 17.8501,
      "eval_samples_per_second": 56.022,
      "eval_steps_per_second": 14.005,
      "step": 23500
    },
    {
      "epoch": 1.8113991231443736,
      "grad_norm": 0.8289865851402283,
      "learning_rate": 8.031581954652885e-05,
      "loss": 0.7899,
      "step": 23550
    },
    {
      "epoch": 1.8152449811552958,
      "grad_norm": 1.1237398386001587,
      "learning_rate": 8.005609952471237e-05,
      "loss": 0.8005,
      "step": 23600
    },
    {
      "epoch": 1.819090839166218,
      "grad_norm": 0.8594374060630798,
      "learning_rate": 7.979637950289588e-05,
      "loss": 0.7765,
      "step": 23650
    },
    {
      "epoch": 1.8229366971771404,
      "grad_norm": 0.754634439945221,
      "learning_rate": 7.95366594810794e-05,
      "loss": 0.7794,
      "step": 23700
    },
    {
      "epoch": 1.8267825551880623,
      "grad_norm": 1.0647647380828857,
      "learning_rate": 7.927693945926292e-05,
      "loss": 0.7681,
      "step": 23750
    },
    {
      "epoch": 1.8267825551880623,
      "eval_loss": 0.7575324773788452,
      "eval_runtime": 17.8657,
      "eval_samples_per_second": 55.973,
      "eval_steps_per_second": 13.993,
      "step": 23750
    },
    {
      "epoch": 1.8306284131989847,
      "grad_norm": 1.1255161762237549,
      "learning_rate": 7.901721943744644e-05,
      "loss": 0.7551,
      "step": 23800
    },
    {
      "epoch": 1.834474271209907,
      "grad_norm": 0.8209452629089355,
      "learning_rate": 7.875749941562995e-05,
      "loss": 0.8091,
      "step": 23850
    },
    {
      "epoch": 1.838320129220829,
      "grad_norm": 0.40779542922973633,
      "learning_rate": 7.849777939381348e-05,
      "loss": 0.7572,
      "step": 23900
    },
    {
      "epoch": 1.8421659872317515,
      "grad_norm": 0.9186558127403259,
      "learning_rate": 7.823805937199699e-05,
      "loss": 0.7681,
      "step": 23950
    },
    {
      "epoch": 1.8460118452426735,
      "grad_norm": 0.4896409213542938,
      "learning_rate": 7.79783393501805e-05,
      "loss": 0.7989,
      "step": 24000
    },
    {
      "epoch": 1.8460118452426735,
      "eval_loss": 0.7566245198249817,
      "eval_runtime": 17.9055,
      "eval_samples_per_second": 55.849,
      "eval_steps_per_second": 13.962,
      "step": 24000
    },
    {
      "epoch": 1.8498577032535959,
      "grad_norm": 1.1524229049682617,
      "learning_rate": 7.771861932836403e-05,
      "loss": 0.7732,
      "step": 24050
    },
    {
      "epoch": 1.853703561264518,
      "grad_norm": 0.4653956890106201,
      "learning_rate": 7.745889930654755e-05,
      "loss": 0.7961,
      "step": 24100
    },
    {
      "epoch": 1.8575494192754403,
      "grad_norm": 0.8423280119895935,
      "learning_rate": 7.719917928473105e-05,
      "loss": 0.7963,
      "step": 24150
    },
    {
      "epoch": 1.8613952772863627,
      "grad_norm": 0.6979435086250305,
      "learning_rate": 7.693945926291458e-05,
      "loss": 0.7816,
      "step": 24200
    },
    {
      "epoch": 1.8652411352972849,
      "grad_norm": 0.7800914645195007,
      "learning_rate": 7.66797392410981e-05,
      "loss": 0.7563,
      "step": 24250
    },
    {
      "epoch": 1.8652411352972849,
      "eval_loss": 0.7574715614318848,
      "eval_runtime": 17.9043,
      "eval_samples_per_second": 55.852,
      "eval_steps_per_second": 13.963,
      "step": 24250
    },
    {
      "epoch": 1.869086993308207,
      "grad_norm": 0.9678017497062683,
      "learning_rate": 7.642001921928162e-05,
      "loss": 0.7689,
      "step": 24300
    },
    {
      "epoch": 1.8729328513191295,
      "grad_norm": 0.454647421836853,
      "learning_rate": 7.616029919746513e-05,
      "loss": 0.7231,
      "step": 24350
    },
    {
      "epoch": 1.8767787093300514,
      "grad_norm": 0.7899460792541504,
      "learning_rate": 7.590057917564866e-05,
      "loss": 0.7471,
      "step": 24400
    },
    {
      "epoch": 1.8806245673409738,
      "grad_norm": 1.1373926401138306,
      "learning_rate": 7.564085915383217e-05,
      "loss": 0.8342,
      "step": 24450
    },
    {
      "epoch": 1.884470425351896,
      "grad_norm": 1.4272133111953735,
      "learning_rate": 7.538113913201568e-05,
      "loss": 0.7585,
      "step": 24500
    },
    {
      "epoch": 1.884470425351896,
      "eval_loss": 0.7536496520042419,
      "eval_runtime": 17.9401,
      "eval_samples_per_second": 55.741,
      "eval_steps_per_second": 13.935,
      "step": 24500
    },
    {
      "epoch": 1.8883162833628182,
      "grad_norm": 0.8080185055732727,
      "learning_rate": 7.512141911019921e-05,
      "loss": 0.8067,
      "step": 24550
    },
    {
      "epoch": 1.8921621413737406,
      "grad_norm": 0.5850221514701843,
      "learning_rate": 7.486169908838274e-05,
      "loss": 0.7455,
      "step": 24600
    },
    {
      "epoch": 1.8960079993846626,
      "grad_norm": 0.7521384954452515,
      "learning_rate": 7.460197906656624e-05,
      "loss": 0.7604,
      "step": 24650
    },
    {
      "epoch": 1.899853857395585,
      "grad_norm": 0.6376401782035828,
      "learning_rate": 7.434225904474976e-05,
      "loss": 0.7485,
      "step": 24700
    },
    {
      "epoch": 1.9036997154065072,
      "grad_norm": 0.6305235624313354,
      "learning_rate": 7.408253902293329e-05,
      "loss": 0.7567,
      "step": 24750
    },
    {
      "epoch": 1.9036997154065072,
      "eval_loss": 0.7527515888214111,
      "eval_runtime": 17.967,
      "eval_samples_per_second": 55.658,
      "eval_steps_per_second": 13.914,
      "step": 24750
    },
    {
      "epoch": 1.9075455734174294,
      "grad_norm": 0.5975210666656494,
      "learning_rate": 7.382281900111679e-05,
      "loss": 0.7855,
      "step": 24800
    },
    {
      "epoch": 1.9113914314283518,
      "grad_norm": 0.41196370124816895,
      "learning_rate": 7.356309897930031e-05,
      "loss": 0.7711,
      "step": 24850
    },
    {
      "epoch": 1.9152372894392737,
      "grad_norm": 1.1755207777023315,
      "learning_rate": 7.330337895748384e-05,
      "loss": 0.8114,
      "step": 24900
    },
    {
      "epoch": 1.9190831474501961,
      "grad_norm": 0.37193945050239563,
      "learning_rate": 7.304365893566735e-05,
      "loss": 0.7437,
      "step": 24950
    },
    {
      "epoch": 1.9229290054611183,
      "grad_norm": 0.6753848195075989,
      "learning_rate": 7.278393891385087e-05,
      "loss": 0.8289,
      "step": 25000
    },
    {
      "epoch": 1.9229290054611183,
      "eval_loss": 0.7534742951393127,
      "eval_runtime": 17.8978,
      "eval_samples_per_second": 55.873,
      "eval_steps_per_second": 13.968,
      "step": 25000
    },
    {
      "epoch": 1.9267748634720405,
      "grad_norm": 0.6314563751220703,
      "learning_rate": 7.25242188920344e-05,
      "loss": 0.8278,
      "step": 25050
    },
    {
      "epoch": 1.930620721482963,
      "grad_norm": 0.38249602913856506,
      "learning_rate": 7.22644988702179e-05,
      "loss": 0.7948,
      "step": 25100
    },
    {
      "epoch": 1.9344665794938851,
      "grad_norm": 0.8241211771965027,
      "learning_rate": 7.200477884840142e-05,
      "loss": 0.8152,
      "step": 25150
    },
    {
      "epoch": 1.9383124375048073,
      "grad_norm": 0.4248273968696594,
      "learning_rate": 7.175025322702127e-05,
      "loss": 0.8083,
      "step": 25200
    },
    {
      "epoch": 1.9421582955157297,
      "grad_norm": 1.0574986934661865,
      "learning_rate": 7.14905332052048e-05,
      "loss": 0.7644,
      "step": 25250
    },
    {
      "epoch": 1.9421582955157297,
      "eval_loss": 0.7520478963851929,
      "eval_runtime": 17.9125,
      "eval_samples_per_second": 55.827,
      "eval_steps_per_second": 13.957,
      "step": 25250
    },
    {
      "epoch": 1.9460041535266517,
      "grad_norm": 0.957831084728241,
      "learning_rate": 7.123081318338831e-05,
      "loss": 0.8145,
      "step": 25300
    },
    {
      "epoch": 1.949850011537574,
      "grad_norm": 0.7300383448600769,
      "learning_rate": 7.097109316157184e-05,
      "loss": 0.8021,
      "step": 25350
    },
    {
      "epoch": 1.9536958695484963,
      "grad_norm": 0.6103696227073669,
      "learning_rate": 7.071137313975535e-05,
      "loss": 0.7561,
      "step": 25400
    },
    {
      "epoch": 1.9575417275594185,
      "grad_norm": 0.9353188276290894,
      "learning_rate": 7.045165311793886e-05,
      "loss": 0.76,
      "step": 25450
    },
    {
      "epoch": 1.9613875855703409,
      "grad_norm": 0.7097103595733643,
      "learning_rate": 7.019193309612239e-05,
      "loss": 0.8125,
      "step": 25500
    },
    {
      "epoch": 1.9613875855703409,
      "eval_loss": 0.7521655559539795,
      "eval_runtime": 17.8845,
      "eval_samples_per_second": 55.914,
      "eval_steps_per_second": 13.979,
      "step": 25500
    },
    {
      "epoch": 1.9652334435812628,
      "grad_norm": 1.2154541015625,
      "learning_rate": 6.99322130743059e-05,
      "loss": 0.7644,
      "step": 25550
    },
    {
      "epoch": 1.9690793015921852,
      "grad_norm": 0.715004026889801,
      "learning_rate": 6.967249305248941e-05,
      "loss": 0.7235,
      "step": 25600
    },
    {
      "epoch": 1.9729251596031074,
      "grad_norm": 0.44530218839645386,
      "learning_rate": 6.941277303067294e-05,
      "loss": 0.7591,
      "step": 25650
    },
    {
      "epoch": 1.9767710176140296,
      "grad_norm": 0.7103247046470642,
      "learning_rate": 6.915305300885647e-05,
      "loss": 0.7203,
      "step": 25700
    },
    {
      "epoch": 1.980616875624952,
      "grad_norm": 0.6260993480682373,
      "learning_rate": 6.88985273874763e-05,
      "loss": 0.8008,
      "step": 25750
    },
    {
      "epoch": 1.980616875624952,
      "eval_loss": 0.7489978075027466,
      "eval_runtime": 17.8489,
      "eval_samples_per_second": 56.026,
      "eval_steps_per_second": 14.006,
      "step": 25750
    },
    {
      "epoch": 1.984462733635874,
      "grad_norm": 0.8690526485443115,
      "learning_rate": 6.863880736565982e-05,
      "loss": 0.7445,
      "step": 25800
    },
    {
      "epoch": 1.9883085916467964,
      "grad_norm": 0.8287826776504517,
      "learning_rate": 6.837908734384334e-05,
      "loss": 0.7606,
      "step": 25850
    },
    {
      "epoch": 1.9921544496577186,
      "grad_norm": 0.9105169773101807,
      "learning_rate": 6.811936732202686e-05,
      "loss": 0.7742,
      "step": 25900
    },
    {
      "epoch": 1.9960003076686408,
      "grad_norm": 0.6223366856575012,
      "learning_rate": 6.785964730021037e-05,
      "loss": 0.8069,
      "step": 25950
    },
    {
      "epoch": 1.9998461656795632,
      "grad_norm": 0.848816454410553,
      "learning_rate": 6.75999272783939e-05,
      "loss": 0.8126,
      "step": 26000
    },
    {
      "epoch": 1.9998461656795632,
      "eval_loss": 0.7482460737228394,
      "eval_runtime": 17.8823,
      "eval_samples_per_second": 55.921,
      "eval_steps_per_second": 13.98,
      "step": 26000
    },
    {
      "epoch": 2.003692023690485,
      "grad_norm": 0.706822395324707,
      "learning_rate": 6.734020725657741e-05,
      "loss": 0.7,
      "step": 26050
    },
    {
      "epoch": 2.0075378817014076,
      "grad_norm": 1.503631830215454,
      "learning_rate": 6.708048723476093e-05,
      "loss": 0.7347,
      "step": 26100
    },
    {
      "epoch": 2.01138373971233,
      "grad_norm": 0.8511216044425964,
      "learning_rate": 6.682076721294445e-05,
      "loss": 0.6917,
      "step": 26150
    },
    {
      "epoch": 2.015229597723252,
      "grad_norm": 0.7063366174697876,
      "learning_rate": 6.656104719112797e-05,
      "loss": 0.7188,
      "step": 26200
    },
    {
      "epoch": 2.0190754557341744,
      "grad_norm": 0.7650218605995178,
      "learning_rate": 6.630132716931149e-05,
      "loss": 0.701,
      "step": 26250
    },
    {
      "epoch": 2.0190754557341744,
      "eval_loss": 0.7444872856140137,
      "eval_runtime": 17.8222,
      "eval_samples_per_second": 56.11,
      "eval_steps_per_second": 14.027,
      "step": 26250
    },
    {
      "epoch": 2.0229213137450963,
      "grad_norm": 0.7015202045440674,
      "learning_rate": 6.6041607147495e-05,
      "loss": 0.7442,
      "step": 26300
    },
    {
      "epoch": 2.0267671717560187,
      "grad_norm": 0.826304018497467,
      "learning_rate": 6.578188712567853e-05,
      "loss": 0.7003,
      "step": 26350
    },
    {
      "epoch": 2.030613029766941,
      "grad_norm": 0.5597676038742065,
      "learning_rate": 6.552216710386204e-05,
      "loss": 0.7108,
      "step": 26400
    },
    {
      "epoch": 2.034458887777863,
      "grad_norm": 0.738636314868927,
      "learning_rate": 6.526244708204557e-05,
      "loss": 0.7407,
      "step": 26450
    },
    {
      "epoch": 2.0383047457887855,
      "grad_norm": 0.6629013419151306,
      "learning_rate": 6.500272706022908e-05,
      "loss": 0.7227,
      "step": 26500
    },
    {
      "epoch": 2.0383047457887855,
      "eval_loss": 0.7432363033294678,
      "eval_runtime": 17.8543,
      "eval_samples_per_second": 56.009,
      "eval_steps_per_second": 14.002,
      "step": 26500
    },
    {
      "epoch": 2.042150603799708,
      "grad_norm": 0.5691978931427002,
      "learning_rate": 6.474300703841259e-05,
      "loss": 0.6752,
      "step": 26550
    },
    {
      "epoch": 2.04599646181063,
      "grad_norm": 0.5654874444007874,
      "learning_rate": 6.448328701659612e-05,
      "loss": 0.7282,
      "step": 26600
    },
    {
      "epoch": 2.0498423198215523,
      "grad_norm": 1.0112574100494385,
      "learning_rate": 6.422356699477963e-05,
      "loss": 0.7343,
      "step": 26650
    },
    {
      "epoch": 2.0536881778324743,
      "grad_norm": 0.2508319318294525,
      "learning_rate": 6.396384697296314e-05,
      "loss": 0.6796,
      "step": 26700
    },
    {
      "epoch": 2.0575340358433967,
      "grad_norm": 1.014090895652771,
      "learning_rate": 6.370412695114667e-05,
      "loss": 0.7276,
      "step": 26750
    },
    {
      "epoch": 2.0575340358433967,
      "eval_loss": 0.7409418821334839,
      "eval_runtime": 18.0076,
      "eval_samples_per_second": 55.532,
      "eval_steps_per_second": 13.883,
      "step": 26750
    },
    {
      "epoch": 2.061379893854319,
      "grad_norm": 0.6783422827720642,
      "learning_rate": 6.344440692933018e-05,
      "loss": 0.7233,
      "step": 26800
    },
    {
      "epoch": 2.065225751865241,
      "grad_norm": 0.9281295537948608,
      "learning_rate": 6.31846869075137e-05,
      "loss": 0.7463,
      "step": 26850
    },
    {
      "epoch": 2.0690716098761635,
      "grad_norm": 0.9024075269699097,
      "learning_rate": 6.293016128613355e-05,
      "loss": 0.7499,
      "step": 26900
    },
    {
      "epoch": 2.0729174678870854,
      "grad_norm": 0.9747761487960815,
      "learning_rate": 6.267044126431707e-05,
      "loss": 0.7736,
      "step": 26950
    },
    {
      "epoch": 2.076763325898008,
      "grad_norm": 0.9101582169532776,
      "learning_rate": 6.241072124250059e-05,
      "loss": 0.7343,
      "step": 27000
    },
    {
      "epoch": 2.076763325898008,
      "eval_loss": 0.7426913380622864,
      "eval_runtime": 18.0363,
      "eval_samples_per_second": 55.444,
      "eval_steps_per_second": 13.861,
      "step": 27000
    },
    {
      "epoch": 2.0806091839089302,
      "grad_norm": 0.4836456775665283,
      "learning_rate": 6.21510012206841e-05,
      "loss": 0.7725,
      "step": 27050
    },
    {
      "epoch": 2.084455041919852,
      "grad_norm": 0.9285927414894104,
      "learning_rate": 6.189128119886762e-05,
      "loss": 0.7167,
      "step": 27100
    },
    {
      "epoch": 2.0883008999307746,
      "grad_norm": 0.8826911449432373,
      "learning_rate": 6.163156117705115e-05,
      "loss": 0.7204,
      "step": 27150
    },
    {
      "epoch": 2.0921467579416966,
      "grad_norm": 1.0057801008224487,
      "learning_rate": 6.137184115523465e-05,
      "loss": 0.713,
      "step": 27200
    },
    {
      "epoch": 2.095992615952619,
      "grad_norm": 1.2974227666854858,
      "learning_rate": 6.111212113341818e-05,
      "loss": 0.6974,
      "step": 27250
    },
    {
      "epoch": 2.095992615952619,
      "eval_loss": 0.7402239441871643,
      "eval_runtime": 17.8723,
      "eval_samples_per_second": 55.953,
      "eval_steps_per_second": 13.988,
      "step": 27250
    },
    {
      "epoch": 2.0998384739635414,
      "grad_norm": 0.7047484517097473,
      "learning_rate": 6.0852401111601696e-05,
      "loss": 0.7377,
      "step": 27300
    },
    {
      "epoch": 2.1036843319744634,
      "grad_norm": 1.0780186653137207,
      "learning_rate": 6.0592681089785216e-05,
      "loss": 0.7153,
      "step": 27350
    },
    {
      "epoch": 2.1075301899853858,
      "grad_norm": 0.6844059228897095,
      "learning_rate": 6.033296106796873e-05,
      "loss": 0.7077,
      "step": 27400
    },
    {
      "epoch": 2.111376047996308,
      "grad_norm": 0.7109357118606567,
      "learning_rate": 6.007324104615225e-05,
      "loss": 0.6574,
      "step": 27450
    },
    {
      "epoch": 2.11522190600723,
      "grad_norm": 1.0174553394317627,
      "learning_rate": 5.9813521024335775e-05,
      "loss": 0.7269,
      "step": 27500
    },
    {
      "epoch": 2.11522190600723,
      "eval_loss": 0.7388279438018799,
      "eval_runtime": 17.9018,
      "eval_samples_per_second": 55.86,
      "eval_steps_per_second": 13.965,
      "step": 27500
    },
    {
      "epoch": 2.1190677640181526,
      "grad_norm": 0.8702675104141235,
      "learning_rate": 5.955380100251928e-05,
      "loss": 0.7034,
      "step": 27550
    },
    {
      "epoch": 2.1229136220290745,
      "grad_norm": 0.9476292729377747,
      "learning_rate": 5.929408098070281e-05,
      "loss": 0.7075,
      "step": 27600
    },
    {
      "epoch": 2.126759480039997,
      "grad_norm": 0.7763323783874512,
      "learning_rate": 5.903436095888633e-05,
      "loss": 0.7401,
      "step": 27650
    },
    {
      "epoch": 2.1306053380509193,
      "grad_norm": 0.9647789001464844,
      "learning_rate": 5.8774640937069847e-05,
      "loss": 0.7308,
      "step": 27700
    },
    {
      "epoch": 2.1344511960618413,
      "grad_norm": 0.7125420570373535,
      "learning_rate": 5.851492091525336e-05,
      "loss": 0.7337,
      "step": 27750
    },
    {
      "epoch": 2.1344511960618413,
      "eval_loss": 0.736056387424469,
      "eval_runtime": 18.0862,
      "eval_samples_per_second": 55.291,
      "eval_steps_per_second": 13.823,
      "step": 27750
    },
    {
      "epoch": 2.1382970540727637,
      "grad_norm": 1.1053630113601685,
      "learning_rate": 5.825520089343688e-05,
      "loss": 0.7434,
      "step": 27800
    },
    {
      "epoch": 2.1421429120836857,
      "grad_norm": 0.751206636428833,
      "learning_rate": 5.79954808716204e-05,
      "loss": 0.7208,
      "step": 27850
    },
    {
      "epoch": 2.145988770094608,
      "grad_norm": 1.0049408674240112,
      "learning_rate": 5.773576084980391e-05,
      "loss": 0.7265,
      "step": 27900
    },
    {
      "epoch": 2.1498346281055305,
      "grad_norm": 0.9738804697990417,
      "learning_rate": 5.747604082798743e-05,
      "loss": 0.7515,
      "step": 27950
    },
    {
      "epoch": 2.1536804861164525,
      "grad_norm": 0.7807592153549194,
      "learning_rate": 5.721632080617095e-05,
      "loss": 0.7309,
      "step": 28000
    },
    {
      "epoch": 2.1536804861164525,
      "eval_loss": 0.736193835735321,
      "eval_runtime": 17.9436,
      "eval_samples_per_second": 55.73,
      "eval_steps_per_second": 13.933,
      "step": 28000
    },
    {
      "epoch": 2.157526344127375,
      "grad_norm": 0.9337176084518433,
      "learning_rate": 5.6956600784354464e-05,
      "loss": 0.7844,
      "step": 28050
    },
    {
      "epoch": 2.1613722021382973,
      "grad_norm": 0.7867174744606018,
      "learning_rate": 5.669688076253798e-05,
      "loss": 0.7634,
      "step": 28100
    },
    {
      "epoch": 2.1652180601492192,
      "grad_norm": 0.6526890397071838,
      "learning_rate": 5.643716074072151e-05,
      "loss": 0.743,
      "step": 28150
    },
    {
      "epoch": 2.1690639181601417,
      "grad_norm": 1.1720079183578491,
      "learning_rate": 5.617744071890503e-05,
      "loss": 0.7165,
      "step": 28200
    },
    {
      "epoch": 2.1729097761710636,
      "grad_norm": 1.029062032699585,
      "learning_rate": 5.5917720697088535e-05,
      "loss": 0.6854,
      "step": 28250
    },
    {
      "epoch": 2.1729097761710636,
      "eval_loss": 0.7313055396080017,
      "eval_runtime": 18.0992,
      "eval_samples_per_second": 55.251,
      "eval_steps_per_second": 13.813,
      "step": 28250
    },
    {
      "epoch": 2.176755634181986,
      "grad_norm": 0.821443498134613,
      "learning_rate": 5.565800067527206e-05,
      "loss": 0.7209,
      "step": 28300
    },
    {
      "epoch": 2.1806014921929084,
      "grad_norm": 0.6121924519538879,
      "learning_rate": 5.539828065345558e-05,
      "loss": 0.7371,
      "step": 28350
    },
    {
      "epoch": 2.1844473502038304,
      "grad_norm": 0.8862821459770203,
      "learning_rate": 5.5138560631639094e-05,
      "loss": 0.684,
      "step": 28400
    },
    {
      "epoch": 2.188293208214753,
      "grad_norm": 0.8110325336456299,
      "learning_rate": 5.4878840609822614e-05,
      "loss": 0.6843,
      "step": 28450
    },
    {
      "epoch": 2.192139066225675,
      "grad_norm": 0.732751727104187,
      "learning_rate": 5.461912058800613e-05,
      "loss": 0.7229,
      "step": 28500
    },
    {
      "epoch": 2.192139066225675,
      "eval_loss": 0.7322584986686707,
      "eval_runtime": 18.025,
      "eval_samples_per_second": 55.478,
      "eval_steps_per_second": 13.87,
      "step": 28500
    },
    {
      "epoch": 2.195984924236597,
      "grad_norm": 0.8648149371147156,
      "learning_rate": 5.435940056618965e-05,
      "loss": 0.6874,
      "step": 28550
    },
    {
      "epoch": 2.1998307822475196,
      "grad_norm": 0.6547895073890686,
      "learning_rate": 5.4099680544373166e-05,
      "loss": 0.7115,
      "step": 28600
    },
    {
      "epoch": 2.2036766402584416,
      "grad_norm": 0.7705133557319641,
      "learning_rate": 5.3839960522556685e-05,
      "loss": 0.7498,
      "step": 28650
    },
    {
      "epoch": 2.207522498269364,
      "grad_norm": 1.2679988145828247,
      "learning_rate": 5.358024050074021e-05,
      "loss": 0.6956,
      "step": 28700
    },
    {
      "epoch": 2.211368356280286,
      "grad_norm": 0.4992118775844574,
      "learning_rate": 5.332052047892372e-05,
      "loss": 0.7124,
      "step": 28750
    },
    {
      "epoch": 2.211368356280286,
      "eval_loss": 0.7316113710403442,
      "eval_runtime": 17.8507,
      "eval_samples_per_second": 56.02,
      "eval_steps_per_second": 14.005,
      "step": 28750
    },
    {
      "epoch": 2.2152142142912084,
      "grad_norm": 0.8134187459945679,
      "learning_rate": 5.306080045710724e-05,
      "loss": 0.733,
      "step": 28800
    },
    {
      "epoch": 2.2190600723021308,
      "grad_norm": 0.5038111805915833,
      "learning_rate": 5.2801080435290764e-05,
      "loss": 0.7239,
      "step": 28850
    },
    {
      "epoch": 2.2229059303130527,
      "grad_norm": 1.0252164602279663,
      "learning_rate": 5.254136041347427e-05,
      "loss": 0.7102,
      "step": 28900
    },
    {
      "epoch": 2.226751788323975,
      "grad_norm": 0.823451817035675,
      "learning_rate": 5.2281640391657796e-05,
      "loss": 0.709,
      "step": 28950
    },
    {
      "epoch": 2.230597646334897,
      "grad_norm": 0.7642868161201477,
      "learning_rate": 5.2021920369841316e-05,
      "loss": 0.7326,
      "step": 29000
    },
    {
      "epoch": 2.230597646334897,
      "eval_loss": 0.7284311652183533,
      "eval_runtime": 17.8949,
      "eval_samples_per_second": 55.882,
      "eval_steps_per_second": 13.97,
      "step": 29000
    },
    {
      "epoch": 2.2344435043458195,
      "grad_norm": 0.45148393511772156,
      "learning_rate": 5.1762200348024836e-05,
      "loss": 0.7068,
      "step": 29050
    },
    {
      "epoch": 2.238289362356742,
      "grad_norm": 0.4047794044017792,
      "learning_rate": 5.150248032620835e-05,
      "loss": 0.7015,
      "step": 29100
    },
    {
      "epoch": 2.242135220367664,
      "grad_norm": 0.8585315942764282,
      "learning_rate": 5.124276030439187e-05,
      "loss": 0.7028,
      "step": 29150
    },
    {
      "epoch": 2.2459810783785863,
      "grad_norm": 0.5665230751037598,
      "learning_rate": 5.098304028257539e-05,
      "loss": 0.658,
      "step": 29200
    },
    {
      "epoch": 2.2498269363895087,
      "grad_norm": 0.7042600512504578,
      "learning_rate": 5.07233202607589e-05,
      "loss": 0.6938,
      "step": 29250
    },
    {
      "epoch": 2.2498269363895087,
      "eval_loss": 0.7279431223869324,
      "eval_runtime": 17.8253,
      "eval_samples_per_second": 56.1,
      "eval_steps_per_second": 14.025,
      "step": 29250
    },
    {
      "epoch": 2.2536727944004307,
      "grad_norm": 0.7422949075698853,
      "learning_rate": 5.046360023894242e-05,
      "loss": 0.6996,
      "step": 29300
    },
    {
      "epoch": 2.257518652411353,
      "grad_norm": 1.0139210224151611,
      "learning_rate": 5.020388021712594e-05,
      "loss": 0.7311,
      "step": 29350
    },
    {
      "epoch": 2.261364510422275,
      "grad_norm": 0.5937057137489319,
      "learning_rate": 4.994416019530946e-05,
      "loss": 0.704,
      "step": 29400
    },
    {
      "epoch": 2.2652103684331975,
      "grad_norm": 1.064329981803894,
      "learning_rate": 4.968444017349297e-05,
      "loss": 0.6792,
      "step": 29450
    },
    {
      "epoch": 2.26905622644412,
      "grad_norm": 0.9638292789459229,
      "learning_rate": 4.94247201516765e-05,
      "loss": 0.74,
      "step": 29500
    },
    {
      "epoch": 2.26905622644412,
      "eval_loss": 0.7287396192550659,
      "eval_runtime": 17.8735,
      "eval_samples_per_second": 55.949,
      "eval_steps_per_second": 13.987,
      "step": 29500
    },
    {
      "epoch": 2.272902084455042,
      "grad_norm": 0.7172055840492249,
      "learning_rate": 4.916500012986001e-05,
      "loss": 0.6984,
      "step": 29550
    },
    {
      "epoch": 2.2767479424659642,
      "grad_norm": 0.6817266345024109,
      "learning_rate": 4.890528010804353e-05,
      "loss": 0.7059,
      "step": 29600
    },
    {
      "epoch": 2.280593800476886,
      "grad_norm": 1.0110056400299072,
      "learning_rate": 4.864556008622705e-05,
      "loss": 0.7304,
      "step": 29650
    },
    {
      "epoch": 2.2844396584878086,
      "grad_norm": 0.6823923587799072,
      "learning_rate": 4.838584006441057e-05,
      "loss": 0.7462,
      "step": 29700
    },
    {
      "epoch": 2.288285516498731,
      "grad_norm": 0.7316113710403442,
      "learning_rate": 4.812612004259408e-05,
      "loss": 0.7068,
      "step": 29750
    },
    {
      "epoch": 2.288285516498731,
      "eval_loss": 0.7273894548416138,
      "eval_runtime": 17.8859,
      "eval_samples_per_second": 55.91,
      "eval_steps_per_second": 13.978,
      "step": 29750
    },
    {
      "epoch": 2.292131374509653,
      "grad_norm": 1.200492024421692,
      "learning_rate": 4.78664000207776e-05,
      "loss": 0.6887,
      "step": 29800
    },
    {
      "epoch": 2.2959772325205754,
      "grad_norm": 1.2268471717834473,
      "learning_rate": 4.760667999896112e-05,
      "loss": 0.7259,
      "step": 29850
    },
    {
      "epoch": 2.299823090531498,
      "grad_norm": 0.7251473069190979,
      "learning_rate": 4.734695997714464e-05,
      "loss": 0.6931,
      "step": 29900
    },
    {
      "epoch": 2.3036689485424198,
      "grad_norm": 0.5327921509742737,
      "learning_rate": 4.708723995532816e-05,
      "loss": 0.7447,
      "step": 29950
    },
    {
      "epoch": 2.307514806553342,
      "grad_norm": 0.7111462950706482,
      "learning_rate": 4.6827519933511674e-05,
      "loss": 0.7221,
      "step": 30000
    },
    {
      "epoch": 2.307514806553342,
      "eval_loss": 0.7264938354492188,
      "eval_runtime": 17.7847,
      "eval_samples_per_second": 56.228,
      "eval_steps_per_second": 14.057,
      "step": 30000
    },
    {
      "epoch": 2.311360664564264,
      "grad_norm": 0.861571729183197,
      "learning_rate": 4.6567799911695194e-05,
      "loss": 0.7052,
      "step": 30050
    },
    {
      "epoch": 2.3152065225751866,
      "grad_norm": 0.9279738068580627,
      "learning_rate": 4.6308079889878714e-05,
      "loss": 0.715,
      "step": 30100
    },
    {
      "epoch": 2.3190523805861085,
      "grad_norm": 0.8576169013977051,
      "learning_rate": 4.6048359868062226e-05,
      "loss": 0.7125,
      "step": 30150
    },
    {
      "epoch": 2.322898238597031,
      "grad_norm": 1.3127994537353516,
      "learning_rate": 4.578863984624575e-05,
      "loss": 0.6703,
      "step": 30200
    },
    {
      "epoch": 2.3267440966079533,
      "grad_norm": 0.5930036306381226,
      "learning_rate": 4.5528919824429266e-05,
      "loss": 0.6927,
      "step": 30250
    },
    {
      "epoch": 2.3267440966079533,
      "eval_loss": 0.7252874970436096,
      "eval_runtime": 17.8506,
      "eval_samples_per_second": 56.021,
      "eval_steps_per_second": 14.005,
      "step": 30250
    },
    {
      "epoch": 2.3305899546188753,
      "grad_norm": 0.6445633769035339,
      "learning_rate": 4.5269199802612785e-05,
      "loss": 0.6804,
      "step": 30300
    },
    {
      "epoch": 2.3344358126297977,
      "grad_norm": 0.9251648783683777,
      "learning_rate": 4.5009479780796305e-05,
      "loss": 0.7226,
      "step": 30350
    },
    {
      "epoch": 2.33828167064072,
      "grad_norm": 1.2103322744369507,
      "learning_rate": 4.474975975897982e-05,
      "loss": 0.7179,
      "step": 30400
    },
    {
      "epoch": 2.342127528651642,
      "grad_norm": 1.0718954801559448,
      "learning_rate": 4.4490039737163344e-05,
      "loss": 0.7638,
      "step": 30450
    },
    {
      "epoch": 2.3459733866625645,
      "grad_norm": 0.8156006336212158,
      "learning_rate": 4.423031971534686e-05,
      "loss": 0.6733,
      "step": 30500
    },
    {
      "epoch": 2.3459733866625645,
      "eval_loss": 0.7244414687156677,
      "eval_runtime": 17.9373,
      "eval_samples_per_second": 55.75,
      "eval_steps_per_second": 13.937,
      "step": 30500
    },
    {
      "epoch": 2.3498192446734865,
      "grad_norm": 0.9593235850334167,
      "learning_rate": 4.3970599693530377e-05,
      "loss": 0.7266,
      "step": 30550
    },
    {
      "epoch": 2.353665102684409,
      "grad_norm": 0.4570913016796112,
      "learning_rate": 4.3710879671713896e-05,
      "loss": 0.7026,
      "step": 30600
    },
    {
      "epoch": 2.3575109606953313,
      "grad_norm": 0.8020208477973938,
      "learning_rate": 4.345115964989741e-05,
      "loss": 0.7345,
      "step": 30650
    },
    {
      "epoch": 2.3613568187062532,
      "grad_norm": 0.7795267701148987,
      "learning_rate": 4.319143962808093e-05,
      "loss": 0.7351,
      "step": 30700
    },
    {
      "epoch": 2.3652026767171757,
      "grad_norm": 0.6240664720535278,
      "learning_rate": 4.293171960626445e-05,
      "loss": 0.7029,
      "step": 30750
    },
    {
      "epoch": 2.3652026767171757,
      "eval_loss": 0.7251197099685669,
      "eval_runtime": 17.8892,
      "eval_samples_per_second": 55.9,
      "eval_steps_per_second": 13.975,
      "step": 30750
    },
    {
      "epoch": 2.3690485347280976,
      "grad_norm": 0.773654043674469,
      "learning_rate": 4.267199958444797e-05,
      "loss": 0.7105,
      "step": 30800
    },
    {
      "epoch": 2.37289439273902,
      "grad_norm": 1.1365927457809448,
      "learning_rate": 4.241227956263149e-05,
      "loss": 0.7019,
      "step": 30850
    },
    {
      "epoch": 2.3767402507499424,
      "grad_norm": 0.6990851759910583,
      "learning_rate": 4.2152559540815e-05,
      "loss": 0.6914,
      "step": 30900
    },
    {
      "epoch": 2.3805861087608644,
      "grad_norm": 0.8598945140838623,
      "learning_rate": 4.189283951899852e-05,
      "loss": 0.7087,
      "step": 30950
    },
    {
      "epoch": 2.384431966771787,
      "grad_norm": 0.9121548533439636,
      "learning_rate": 4.163311949718204e-05,
      "loss": 0.7212,
      "step": 31000
    },
    {
      "epoch": 2.384431966771787,
      "eval_loss": 0.7226839661598206,
      "eval_runtime": 17.8913,
      "eval_samples_per_second": 55.893,
      "eval_steps_per_second": 13.973,
      "step": 31000
    },
    {
      "epoch": 2.3882778247827092,
      "grad_norm": 0.6950593590736389,
      "learning_rate": 4.137339947536556e-05,
      "loss": 0.7201,
      "step": 31050
    },
    {
      "epoch": 2.392123682793631,
      "grad_norm": 0.7376019358634949,
      "learning_rate": 4.111367945354908e-05,
      "loss": 0.7157,
      "step": 31100
    },
    {
      "epoch": 2.3959695408045536,
      "grad_norm": 1.286970853805542,
      "learning_rate": 4.08539594317326e-05,
      "loss": 0.7325,
      "step": 31150
    },
    {
      "epoch": 2.3998153988154756,
      "grad_norm": 0.5061975121498108,
      "learning_rate": 4.059423940991611e-05,
      "loss": 0.7206,
      "step": 31200
    },
    {
      "epoch": 2.403661256826398,
      "grad_norm": 0.7503495216369629,
      "learning_rate": 4.033451938809963e-05,
      "loss": 0.7197,
      "step": 31250
    },
    {
      "epoch": 2.403661256826398,
      "eval_loss": 0.7893036007881165,
      "eval_runtime": 17.9552,
      "eval_samples_per_second": 55.694,
      "eval_steps_per_second": 13.924,
      "step": 31250
    },
    {
      "epoch": 2.4075071148373204,
      "grad_norm": 0.6520366668701172,
      "learning_rate": 4.007479936628315e-05,
      "loss": 0.7496,
      "step": 31300
    },
    {
      "epoch": 2.4113529728482423,
      "grad_norm": 0.7475297451019287,
      "learning_rate": 3.981507934446666e-05,
      "loss": 0.7283,
      "step": 31350
    },
    {
      "epoch": 2.4151988308591648,
      "grad_norm": 1.0714281797409058,
      "learning_rate": 3.955535932265019e-05,
      "loss": 0.7553,
      "step": 31400
    },
    {
      "epoch": 2.4190446888700867,
      "grad_norm": 0.6734263300895691,
      "learning_rate": 3.92956393008337e-05,
      "loss": 0.698,
      "step": 31450
    },
    {
      "epoch": 2.422890546881009,
      "grad_norm": 0.7820257544517517,
      "learning_rate": 3.903591927901722e-05,
      "loss": 0.7346,
      "step": 31500
    },
    {
      "epoch": 2.422890546881009,
      "eval_loss": 0.7877212762832642,
      "eval_runtime": 17.9806,
      "eval_samples_per_second": 55.616,
      "eval_steps_per_second": 13.904,
      "step": 31500
    },
    {
      "epoch": 2.4267364048919315,
      "grad_norm": 0.9221381545066833,
      "learning_rate": 3.877619925720074e-05,
      "loss": 0.7077,
      "step": 31550
    },
    {
      "epoch": 2.4305822629028535,
      "grad_norm": 1.1155864000320435,
      "learning_rate": 3.8516479235384255e-05,
      "loss": 0.7358,
      "step": 31600
    },
    {
      "epoch": 2.434428120913776,
      "grad_norm": 0.949946939945221,
      "learning_rate": 3.825675921356778e-05,
      "loss": 0.7314,
      "step": 31650
    },
    {
      "epoch": 2.4382739789246983,
      "grad_norm": 0.7200281023979187,
      "learning_rate": 3.7997039191751294e-05,
      "loss": 0.6781,
      "step": 31700
    },
    {
      "epoch": 2.4421198369356203,
      "grad_norm": 1.1924189329147339,
      "learning_rate": 3.773731916993481e-05,
      "loss": 0.7116,
      "step": 31750
    },
    {
      "epoch": 2.4421198369356203,
      "eval_loss": 0.7858553528785706,
      "eval_runtime": 18.0342,
      "eval_samples_per_second": 55.45,
      "eval_steps_per_second": 13.863,
      "step": 31750
    },
    {
      "epoch": 2.4459656949465427,
      "grad_norm": 0.7993971109390259,
      "learning_rate": 3.747759914811833e-05,
      "loss": 0.7068,
      "step": 31800
    },
    {
      "epoch": 2.4498115529574647,
      "grad_norm": 0.6277671456336975,
      "learning_rate": 3.7217879126301846e-05,
      "loss": 0.7273,
      "step": 31850
    },
    {
      "epoch": 2.453657410968387,
      "grad_norm": 0.8524878621101379,
      "learning_rate": 3.6958159104485365e-05,
      "loss": 0.7537,
      "step": 31900
    },
    {
      "epoch": 2.457503268979309,
      "grad_norm": 0.9068925976753235,
      "learning_rate": 3.6698439082668885e-05,
      "loss": 0.7192,
      "step": 31950
    },
    {
      "epoch": 2.4613491269902315,
      "grad_norm": 0.866385817527771,
      "learning_rate": 3.6438719060852405e-05,
      "loss": 0.7145,
      "step": 32000
    },
    {
      "epoch": 2.4613491269902315,
      "eval_loss": 0.7852405905723572,
      "eval_runtime": 18.0021,
      "eval_samples_per_second": 55.549,
      "eval_steps_per_second": 13.887,
      "step": 32000
    },
    {
      "epoch": 2.465194985001154,
      "grad_norm": 0.8729520440101624,
      "learning_rate": 3.6178999039035924e-05,
      "loss": 0.7121,
      "step": 32050
    },
    {
      "epoch": 2.469040843012076,
      "grad_norm": 1.2588157653808594,
      "learning_rate": 3.591927901721944e-05,
      "loss": 0.7145,
      "step": 32100
    },
    {
      "epoch": 2.4728867010229982,
      "grad_norm": 1.0234293937683105,
      "learning_rate": 3.565955899540296e-05,
      "loss": 0.7173,
      "step": 32150
    },
    {
      "epoch": 2.4767325590339206,
      "grad_norm": 0.6210401058197021,
      "learning_rate": 3.5399838973586476e-05,
      "loss": 0.7404,
      "step": 32200
    },
    {
      "epoch": 2.4805784170448426,
      "grad_norm": 1.0649775266647339,
      "learning_rate": 3.5140118951769996e-05,
      "loss": 0.713,
      "step": 32250
    },
    {
      "epoch": 2.4805784170448426,
      "eval_loss": 0.7857936024665833,
      "eval_runtime": 17.9415,
      "eval_samples_per_second": 55.737,
      "eval_steps_per_second": 13.934,
      "step": 32250
    },
    {
      "epoch": 2.484424275055765,
      "grad_norm": 0.6743142604827881,
      "learning_rate": 3.488039892995351e-05,
      "loss": 0.7213,
      "step": 32300
    },
    {
      "epoch": 2.488270133066687,
      "grad_norm": 0.7584249377250671,
      "learning_rate": 3.462067890813703e-05,
      "loss": 0.7175,
      "step": 32350
    },
    {
      "epoch": 2.4921159910776094,
      "grad_norm": 1.7324374914169312,
      "learning_rate": 3.436095888632055e-05,
      "loss": 0.7255,
      "step": 32400
    },
    {
      "epoch": 2.495961849088532,
      "grad_norm": 1.0071933269500732,
      "learning_rate": 3.410123886450407e-05,
      "loss": 0.6905,
      "step": 32450
    },
    {
      "epoch": 2.4998077070994538,
      "grad_norm": 0.8606531023979187,
      "learning_rate": 3.384151884268759e-05,
      "loss": 0.7235,
      "step": 32500
    },
    {
      "epoch": 2.4998077070994538,
      "eval_loss": 0.7832362651824951,
      "eval_runtime": 17.9726,
      "eval_samples_per_second": 55.64,
      "eval_steps_per_second": 13.91,
      "step": 32500
    },
    {
      "epoch": 2.503653565110376,
      "grad_norm": 0.7658631205558777,
      "learning_rate": 3.35817988208711e-05,
      "loss": 0.7258,
      "step": 32550
    },
    {
      "epoch": 2.507499423121298,
      "grad_norm": 1.3583028316497803,
      "learning_rate": 3.3322078799054627e-05,
      "loss": 0.7414,
      "step": 32600
    },
    {
      "epoch": 2.5113452811322206,
      "grad_norm": 0.7680505514144897,
      "learning_rate": 3.306235877723814e-05,
      "loss": 0.7126,
      "step": 32650
    },
    {
      "epoch": 2.515191139143143,
      "grad_norm": 0.9117040634155273,
      "learning_rate": 3.280263875542165e-05,
      "loss": 0.7019,
      "step": 32700
    },
    {
      "epoch": 2.519036997154065,
      "grad_norm": 0.995895504951477,
      "learning_rate": 3.254291873360518e-05,
      "loss": 0.7272,
      "step": 32750
    },
    {
      "epoch": 2.519036997154065,
      "eval_loss": 0.7824276089668274,
      "eval_runtime": 17.8488,
      "eval_samples_per_second": 56.026,
      "eval_steps_per_second": 14.007,
      "step": 32750
    },
    {
      "epoch": 2.5228828551649873,
      "grad_norm": 0.6426506042480469,
      "learning_rate": 3.228319871178869e-05,
      "loss": 0.7309,
      "step": 32800
    },
    {
      "epoch": 2.5267287131759097,
      "grad_norm": 1.327599048614502,
      "learning_rate": 3.202347868997221e-05,
      "loss": 0.7354,
      "step": 32850
    },
    {
      "epoch": 2.5305745711868317,
      "grad_norm": 0.9184108972549438,
      "learning_rate": 3.176375866815573e-05,
      "loss": 0.7053,
      "step": 32900
    },
    {
      "epoch": 2.534420429197754,
      "grad_norm": 0.9938299059867859,
      "learning_rate": 3.1504038646339244e-05,
      "loss": 0.7374,
      "step": 32950
    },
    {
      "epoch": 2.5382662872086765,
      "grad_norm": 1.230980396270752,
      "learning_rate": 3.124431862452277e-05,
      "loss": 0.7261,
      "step": 33000
    },
    {
      "epoch": 2.5382662872086765,
      "eval_loss": 0.7817492485046387,
      "eval_runtime": 17.9221,
      "eval_samples_per_second": 55.797,
      "eval_steps_per_second": 13.949,
      "step": 33000
    },
    {
      "epoch": 2.5421121452195985,
      "grad_norm": 0.7762789726257324,
      "learning_rate": 3.098459860270628e-05,
      "loss": 0.7279,
      "step": 33050
    },
    {
      "epoch": 2.5459580032305205,
      "grad_norm": 0.8807786703109741,
      "learning_rate": 3.07248785808898e-05,
      "loss": 0.6957,
      "step": 33100
    },
    {
      "epoch": 2.549803861241443,
      "grad_norm": 0.8823468089103699,
      "learning_rate": 3.046515855907332e-05,
      "loss": 0.7584,
      "step": 33150
    },
    {
      "epoch": 2.5536497192523653,
      "grad_norm": 0.6461008191108704,
      "learning_rate": 3.0205438537256835e-05,
      "loss": 0.7231,
      "step": 33200
    },
    {
      "epoch": 2.5574955772632872,
      "grad_norm": 0.9959568977355957,
      "learning_rate": 2.9945718515440358e-05,
      "loss": 0.7097,
      "step": 33250
    },
    {
      "epoch": 2.5574955772632872,
      "eval_loss": 0.7804549336433411,
      "eval_runtime": 17.9765,
      "eval_samples_per_second": 55.628,
      "eval_steps_per_second": 13.907,
      "step": 33250
    },
    {
      "epoch": 2.5613414352742097,
      "grad_norm": 1.0902256965637207,
      "learning_rate": 2.9685998493623874e-05,
      "loss": 0.7259,
      "step": 33300
    },
    {
      "epoch": 2.565187293285132,
      "grad_norm": 0.8527780771255493,
      "learning_rate": 2.9426278471807394e-05,
      "loss": 0.7138,
      "step": 33350
    },
    {
      "epoch": 2.569033151296054,
      "grad_norm": 0.7497609257698059,
      "learning_rate": 2.916655844999091e-05,
      "loss": 0.6676,
      "step": 33400
    },
    {
      "epoch": 2.5728790093069764,
      "grad_norm": 1.252274751663208,
      "learning_rate": 2.8906838428174433e-05,
      "loss": 0.7545,
      "step": 33450
    },
    {
      "epoch": 2.576724867317899,
      "grad_norm": 0.8742374777793884,
      "learning_rate": 2.864711840635795e-05,
      "loss": 0.6727,
      "step": 33500
    },
    {
      "epoch": 2.576724867317899,
      "eval_loss": 0.7791668176651001,
      "eval_runtime": 17.8738,
      "eval_samples_per_second": 55.948,
      "eval_steps_per_second": 13.987,
      "step": 33500
    },
    {
      "epoch": 2.580570725328821,
      "grad_norm": 0.5088424682617188,
      "learning_rate": 2.8387398384541465e-05,
      "loss": 0.7113,
      "step": 33550
    },
    {
      "epoch": 2.5844165833397432,
      "grad_norm": 0.6116564273834229,
      "learning_rate": 2.8127678362724985e-05,
      "loss": 0.7523,
      "step": 33600
    },
    {
      "epoch": 2.588262441350665,
      "grad_norm": 0.6378856301307678,
      "learning_rate": 2.78679583409085e-05,
      "loss": 0.6924,
      "step": 33650
    },
    {
      "epoch": 2.5921082993615876,
      "grad_norm": 1.1341512203216553,
      "learning_rate": 2.760823831909202e-05,
      "loss": 0.7089,
      "step": 33700
    },
    {
      "epoch": 2.5959541573725096,
      "grad_norm": 0.8231232762336731,
      "learning_rate": 2.7348518297275537e-05,
      "loss": 0.7309,
      "step": 33750
    },
    {
      "epoch": 2.5959541573725096,
      "eval_loss": 0.7776284217834473,
      "eval_runtime": 17.9143,
      "eval_samples_per_second": 55.821,
      "eval_steps_per_second": 13.955,
      "step": 33750
    },
    {
      "epoch": 2.599800015383432,
      "grad_norm": 0.7154203653335571,
      "learning_rate": 2.7088798275459053e-05,
      "loss": 0.7273,
      "step": 33800
    },
    {
      "epoch": 2.6036458733943544,
      "grad_norm": 0.9213638305664062,
      "learning_rate": 2.6829078253642576e-05,
      "loss": 0.7628,
      "step": 33850
    },
    {
      "epoch": 2.6074917314052763,
      "grad_norm": 1.260438084602356,
      "learning_rate": 2.6569358231826092e-05,
      "loss": 0.7004,
      "step": 33900
    },
    {
      "epoch": 2.6113375894161988,
      "grad_norm": 0.9463502764701843,
      "learning_rate": 2.6309638210009612e-05,
      "loss": 0.6771,
      "step": 33950
    },
    {
      "epoch": 2.615183447427121,
      "grad_norm": 0.7610837817192078,
      "learning_rate": 2.604991818819313e-05,
      "loss": 0.7217,
      "step": 34000
    },
    {
      "epoch": 2.615183447427121,
      "eval_loss": 0.7775171995162964,
      "eval_runtime": 17.9057,
      "eval_samples_per_second": 55.848,
      "eval_steps_per_second": 13.962,
      "step": 34000
    },
    {
      "epoch": 2.619029305438043,
      "grad_norm": 0.4978080093860626,
      "learning_rate": 2.5790198166376645e-05,
      "loss": 0.7309,
      "step": 34050
    },
    {
      "epoch": 2.6228751634489655,
      "grad_norm": 0.779080331325531,
      "learning_rate": 2.5530478144560168e-05,
      "loss": 0.7105,
      "step": 34100
    },
    {
      "epoch": 2.626721021459888,
      "grad_norm": 0.5153629779815674,
      "learning_rate": 2.5275952523180012e-05,
      "loss": 0.7377,
      "step": 34150
    },
    {
      "epoch": 2.63056687947081,
      "grad_norm": 0.8356613516807556,
      "learning_rate": 2.5016232501363528e-05,
      "loss": 0.7135,
      "step": 34200
    },
    {
      "epoch": 2.6344127374817323,
      "grad_norm": 0.5202348232269287,
      "learning_rate": 2.475651247954705e-05,
      "loss": 0.7333,
      "step": 34250
    },
    {
      "epoch": 2.6344127374817323,
      "eval_loss": 0.7767261862754822,
      "eval_runtime": 17.8651,
      "eval_samples_per_second": 55.975,
      "eval_steps_per_second": 13.994,
      "step": 34250
    },
    {
      "epoch": 2.6382585954926543,
      "grad_norm": 0.9579488039016724,
      "learning_rate": 2.4496792457730567e-05,
      "loss": 0.7823,
      "step": 34300
    },
    {
      "epoch": 2.6421044535035767,
      "grad_norm": 0.7704477906227112,
      "learning_rate": 2.4237072435914084e-05,
      "loss": 0.6979,
      "step": 34350
    },
    {
      "epoch": 2.6459503115144987,
      "grad_norm": 0.8563690781593323,
      "learning_rate": 2.3977352414097603e-05,
      "loss": 0.7049,
      "step": 34400
    },
    {
      "epoch": 2.649796169525421,
      "grad_norm": 0.663038432598114,
      "learning_rate": 2.3717632392281123e-05,
      "loss": 0.751,
      "step": 34450
    },
    {
      "epoch": 2.6536420275363435,
      "grad_norm": 0.8598125576972961,
      "learning_rate": 2.345791237046464e-05,
      "loss": 0.6982,
      "step": 34500
    },
    {
      "epoch": 2.6536420275363435,
      "eval_loss": 0.7752255201339722,
      "eval_runtime": 17.9612,
      "eval_samples_per_second": 55.675,
      "eval_steps_per_second": 13.919,
      "step": 34500
    },
    {
      "epoch": 2.6574878855472654,
      "grad_norm": 1.2697360515594482,
      "learning_rate": 2.319819234864816e-05,
      "loss": 0.7289,
      "step": 34550
    },
    {
      "epoch": 2.661333743558188,
      "grad_norm": 0.6098369359970093,
      "learning_rate": 2.2938472326831675e-05,
      "loss": 0.6948,
      "step": 34600
    },
    {
      "epoch": 2.6651796015691103,
      "grad_norm": 0.48443081974983215,
      "learning_rate": 2.2678752305015195e-05,
      "loss": 0.7159,
      "step": 34650
    },
    {
      "epoch": 2.6690254595800322,
      "grad_norm": 0.7432298064231873,
      "learning_rate": 2.2419032283198714e-05,
      "loss": 0.7089,
      "step": 34700
    },
    {
      "epoch": 2.6728713175909546,
      "grad_norm": 0.7649087309837341,
      "learning_rate": 2.215931226138223e-05,
      "loss": 0.7107,
      "step": 34750
    },
    {
      "epoch": 2.6728713175909546,
      "eval_loss": 0.774026095867157,
      "eval_runtime": 18.0048,
      "eval_samples_per_second": 55.541,
      "eval_steps_per_second": 13.885,
      "step": 34750
    },
    {
      "epoch": 2.676717175601877,
      "grad_norm": 0.984624445438385,
      "learning_rate": 2.189959223956575e-05,
      "loss": 0.749,
      "step": 34800
    },
    {
      "epoch": 2.680563033612799,
      "grad_norm": 0.7625775933265686,
      "learning_rate": 2.163987221774927e-05,
      "loss": 0.6645,
      "step": 34850
    },
    {
      "epoch": 2.684408891623721,
      "grad_norm": 0.846238374710083,
      "learning_rate": 2.1380152195932786e-05,
      "loss": 0.7468,
      "step": 34900
    },
    {
      "epoch": 2.6882547496346434,
      "grad_norm": 1.1688568592071533,
      "learning_rate": 2.1120432174116302e-05,
      "loss": 0.7105,
      "step": 34950
    },
    {
      "epoch": 2.692100607645566,
      "grad_norm": 0.9417968392372131,
      "learning_rate": 2.0860712152299822e-05,
      "loss": 0.6826,
      "step": 35000
    },
    {
      "epoch": 2.692100607645566,
      "eval_loss": 0.7743579149246216,
      "eval_runtime": 18.0822,
      "eval_samples_per_second": 55.303,
      "eval_steps_per_second": 13.826,
      "step": 35000
    },
    {
      "epoch": 2.6959464656564878,
      "grad_norm": 1.1616791486740112,
      "learning_rate": 2.060099213048334e-05,
      "loss": 0.7242,
      "step": 35050
    },
    {
      "epoch": 2.69979232366741,
      "grad_norm": 0.9195474982261658,
      "learning_rate": 2.0341272108666858e-05,
      "loss": 0.7112,
      "step": 35100
    },
    {
      "epoch": 2.7036381816783326,
      "grad_norm": 1.168445110321045,
      "learning_rate": 2.0081552086850377e-05,
      "loss": 0.7331,
      "step": 35150
    },
    {
      "epoch": 2.7074840396892546,
      "grad_norm": 1.3413971662521362,
      "learning_rate": 1.9821832065033893e-05,
      "loss": 0.7627,
      "step": 35200
    },
    {
      "epoch": 2.711329897700177,
      "grad_norm": 0.9387266039848328,
      "learning_rate": 1.9562112043217413e-05,
      "loss": 0.6968,
      "step": 35250
    },
    {
      "epoch": 2.711329897700177,
      "eval_loss": 0.7742106914520264,
      "eval_runtime": 18.076,
      "eval_samples_per_second": 55.322,
      "eval_steps_per_second": 13.83,
      "step": 35250
    },
    {
      "epoch": 2.7151757557110994,
      "grad_norm": 0.8906998634338379,
      "learning_rate": 1.930239202140093e-05,
      "loss": 0.7206,
      "step": 35300
    },
    {
      "epoch": 2.7190216137220213,
      "grad_norm": 0.8276380896568298,
      "learning_rate": 1.904267199958445e-05,
      "loss": 0.7101,
      "step": 35350
    },
    {
      "epoch": 2.7228674717329437,
      "grad_norm": 0.8341213464736938,
      "learning_rate": 1.878295197776797e-05,
      "loss": 0.7631,
      "step": 35400
    },
    {
      "epoch": 2.7267133297438657,
      "grad_norm": 0.9501305222511292,
      "learning_rate": 1.8523231955951485e-05,
      "loss": 0.7138,
      "step": 35450
    },
    {
      "epoch": 2.730559187754788,
      "grad_norm": 0.9375068545341492,
      "learning_rate": 1.8263511934135e-05,
      "loss": 0.7556,
      "step": 35500
    },
    {
      "epoch": 2.730559187754788,
      "eval_loss": 0.7727531790733337,
      "eval_runtime": 18.0286,
      "eval_samples_per_second": 55.467,
      "eval_steps_per_second": 13.867,
      "step": 35500
    },
    {
      "epoch": 2.73440504576571,
      "grad_norm": 0.5093711018562317,
      "learning_rate": 1.800379191231852e-05,
      "loss": 0.6425,
      "step": 35550
    },
    {
      "epoch": 2.7382509037766325,
      "grad_norm": 1.0487879514694214,
      "learning_rate": 1.774407189050204e-05,
      "loss": 0.6846,
      "step": 35600
    },
    {
      "epoch": 2.742096761787555,
      "grad_norm": 0.6705742478370667,
      "learning_rate": 1.748435186868556e-05,
      "loss": 0.7464,
      "step": 35650
    },
    {
      "epoch": 2.745942619798477,
      "grad_norm": 0.43706628680229187,
      "learning_rate": 1.7224631846869076e-05,
      "loss": 0.7325,
      "step": 35700
    },
    {
      "epoch": 2.7497884778093993,
      "grad_norm": 1.1549192667007446,
      "learning_rate": 1.6964911825052592e-05,
      "loss": 0.7083,
      "step": 35750
    },
    {
      "epoch": 2.7497884778093993,
      "eval_loss": 0.7717772126197815,
      "eval_runtime": 18.085,
      "eval_samples_per_second": 55.295,
      "eval_steps_per_second": 13.824,
      "step": 35750
    },
    {
      "epoch": 2.7536343358203217,
      "grad_norm": 0.5367007255554199,
      "learning_rate": 1.6705191803236112e-05,
      "loss": 0.7054,
      "step": 35800
    },
    {
      "epoch": 2.7574801938312437,
      "grad_norm": 0.8213953971862793,
      "learning_rate": 1.644547178141963e-05,
      "loss": 0.6764,
      "step": 35850
    },
    {
      "epoch": 2.761326051842166,
      "grad_norm": 0.9012633562088013,
      "learning_rate": 1.6185751759603148e-05,
      "loss": 0.725,
      "step": 35900
    },
    {
      "epoch": 2.7651719098530885,
      "grad_norm": 0.656104326248169,
      "learning_rate": 1.5926031737786667e-05,
      "loss": 0.7279,
      "step": 35950
    },
    {
      "epoch": 2.7690177678640104,
      "grad_norm": 0.901136040687561,
      "learning_rate": 1.5666311715970187e-05,
      "loss": 0.7354,
      "step": 36000
    },
    {
      "epoch": 2.7690177678640104,
      "eval_loss": 0.7717016935348511,
      "eval_runtime": 17.9512,
      "eval_samples_per_second": 55.707,
      "eval_steps_per_second": 13.927,
      "step": 36000
    },
    {
      "epoch": 2.772863625874933,
      "grad_norm": 0.6917023658752441,
      "learning_rate": 1.5406591694153703e-05,
      "loss": 0.6798,
      "step": 36050
    },
    {
      "epoch": 2.776709483885855,
      "grad_norm": 0.9695160388946533,
      "learning_rate": 1.5146871672337221e-05,
      "loss": 0.7063,
      "step": 36100
    },
    {
      "epoch": 2.780555341896777,
      "grad_norm": 1.0134507417678833,
      "learning_rate": 1.4887151650520739e-05,
      "loss": 0.7325,
      "step": 36150
    },
    {
      "epoch": 2.784401199907699,
      "grad_norm": 0.8022010922431946,
      "learning_rate": 1.4627431628704259e-05,
      "loss": 0.7095,
      "step": 36200
    },
    {
      "epoch": 2.7882470579186216,
      "grad_norm": 0.8629682660102844,
      "learning_rate": 1.4367711606887777e-05,
      "loss": 0.6793,
      "step": 36250
    },
    {
      "epoch": 2.7882470579186216,
      "eval_loss": 0.7713639736175537,
      "eval_runtime": 17.9977,
      "eval_samples_per_second": 55.563,
      "eval_steps_per_second": 13.891,
      "step": 36250
    },
    {
      "epoch": 2.792092915929544,
      "grad_norm": 0.8491897583007812,
      "learning_rate": 1.4107991585071294e-05,
      "loss": 0.6899,
      "step": 36300
    },
    {
      "epoch": 2.795938773940466,
      "grad_norm": 1.0382113456726074,
      "learning_rate": 1.384827156325481e-05,
      "loss": 0.7362,
      "step": 36350
    },
    {
      "epoch": 2.7997846319513884,
      "grad_norm": 0.7207579016685486,
      "learning_rate": 1.358855154143833e-05,
      "loss": 0.6984,
      "step": 36400
    },
    {
      "epoch": 2.803630489962311,
      "grad_norm": 0.9483594298362732,
      "learning_rate": 1.3328831519621848e-05,
      "loss": 0.713,
      "step": 36450
    },
    {
      "epoch": 2.8074763479732328,
      "grad_norm": 1.0805621147155762,
      "learning_rate": 1.3069111497805366e-05,
      "loss": 0.7235,
      "step": 36500
    },
    {
      "epoch": 2.8074763479732328,
      "eval_loss": 0.7706654667854309,
      "eval_runtime": 17.9083,
      "eval_samples_per_second": 55.84,
      "eval_steps_per_second": 13.96,
      "step": 36500
    },
    {
      "epoch": 2.811322205984155,
      "grad_norm": 0.8592945337295532,
      "learning_rate": 1.2809391475988886e-05,
      "loss": 0.7173,
      "step": 36550
    },
    {
      "epoch": 2.8151680639950776,
      "grad_norm": 1.0562350749969482,
      "learning_rate": 1.2549671454172402e-05,
      "loss": 0.6661,
      "step": 36600
    },
    {
      "epoch": 2.8190139220059995,
      "grad_norm": 1.0829477310180664,
      "learning_rate": 1.2289951432355922e-05,
      "loss": 0.7105,
      "step": 36650
    },
    {
      "epoch": 2.8228597800169215,
      "grad_norm": 1.2846815586090088,
      "learning_rate": 1.203542581097577e-05,
      "loss": 0.7218,
      "step": 36700
    },
    {
      "epoch": 2.826705638027844,
      "grad_norm": 1.3996707201004028,
      "learning_rate": 1.1775705789159287e-05,
      "loss": 0.7348,
      "step": 36750
    },
    {
      "epoch": 2.826705638027844,
      "eval_loss": 0.7698732018470764,
      "eval_runtime": 17.9822,
      "eval_samples_per_second": 55.611,
      "eval_steps_per_second": 13.903,
      "step": 36750
    },
    {
      "epoch": 2.8305514960387663,
      "grad_norm": 1.040479302406311,
      "learning_rate": 1.1515985767342805e-05,
      "loss": 0.7103,
      "step": 36800
    },
    {
      "epoch": 2.8343973540496883,
      "grad_norm": 0.8566408753395081,
      "learning_rate": 1.1256265745526323e-05,
      "loss": 0.7087,
      "step": 36850
    },
    {
      "epoch": 2.8382432120606107,
      "grad_norm": 1.0727367401123047,
      "learning_rate": 1.0996545723709841e-05,
      "loss": 0.6972,
      "step": 36900
    },
    {
      "epoch": 2.842089070071533,
      "grad_norm": 0.8675785064697266,
      "learning_rate": 1.0736825701893359e-05,
      "loss": 0.7318,
      "step": 36950
    },
    {
      "epoch": 2.845934928082455,
      "grad_norm": 1.2655267715454102,
      "learning_rate": 1.0477105680076877e-05,
      "loss": 0.7333,
      "step": 37000
    },
    {
      "epoch": 2.845934928082455,
      "eval_loss": 0.7695651650428772,
      "eval_runtime": 17.7976,
      "eval_samples_per_second": 56.187,
      "eval_steps_per_second": 14.047,
      "step": 37000
    },
    {
      "epoch": 2.8497807860933775,
      "grad_norm": 1.1233916282653809,
      "learning_rate": 1.0217385658260397e-05,
      "loss": 0.7128,
      "step": 37050
    },
    {
      "epoch": 2.8536266441043,
      "grad_norm": 0.917649507522583,
      "learning_rate": 9.957665636443913e-06,
      "loss": 0.7402,
      "step": 37100
    },
    {
      "epoch": 2.857472502115222,
      "grad_norm": 0.8935102820396423,
      "learning_rate": 9.697945614627432e-06,
      "loss": 0.731,
      "step": 37150
    },
    {
      "epoch": 2.8613183601261443,
      "grad_norm": 0.6891331076622009,
      "learning_rate": 9.43822559281095e-06,
      "loss": 0.7331,
      "step": 37200
    },
    {
      "epoch": 2.8651642181370662,
      "grad_norm": 0.7505995631217957,
      "learning_rate": 9.178505570994468e-06,
      "loss": 0.6744,
      "step": 37250
    },
    {
      "epoch": 2.8651642181370662,
      "eval_loss": 0.7693511247634888,
      "eval_runtime": 17.8693,
      "eval_samples_per_second": 55.962,
      "eval_steps_per_second": 13.99,
      "step": 37250
    },
    {
      "epoch": 2.8690100761479886,
      "grad_norm": 1.2373569011688232,
      "learning_rate": 8.918785549177986e-06,
      "loss": 0.6981,
      "step": 37300
    },
    {
      "epoch": 2.8728559341589106,
      "grad_norm": 0.9159016013145447,
      "learning_rate": 8.659065527361506e-06,
      "loss": 0.7601,
      "step": 37350
    },
    {
      "epoch": 2.876701792169833,
      "grad_norm": 0.3170250952243805,
      "learning_rate": 8.399345505545022e-06,
      "loss": 0.7008,
      "step": 37400
    },
    {
      "epoch": 2.8805476501807554,
      "grad_norm": 0.7592608332633972,
      "learning_rate": 8.139625483728542e-06,
      "loss": 0.6966,
      "step": 37450
    },
    {
      "epoch": 2.8843935081916774,
      "grad_norm": 0.7826717495918274,
      "learning_rate": 7.879905461912058e-06,
      "loss": 0.7398,
      "step": 37500
    },
    {
      "epoch": 2.8843935081916774,
      "eval_loss": 0.7694031596183777,
      "eval_runtime": 17.9358,
      "eval_samples_per_second": 55.754,
      "eval_steps_per_second": 13.939,
      "step": 37500
    },
    {
      "epoch": 2.8882393662026,
      "grad_norm": 0.6858485341072083,
      "learning_rate": 7.6201854400955775e-06,
      "loss": 0.7132,
      "step": 37550
    },
    {
      "epoch": 2.892085224213522,
      "grad_norm": 0.7138088345527649,
      "learning_rate": 7.3604654182790955e-06,
      "loss": 0.7082,
      "step": 37600
    },
    {
      "epoch": 2.895931082224444,
      "grad_norm": 0.4927150309085846,
      "learning_rate": 7.100745396462613e-06,
      "loss": 0.7551,
      "step": 37650
    },
    {
      "epoch": 2.8997769402353666,
      "grad_norm": 0.879112720489502,
      "learning_rate": 6.841025374646131e-06,
      "loss": 0.7228,
      "step": 37700
    },
    {
      "epoch": 2.903622798246289,
      "grad_norm": 1.2699699401855469,
      "learning_rate": 6.58130535282965e-06,
      "loss": 0.7208,
      "step": 37750
    },
    {
      "epoch": 2.903622798246289,
      "eval_loss": 0.7685362696647644,
      "eval_runtime": 17.9674,
      "eval_samples_per_second": 55.656,
      "eval_steps_per_second": 13.914,
      "step": 37750
    },
    {
      "epoch": 2.907468656257211,
      "grad_norm": 0.7341476082801819,
      "learning_rate": 6.321585331013168e-06,
      "loss": 0.761,
      "step": 37800
    },
    {
      "epoch": 2.9113145142681334,
      "grad_norm": 0.8890082836151123,
      "learning_rate": 6.061865309196686e-06,
      "loss": 0.6837,
      "step": 37850
    },
    {
      "epoch": 2.9151603722790553,
      "grad_norm": 0.5546180009841919,
      "learning_rate": 5.802145287380204e-06,
      "loss": 0.7126,
      "step": 37900
    },
    {
      "epoch": 2.9190062302899777,
      "grad_norm": 0.7684674263000488,
      "learning_rate": 5.542425265563723e-06,
      "loss": 0.6765,
      "step": 37950
    },
    {
      "epoch": 2.9228520883008997,
      "grad_norm": 0.8968291282653809,
      "learning_rate": 5.2827052437472405e-06,
      "loss": 0.6839,
      "step": 38000
    },
    {
      "epoch": 2.9228520883008997,
      "eval_loss": 0.7687397003173828,
      "eval_runtime": 17.8165,
      "eval_samples_per_second": 56.128,
      "eval_steps_per_second": 14.032,
      "step": 38000
    },
    {
      "epoch": 2.926697946311822,
      "grad_norm": 0.8798107504844666,
      "learning_rate": 5.0229852219307584e-06,
      "loss": 0.7084,
      "step": 38050
    },
    {
      "epoch": 2.9305438043227445,
      "grad_norm": 0.3845706284046173,
      "learning_rate": 4.763265200114277e-06,
      "loss": 0.6764,
      "step": 38100
    },
    {
      "epoch": 2.9343896623336665,
      "grad_norm": 0.6847463846206665,
      "learning_rate": 4.503545178297795e-06,
      "loss": 0.7165,
      "step": 38150
    },
    {
      "epoch": 2.938235520344589,
      "grad_norm": 0.7632951736450195,
      "learning_rate": 4.243825156481313e-06,
      "loss": 0.7311,
      "step": 38200
    },
    {
      "epoch": 2.9420813783555113,
      "grad_norm": 1.3314287662506104,
      "learning_rate": 3.984105134664832e-06,
      "loss": 0.6852,
      "step": 38250
    },
    {
      "epoch": 2.9420813783555113,
      "eval_loss": 0.7683274149894714,
      "eval_runtime": 17.9706,
      "eval_samples_per_second": 55.646,
      "eval_steps_per_second": 13.912,
      "step": 38250
    },
    {
      "epoch": 2.9459272363664333,
      "grad_norm": 1.0179448127746582,
      "learning_rate": 3.7243851128483497e-06,
      "loss": 0.759,
      "step": 38300
    },
    {
      "epoch": 2.9497730943773557,
      "grad_norm": 1.4116487503051758,
      "learning_rate": 3.4646650910318677e-06,
      "loss": 0.7773,
      "step": 38350
    },
    {
      "epoch": 2.953618952388278,
      "grad_norm": 0.6251114010810852,
      "learning_rate": 3.2049450692153856e-06,
      "loss": 0.7016,
      "step": 38400
    },
    {
      "epoch": 2.9574648103992,
      "grad_norm": 0.9810579419136047,
      "learning_rate": 2.945225047398904e-06,
      "loss": 0.6909,
      "step": 38450
    },
    {
      "epoch": 2.961310668410122,
      "grad_norm": 0.7243860363960266,
      "learning_rate": 2.6855050255824223e-06,
      "loss": 0.7305,
      "step": 38500
    },
    {
      "epoch": 2.961310668410122,
      "eval_loss": 0.7680486440658569,
      "eval_runtime": 17.9145,
      "eval_samples_per_second": 55.821,
      "eval_steps_per_second": 13.955,
      "step": 38500
    },
    {
      "epoch": 2.9651565264210444,
      "grad_norm": 0.7657055854797363,
      "learning_rate": 2.4257850037659406e-06,
      "loss": 0.7114,
      "step": 38550
    },
    {
      "epoch": 2.969002384431967,
      "grad_norm": 0.7305043339729309,
      "learning_rate": 2.1660649819494585e-06,
      "loss": 0.7135,
      "step": 38600
    },
    {
      "epoch": 2.972848242442889,
      "grad_norm": 0.7981142401695251,
      "learning_rate": 1.9063449601329769e-06,
      "loss": 0.7328,
      "step": 38650
    },
    {
      "epoch": 2.976694100453811,
      "grad_norm": 0.7305875420570374,
      "learning_rate": 1.6466249383164948e-06,
      "loss": 0.7103,
      "step": 38700
    },
    {
      "epoch": 2.9805399584647336,
      "grad_norm": 1.197097659111023,
      "learning_rate": 1.386904916500013e-06,
      "loss": 0.7148,
      "step": 38750
    },
    {
      "epoch": 2.9805399584647336,
      "eval_loss": 0.7678167819976807,
      "eval_runtime": 17.8827,
      "eval_samples_per_second": 55.92,
      "eval_steps_per_second": 13.98,
      "step": 38750
    },
    {
      "epoch": 2.9843858164756556,
      "grad_norm": 0.8533993363380432,
      "learning_rate": 1.127184894683531e-06,
      "loss": 0.753,
      "step": 38800
    },
    {
      "epoch": 2.988231674486578,
      "grad_norm": 0.7372131943702698,
      "learning_rate": 8.674648728670494e-07,
      "loss": 0.7082,
      "step": 38850
    },
    {
      "epoch": 2.9920775324975004,
      "grad_norm": 1.499084234237671,
      "learning_rate": 6.077448510505675e-07,
      "loss": 0.6937,
      "step": 38900
    },
    {
      "epoch": 2.9959233905084224,
      "grad_norm": 0.5895427465438843,
      "learning_rate": 3.4802482923408566e-07,
      "loss": 0.701,
      "step": 38950
    },
    {
      "epoch": 2.999769248519345,
      "grad_norm": 0.9201724529266357,
      "learning_rate": 8.830480741760382e-08,
      "loss": 0.7665,
      "step": 39000
    },
    {
      "epoch": 2.999769248519345,
      "eval_loss": 0.767805814743042,
      "eval_runtime": 17.9509,
      "eval_samples_per_second": 55.708,
      "eval_steps_per_second": 13.927,
      "step": 39000
    }
  ],
  "logging_steps": 50,
  "max_steps": 39003,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}