{
  "best_global_step": 38000,
  "best_metric": 0.98431396484375,
  "best_model_checkpoint": "./ar-diffusion-checkpoints-fixed/checkpoint-38000",
  "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": 10.111166000366211,
      "learning_rate": 1.76e-05,
      "loss": 13.1594,
      "step": 50
    },
    {
      "epoch": 0.007691716021844474,
      "grad_norm": 5.981163501739502,
      "learning_rate": 3.76e-05,
      "loss": 7.7407,
      "step": 100
    },
    {
      "epoch": 0.01153757403276671,
      "grad_norm": 7.169628143310547,
      "learning_rate": 5.76e-05,
      "loss": 6.5841,
      "step": 150
    },
    {
      "epoch": 0.015383432043688947,
      "grad_norm": 9.481369972229004,
      "learning_rate": 7.76e-05,
      "loss": 5.8929,
      "step": 200
    },
    {
      "epoch": 0.019229290054611183,
      "grad_norm": 5.348191261291504,
      "learning_rate": 9.76e-05,
      "loss": 5.8651,
      "step": 250
    },
    {
      "epoch": 0.019229290054611183,
      "eval_loss": 5.691133975982666,
      "eval_runtime": 19.1859,
      "eval_samples_per_second": 52.122,
      "eval_steps_per_second": 13.03,
      "step": 250
    },
    {
      "epoch": 0.02307514806553342,
      "grad_norm": 9.455660820007324,
      "learning_rate": 0.0001176,
      "loss": 5.3783,
      "step": 300
    },
    {
      "epoch": 0.02692100607645566,
      "grad_norm": 11.51925277709961,
      "learning_rate": 0.00013759999999999998,
      "loss": 5.0764,
      "step": 350
    },
    {
      "epoch": 0.030766864087377895,
      "grad_norm": 5.797102928161621,
      "learning_rate": 0.0001576,
      "loss": 5.2967,
      "step": 400
    },
    {
      "epoch": 0.03461272209830013,
      "grad_norm": 3.018831253051758,
      "learning_rate": 0.0001776,
      "loss": 5.1947,
      "step": 450
    },
    {
      "epoch": 0.038458580109222366,
      "grad_norm": 5.181191921234131,
      "learning_rate": 0.0001976,
      "loss": 5.2159,
      "step": 500
    },
    {
      "epoch": 0.038458580109222366,
      "eval_loss": 5.163902759552002,
      "eval_runtime": 19.7141,
      "eval_samples_per_second": 50.725,
      "eval_steps_per_second": 12.681,
      "step": 500
    },
    {
      "epoch": 0.0423044381201446,
      "grad_norm": 6.368561744689941,
      "learning_rate": 0.0001997714463808015,
      "loss": 5.0723,
      "step": 550
    },
    {
      "epoch": 0.04615029613106684,
      "grad_norm": 4.666464328765869,
      "learning_rate": 0.00019951172635898504,
      "loss": 5.0754,
      "step": 600
    },
    {
      "epoch": 0.04999615414198908,
      "grad_norm": 9.395013809204102,
      "learning_rate": 0.00019925200633716855,
      "loss": 4.7718,
      "step": 650
    },
    {
      "epoch": 0.05384201215291132,
      "grad_norm": 4.279386520385742,
      "learning_rate": 0.00019899228631535204,
      "loss": 4.781,
      "step": 700
    },
    {
      "epoch": 0.05768787016383355,
      "grad_norm": 4.274901866912842,
      "learning_rate": 0.00019873256629353558,
      "loss": 4.9529,
      "step": 750
    },
    {
      "epoch": 0.05768787016383355,
      "eval_loss": 4.844481945037842,
      "eval_runtime": 18.9171,
      "eval_samples_per_second": 52.862,
      "eval_steps_per_second": 13.216,
      "step": 750
    },
    {
      "epoch": 0.06153372817475579,
      "grad_norm": 3.241445541381836,
      "learning_rate": 0.0001984728462717191,
      "loss": 5.0586,
      "step": 800
    },
    {
      "epoch": 0.06537958618567802,
      "grad_norm": 5.880845546722412,
      "learning_rate": 0.00019821312624990263,
      "loss": 4.952,
      "step": 850
    },
    {
      "epoch": 0.06922544419660026,
      "grad_norm": 3.81998610496521,
      "learning_rate": 0.00019795340622808614,
      "loss": 5.0012,
      "step": 900
    },
    {
      "epoch": 0.0730713022075225,
      "grad_norm": 4.52741003036499,
      "learning_rate": 0.00019769368620626966,
      "loss": 4.7775,
      "step": 950
    },
    {
      "epoch": 0.07691716021844473,
      "grad_norm": 27.68866729736328,
      "learning_rate": 0.00019743916058488948,
      "loss": 4.9852,
      "step": 1000
    },
    {
      "epoch": 0.07691716021844473,
      "eval_loss": 4.836514472961426,
      "eval_runtime": 19.7103,
      "eval_samples_per_second": 50.735,
      "eval_steps_per_second": 12.684,
      "step": 1000
    },
    {
      "epoch": 0.08076301822936698,
      "grad_norm": 5.79191255569458,
      "learning_rate": 0.00019717944056307302,
      "loss": 4.6731,
      "step": 1050
    },
    {
      "epoch": 0.0846088762402892,
      "grad_norm": 4.957877159118652,
      "learning_rate": 0.00019691972054125653,
      "loss": 4.7966,
      "step": 1100
    },
    {
      "epoch": 0.08845473425121145,
      "grad_norm": 3.2968597412109375,
      "learning_rate": 0.00019666000051944005,
      "loss": 4.752,
      "step": 1150
    },
    {
      "epoch": 0.09230059226213368,
      "grad_norm": 6.059363842010498,
      "learning_rate": 0.00019640028049762359,
      "loss": 4.7368,
      "step": 1200
    },
    {
      "epoch": 0.09614645027305592,
      "grad_norm": 5.9793171882629395,
      "learning_rate": 0.0001961405604758071,
      "loss": 4.9613,
      "step": 1250
    },
    {
      "epoch": 0.09614645027305592,
      "eval_loss": 4.76948356628418,
      "eval_runtime": 18.6703,
      "eval_samples_per_second": 53.561,
      "eval_steps_per_second": 13.39,
      "step": 1250
    },
    {
      "epoch": 0.09999230828397816,
      "grad_norm": 7.604544162750244,
      "learning_rate": 0.0001958808404539906,
      "loss": 4.8365,
      "step": 1300
    },
    {
      "epoch": 0.10383816629490039,
      "grad_norm": 6.4756083488464355,
      "learning_rate": 0.00019562112043217412,
      "loss": 4.8381,
      "step": 1350
    },
    {
      "epoch": 0.10768402430582263,
      "grad_norm": 3.605341672897339,
      "learning_rate": 0.00019536140041035764,
      "loss": 4.7004,
      "step": 1400
    },
    {
      "epoch": 0.11152988231674486,
      "grad_norm": 3.4805853366851807,
      "learning_rate": 0.00019510168038854115,
      "loss": 4.7293,
      "step": 1450
    },
    {
      "epoch": 0.1153757403276671,
      "grad_norm": 4.5123796463012695,
      "learning_rate": 0.0001948419603667247,
      "loss": 4.8315,
      "step": 1500
    },
    {
      "epoch": 0.1153757403276671,
      "eval_loss": 4.784451961517334,
      "eval_runtime": 18.5259,
      "eval_samples_per_second": 53.979,
      "eval_steps_per_second": 13.495,
      "step": 1500
    },
    {
      "epoch": 0.11922159833858934,
      "grad_norm": 3.392902135848999,
      "learning_rate": 0.0001945822403449082,
      "loss": 4.859,
      "step": 1550
    },
    {
      "epoch": 0.12306745634951158,
      "grad_norm": 4.783381938934326,
      "learning_rate": 0.00019432252032309172,
      "loss": 4.7791,
      "step": 1600
    },
    {
      "epoch": 0.12691331436043382,
      "grad_norm": 3.96708345413208,
      "learning_rate": 0.00019406280030127523,
      "loss": 4.7743,
      "step": 1650
    },
    {
      "epoch": 0.13075917237135604,
      "grad_norm": 5.381892681121826,
      "learning_rate": 0.00019380308027945874,
      "loss": 4.8915,
      "step": 1700
    },
    {
      "epoch": 0.13460503038227828,
      "grad_norm": 3.269899606704712,
      "learning_rate": 0.00019354336025764228,
      "loss": 4.8823,
      "step": 1750
    },
    {
      "epoch": 0.13460503038227828,
      "eval_loss": 4.689135551452637,
      "eval_runtime": 18.4665,
      "eval_samples_per_second": 54.152,
      "eval_steps_per_second": 13.538,
      "step": 1750
    },
    {
      "epoch": 0.13845088839320052,
      "grad_norm": 2.5163233280181885,
      "learning_rate": 0.0001932836402358258,
      "loss": 4.8513,
      "step": 1800
    },
    {
      "epoch": 0.14229674640412276,
      "grad_norm": 4.244402885437012,
      "learning_rate": 0.0001930239202140093,
      "loss": 4.6855,
      "step": 1850
    },
    {
      "epoch": 0.146142604415045,
      "grad_norm": 6.551025867462158,
      "learning_rate": 0.00019276420019219285,
      "loss": 4.8173,
      "step": 1900
    },
    {
      "epoch": 0.14998846242596722,
      "grad_norm": 2.791435956954956,
      "learning_rate": 0.00019250448017037633,
      "loss": 4.6983,
      "step": 1950
    },
    {
      "epoch": 0.15383432043688947,
      "grad_norm": 3.4012036323547363,
      "learning_rate": 0.00019224476014855984,
      "loss": 4.8178,
      "step": 2000
    },
    {
      "epoch": 0.15383432043688947,
      "eval_loss": 4.666446208953857,
      "eval_runtime": 18.3775,
      "eval_samples_per_second": 54.414,
      "eval_steps_per_second": 13.604,
      "step": 2000
    },
    {
      "epoch": 0.1576801784478117,
      "grad_norm": 6.46567964553833,
      "learning_rate": 0.00019198504012674338,
      "loss": 4.8124,
      "step": 2050
    },
    {
      "epoch": 0.16152603645873395,
      "grad_norm": 4.300732135772705,
      "learning_rate": 0.0001917253201049269,
      "loss": 4.7302,
      "step": 2100
    },
    {
      "epoch": 0.16537189446965617,
      "grad_norm": 4.140190601348877,
      "learning_rate": 0.0001914656000831104,
      "loss": 4.6498,
      "step": 2150
    },
    {
      "epoch": 0.1692177524805784,
      "grad_norm": 3.221662998199463,
      "learning_rate": 0.00019120588006129395,
      "loss": 4.6428,
      "step": 2200
    },
    {
      "epoch": 0.17306361049150065,
      "grad_norm": 2.6079111099243164,
      "learning_rate": 0.00019094616003947746,
      "loss": 4.6115,
      "step": 2250
    },
    {
      "epoch": 0.17306361049150065,
      "eval_loss": 4.661706924438477,
      "eval_runtime": 18.7413,
      "eval_samples_per_second": 53.358,
      "eval_steps_per_second": 13.34,
      "step": 2250
    },
    {
      "epoch": 0.1769094685024229,
      "grad_norm": 4.190133094787598,
      "learning_rate": 0.00019068644001766095,
      "loss": 4.7041,
      "step": 2300
    },
    {
      "epoch": 0.18075532651334514,
      "grad_norm": 5.242035388946533,
      "learning_rate": 0.0001904267199958445,
      "loss": 4.6656,
      "step": 2350
    },
    {
      "epoch": 0.18460118452426735,
      "grad_norm": 6.203541278839111,
      "learning_rate": 0.000190166999974028,
      "loss": 4.6835,
      "step": 2400
    },
    {
      "epoch": 0.1884470425351896,
      "grad_norm": 2.8591034412384033,
      "learning_rate": 0.00018990727995221151,
      "loss": 4.6751,
      "step": 2450
    },
    {
      "epoch": 0.19229290054611184,
      "grad_norm": 5.204433441162109,
      "learning_rate": 0.00018964755993039505,
      "loss": 4.7301,
      "step": 2500
    },
    {
      "epoch": 0.19229290054611184,
      "eval_loss": 4.597048282623291,
      "eval_runtime": 18.5098,
      "eval_samples_per_second": 54.025,
      "eval_steps_per_second": 13.506,
      "step": 2500
    },
    {
      "epoch": 0.19613875855703408,
      "grad_norm": 2.59525465965271,
      "learning_rate": 0.00018938783990857857,
      "loss": 4.4364,
      "step": 2550
    },
    {
      "epoch": 0.19998461656795632,
      "grad_norm": 3.844686985015869,
      "learning_rate": 0.00018912811988676208,
      "loss": 4.6437,
      "step": 2600
    },
    {
      "epoch": 0.20383047457887854,
      "grad_norm": 3.4633946418762207,
      "learning_rate": 0.0001888683998649456,
      "loss": 4.5016,
      "step": 2650
    },
    {
      "epoch": 0.20767633258980078,
      "grad_norm": 3.7852296829223633,
      "learning_rate": 0.0001886086798431291,
      "loss": 4.5845,
      "step": 2700
    },
    {
      "epoch": 0.21152219060072303,
      "grad_norm": 3.8716065883636475,
      "learning_rate": 0.00018834895982131265,
      "loss": 4.3669,
      "step": 2750
    },
    {
      "epoch": 0.21152219060072303,
      "eval_loss": 4.602295875549316,
      "eval_runtime": 18.4747,
      "eval_samples_per_second": 54.128,
      "eval_steps_per_second": 13.532,
      "step": 2750
    },
    {
      "epoch": 0.21536804861164527,
      "grad_norm": 3.9932167530059814,
      "learning_rate": 0.00018808923979949616,
      "loss": 4.4545,
      "step": 2800
    },
    {
      "epoch": 0.21921390662256748,
      "grad_norm": 4.182176113128662,
      "learning_rate": 0.00018782951977767967,
      "loss": 4.5355,
      "step": 2850
    },
    {
      "epoch": 0.22305976463348973,
      "grad_norm": 3.776895523071289,
      "learning_rate": 0.00018756979975586318,
      "loss": 4.6859,
      "step": 2900
    },
    {
      "epoch": 0.22690562264441197,
      "grad_norm": 3.9219324588775635,
      "learning_rate": 0.0001873100797340467,
      "loss": 4.5186,
      "step": 2950
    },
    {
      "epoch": 0.2307514806553342,
      "grad_norm": 3.7879323959350586,
      "learning_rate": 0.0001870503597122302,
      "loss": 4.612,
      "step": 3000
    },
    {
      "epoch": 0.2307514806553342,
      "eval_loss": 4.579595565795898,
      "eval_runtime": 18.476,
      "eval_samples_per_second": 54.124,
      "eval_steps_per_second": 13.531,
      "step": 3000
    },
    {
      "epoch": 0.23459733866625646,
      "grad_norm": 6.408226490020752,
      "learning_rate": 0.00018679063969041375,
      "loss": 4.6614,
      "step": 3050
    },
    {
      "epoch": 0.23844319667717867,
      "grad_norm": 6.529899597167969,
      "learning_rate": 0.00018653091966859726,
      "loss": 4.5576,
      "step": 3100
    },
    {
      "epoch": 0.2422890546881009,
      "grad_norm": 4.417705059051514,
      "learning_rate": 0.00018627119964678078,
      "loss": 4.5466,
      "step": 3150
    },
    {
      "epoch": 0.24613491269902316,
      "grad_norm": 3.052746295928955,
      "learning_rate": 0.0001860114796249643,
      "loss": 4.6959,
      "step": 3200
    },
    {
      "epoch": 0.2499807707099454,
      "grad_norm": 5.121955871582031,
      "learning_rate": 0.0001857517596031478,
      "loss": 4.6796,
      "step": 3250
    },
    {
      "epoch": 0.2499807707099454,
      "eval_loss": 4.576270580291748,
      "eval_runtime": 18.8896,
      "eval_samples_per_second": 52.939,
      "eval_steps_per_second": 13.235,
      "step": 3250
    },
    {
      "epoch": 0.25382662872086764,
      "grad_norm": 2.772292375564575,
      "learning_rate": 0.00018549203958133131,
      "loss": 4.534,
      "step": 3300
    },
    {
      "epoch": 0.2576724867317899,
      "grad_norm": 6.105401039123535,
      "learning_rate": 0.00018523231955951485,
      "loss": 4.4958,
      "step": 3350
    },
    {
      "epoch": 0.26151834474271207,
      "grad_norm": 2.742037057876587,
      "learning_rate": 0.00018497259953769837,
      "loss": 4.5835,
      "step": 3400
    },
    {
      "epoch": 0.2653642027536343,
      "grad_norm": 3.6125235557556152,
      "learning_rate": 0.0001847128795158819,
      "loss": 4.5917,
      "step": 3450
    },
    {
      "epoch": 0.26921006076455656,
      "grad_norm": 3.8596699237823486,
      "learning_rate": 0.00018445315949406542,
      "loss": 4.4909,
      "step": 3500
    },
    {
      "epoch": 0.26921006076455656,
      "eval_loss": 4.520543098449707,
      "eval_runtime": 18.7111,
      "eval_samples_per_second": 53.444,
      "eval_steps_per_second": 13.361,
      "step": 3500
    },
    {
      "epoch": 0.2730559187754788,
      "grad_norm": 3.639690399169922,
      "learning_rate": 0.0001841934394722489,
      "loss": 4.5127,
      "step": 3550
    },
    {
      "epoch": 0.27690177678640104,
      "grad_norm": 3.864473342895508,
      "learning_rate": 0.00018393371945043245,
      "loss": 4.599,
      "step": 3600
    },
    {
      "epoch": 0.2807476347973233,
      "grad_norm": 4.662705421447754,
      "learning_rate": 0.00018367399942861596,
      "loss": 4.5951,
      "step": 3650
    },
    {
      "epoch": 0.28459349280824553,
      "grad_norm": 3.066333532333374,
      "learning_rate": 0.00018341427940679947,
      "loss": 4.5583,
      "step": 3700
    },
    {
      "epoch": 0.2884393508191678,
      "grad_norm": 4.054274082183838,
      "learning_rate": 0.000183154559384983,
      "loss": 4.5732,
      "step": 3750
    },
    {
      "epoch": 0.2884393508191678,
      "eval_loss": 4.569947242736816,
      "eval_runtime": 18.5242,
      "eval_samples_per_second": 53.983,
      "eval_steps_per_second": 13.496,
      "step": 3750
    },
    {
      "epoch": 0.29228520883009,
      "grad_norm": 6.440130710601807,
      "learning_rate": 0.00018289483936316652,
      "loss": 4.6025,
      "step": 3800
    },
    {
      "epoch": 0.2961310668410122,
      "grad_norm": 5.849060535430908,
      "learning_rate": 0.00018263511934135004,
      "loss": 4.4932,
      "step": 3850
    },
    {
      "epoch": 0.29997692485193445,
      "grad_norm": 7.0537800788879395,
      "learning_rate": 0.00018237539931953355,
      "loss": 4.6986,
      "step": 3900
    },
    {
      "epoch": 0.3038227828628567,
      "grad_norm": 5.134806156158447,
      "learning_rate": 0.00018211567929771706,
      "loss": 4.5074,
      "step": 3950
    },
    {
      "epoch": 0.30766864087377893,
      "grad_norm": 2.872307777404785,
      "learning_rate": 0.00018185595927590058,
      "loss": 4.547,
      "step": 4000
    },
    {
      "epoch": 0.30766864087377893,
      "eval_loss": 4.5459160804748535,
      "eval_runtime": 18.4232,
      "eval_samples_per_second": 54.279,
      "eval_steps_per_second": 13.57,
      "step": 4000
    },
    {
      "epoch": 0.3115144988847012,
      "grad_norm": 5.374965190887451,
      "learning_rate": 0.00018159623925408412,
      "loss": 4.5327,
      "step": 4050
    },
    {
      "epoch": 0.3153603568956234,
      "grad_norm": 4.99652624130249,
      "learning_rate": 0.00018133651923226763,
      "loss": 4.3966,
      "step": 4100
    },
    {
      "epoch": 0.31920621490654566,
      "grad_norm": 6.257124423980713,
      "learning_rate": 0.00018107679921045114,
      "loss": 4.5792,
      "step": 4150
    },
    {
      "epoch": 0.3230520729174679,
      "grad_norm": 8.056533813476562,
      "learning_rate": 0.00018081707918863465,
      "loss": 4.6401,
      "step": 4200
    },
    {
      "epoch": 0.32689793092839015,
      "grad_norm": 4.024567127227783,
      "learning_rate": 0.00018055735916681817,
      "loss": 4.5516,
      "step": 4250
    },
    {
      "epoch": 0.32689793092839015,
      "eval_loss": 4.558110237121582,
      "eval_runtime": 18.5343,
      "eval_samples_per_second": 53.954,
      "eval_steps_per_second": 13.488,
      "step": 4250
    },
    {
      "epoch": 0.33074378893931233,
      "grad_norm": 2.4326066970825195,
      "learning_rate": 0.0001802976391450017,
      "loss": 4.5761,
      "step": 4300
    },
    {
      "epoch": 0.3345896469502346,
      "grad_norm": 2.7615299224853516,
      "learning_rate": 0.00018003791912318522,
      "loss": 4.4291,
      "step": 4350
    },
    {
      "epoch": 0.3384355049611568,
      "grad_norm": 3.9387362003326416,
      "learning_rate": 0.00017977819910136873,
      "loss": 4.5422,
      "step": 4400
    },
    {
      "epoch": 0.34228136297207906,
      "grad_norm": 23.72602653503418,
      "learning_rate": 0.00017951847907955227,
      "loss": 4.4379,
      "step": 4450
    },
    {
      "epoch": 0.3461272209830013,
      "grad_norm": 2.968930959701538,
      "learning_rate": 0.0001792639534581721,
      "loss": 4.433,
      "step": 4500
    },
    {
      "epoch": 0.3461272209830013,
      "eval_loss": 4.5359063148498535,
      "eval_runtime": 18.5422,
      "eval_samples_per_second": 53.931,
      "eval_steps_per_second": 13.483,
      "step": 4500
    },
    {
      "epoch": 0.34997307899392355,
      "grad_norm": 6.404330730438232,
      "learning_rate": 0.0001790042334363556,
      "loss": 4.5673,
      "step": 4550
    },
    {
      "epoch": 0.3538189370048458,
      "grad_norm": 10.212136268615723,
      "learning_rate": 0.00017874451341453912,
      "loss": 4.6249,
      "step": 4600
    },
    {
      "epoch": 0.35766479501576803,
      "grad_norm": 4.401816368103027,
      "learning_rate": 0.00017848479339272266,
      "loss": 4.6305,
      "step": 4650
    },
    {
      "epoch": 0.3615106530266903,
      "grad_norm": 4.710996150970459,
      "learning_rate": 0.00017822507337090617,
      "loss": 4.3731,
      "step": 4700
    },
    {
      "epoch": 0.36535651103761246,
      "grad_norm": 3.150613307952881,
      "learning_rate": 0.0001779653533490897,
      "loss": 4.4491,
      "step": 4750
    },
    {
      "epoch": 0.36535651103761246,
      "eval_loss": 4.564510345458984,
      "eval_runtime": 18.5575,
      "eval_samples_per_second": 53.886,
      "eval_steps_per_second": 13.472,
      "step": 4750
    },
    {
      "epoch": 0.3692023690485347,
      "grad_norm": 4.828207492828369,
      "learning_rate": 0.0001777056333272732,
      "loss": 4.5923,
      "step": 4800
    },
    {
      "epoch": 0.37304822705945695,
      "grad_norm": 3.780848264694214,
      "learning_rate": 0.0001774459133054567,
      "loss": 4.5544,
      "step": 4850
    },
    {
      "epoch": 0.3768940850703792,
      "grad_norm": 4.04913854598999,
      "learning_rate": 0.00017718619328364023,
      "loss": 4.5271,
      "step": 4900
    },
    {
      "epoch": 0.38073994308130144,
      "grad_norm": 4.097137451171875,
      "learning_rate": 0.00017692647326182377,
      "loss": 4.4929,
      "step": 4950
    },
    {
      "epoch": 0.3845858010922237,
      "grad_norm": 4.65788459777832,
      "learning_rate": 0.00017666675324000728,
      "loss": 4.5888,
      "step": 5000
    },
    {
      "epoch": 0.3845858010922237,
      "eval_loss": 4.465761661529541,
      "eval_runtime": 18.6518,
      "eval_samples_per_second": 53.614,
      "eval_steps_per_second": 13.404,
      "step": 5000
    },
    {
      "epoch": 0.3884316591031459,
      "grad_norm": 3.0376453399658203,
      "learning_rate": 0.00017640703321819082,
      "loss": 4.574,
      "step": 5050
    },
    {
      "epoch": 0.39227751711406816,
      "grad_norm": 2.6457693576812744,
      "learning_rate": 0.00017614731319637433,
      "loss": 4.5782,
      "step": 5100
    },
    {
      "epoch": 0.3961233751249904,
      "grad_norm": 4.438416957855225,
      "learning_rate": 0.00017588759317455782,
      "loss": 4.5586,
      "step": 5150
    },
    {
      "epoch": 0.39996923313591265,
      "grad_norm": 5.325882911682129,
      "learning_rate": 0.00017562787315274136,
      "loss": 4.5136,
      "step": 5200
    },
    {
      "epoch": 0.40381509114683484,
      "grad_norm": 1.816029667854309,
      "learning_rate": 0.00017536815313092487,
      "loss": 4.4819,
      "step": 5250
    },
    {
      "epoch": 0.40381509114683484,
      "eval_loss": 4.561609268188477,
      "eval_runtime": 18.5623,
      "eval_samples_per_second": 53.873,
      "eval_steps_per_second": 13.468,
      "step": 5250
    },
    {
      "epoch": 0.4076609491577571,
      "grad_norm": 4.011863708496094,
      "learning_rate": 0.00017510843310910838,
      "loss": 4.4284,
      "step": 5300
    },
    {
      "epoch": 0.4115068071686793,
      "grad_norm": 10.169037818908691,
      "learning_rate": 0.00017484871308729192,
      "loss": 4.5414,
      "step": 5350
    },
    {
      "epoch": 0.41535266517960157,
      "grad_norm": 3.9185502529144287,
      "learning_rate": 0.00017458899306547544,
      "loss": 4.6659,
      "step": 5400
    },
    {
      "epoch": 0.4191985231905238,
      "grad_norm": 2.873530864715576,
      "learning_rate": 0.00017432927304365895,
      "loss": 4.4482,
      "step": 5450
    },
    {
      "epoch": 0.42304438120144605,
      "grad_norm": 4.379590034484863,
      "learning_rate": 0.00017406955302184246,
      "loss": 4.5501,
      "step": 5500
    },
    {
      "epoch": 0.42304438120144605,
      "eval_loss": 4.513929843902588,
      "eval_runtime": 18.6196,
      "eval_samples_per_second": 53.707,
      "eval_steps_per_second": 13.427,
      "step": 5500
    },
    {
      "epoch": 0.4268902392123683,
      "grad_norm": 4.2113752365112305,
      "learning_rate": 0.00017380983300002597,
      "loss": 4.5454,
      "step": 5550
    },
    {
      "epoch": 0.43073609722329054,
      "grad_norm": 4.782048225402832,
      "learning_rate": 0.0001735501129782095,
      "loss": 4.4705,
      "step": 5600
    },
    {
      "epoch": 0.4345819552342128,
      "grad_norm": 2.6753036975860596,
      "learning_rate": 0.00017329039295639303,
      "loss": 4.4592,
      "step": 5650
    },
    {
      "epoch": 0.43842781324513497,
      "grad_norm": 5.012415885925293,
      "learning_rate": 0.00017303067293457654,
      "loss": 4.623,
      "step": 5700
    },
    {
      "epoch": 0.4422736712560572,
      "grad_norm": 3.0666699409484863,
      "learning_rate": 0.00017277095291276005,
      "loss": 4.4325,
      "step": 5750
    },
    {
      "epoch": 0.4422736712560572,
      "eval_loss": 4.489352703094482,
      "eval_runtime": 18.4948,
      "eval_samples_per_second": 54.069,
      "eval_steps_per_second": 13.517,
      "step": 5750
    },
    {
      "epoch": 0.44611952926697945,
      "grad_norm": 6.8570876121521,
      "learning_rate": 0.00017251123289094357,
      "loss": 4.5179,
      "step": 5800
    },
    {
      "epoch": 0.4499653872779017,
      "grad_norm": 7.190755844116211,
      "learning_rate": 0.00017225151286912708,
      "loss": 4.5877,
      "step": 5850
    },
    {
      "epoch": 0.45381124528882394,
      "grad_norm": 4.404886722564697,
      "learning_rate": 0.00017199179284731062,
      "loss": 4.4072,
      "step": 5900
    },
    {
      "epoch": 0.4576571032997462,
      "grad_norm": 3.0543084144592285,
      "learning_rate": 0.00017173207282549413,
      "loss": 4.4656,
      "step": 5950
    },
    {
      "epoch": 0.4615029613106684,
      "grad_norm": 6.7454514503479,
      "learning_rate": 0.00017147235280367764,
      "loss": 4.5688,
      "step": 6000
    },
    {
      "epoch": 0.4615029613106684,
      "eval_loss": 4.480144023895264,
      "eval_runtime": 18.5584,
      "eval_samples_per_second": 53.884,
      "eval_steps_per_second": 13.471,
      "step": 6000
    },
    {
      "epoch": 0.46534881932159067,
      "grad_norm": 5.5196661949157715,
      "learning_rate": 0.00017121263278186118,
      "loss": 4.5703,
      "step": 6050
    },
    {
      "epoch": 0.4691946773325129,
      "grad_norm": 4.253966331481934,
      "learning_rate": 0.00017095291276004467,
      "loss": 4.6012,
      "step": 6100
    },
    {
      "epoch": 0.4730405353434351,
      "grad_norm": 2.459376096725464,
      "learning_rate": 0.00017069319273822818,
      "loss": 4.5002,
      "step": 6150
    },
    {
      "epoch": 0.47688639335435734,
      "grad_norm": 4.933450698852539,
      "learning_rate": 0.00017043347271641172,
      "loss": 4.5703,
      "step": 6200
    },
    {
      "epoch": 0.4807322513652796,
      "grad_norm": 4.511186599731445,
      "learning_rate": 0.00017017375269459524,
      "loss": 4.4665,
      "step": 6250
    },
    {
      "epoch": 0.4807322513652796,
      "eval_loss": 4.485811233520508,
      "eval_runtime": 18.522,
      "eval_samples_per_second": 53.99,
      "eval_steps_per_second": 13.497,
      "step": 6250
    },
    {
      "epoch": 0.4845781093762018,
      "grad_norm": 5.634074687957764,
      "learning_rate": 0.00016991403267277875,
      "loss": 4.4616,
      "step": 6300
    },
    {
      "epoch": 0.48842396738712407,
      "grad_norm": 3.319650650024414,
      "learning_rate": 0.0001696543126509623,
      "loss": 4.4836,
      "step": 6350
    },
    {
      "epoch": 0.4922698253980463,
      "grad_norm": 3.306976079940796,
      "learning_rate": 0.00016939459262914577,
      "loss": 4.5256,
      "step": 6400
    },
    {
      "epoch": 0.49611568340896856,
      "grad_norm": 4.1797308921813965,
      "learning_rate": 0.0001691348726073293,
      "loss": 4.3822,
      "step": 6450
    },
    {
      "epoch": 0.4999615414198908,
      "grad_norm": 3.2349929809570312,
      "learning_rate": 0.00016887515258551283,
      "loss": 4.4384,
      "step": 6500
    },
    {
      "epoch": 0.4999615414198908,
      "eval_loss": 4.485826015472412,
      "eval_runtime": 18.614,
      "eval_samples_per_second": 53.723,
      "eval_steps_per_second": 13.431,
      "step": 6500
    },
    {
      "epoch": 0.503807399430813,
      "grad_norm": 3.325056791305542,
      "learning_rate": 0.00016861543256369634,
      "loss": 4.618,
      "step": 6550
    },
    {
      "epoch": 0.5076532574417353,
      "grad_norm": 4.026259899139404,
      "learning_rate": 0.0001683609069423162,
      "loss": 4.5485,
      "step": 6600
    },
    {
      "epoch": 0.5114991154526575,
      "grad_norm": 3.1270413398742676,
      "learning_rate": 0.00016810118692049973,
      "loss": 4.4426,
      "step": 6650
    },
    {
      "epoch": 0.5153449734635798,
      "grad_norm": 5.264435768127441,
      "learning_rate": 0.00016784146689868324,
      "loss": 4.3544,
      "step": 6700
    },
    {
      "epoch": 0.519190831474502,
      "grad_norm": 4.349465847015381,
      "learning_rate": 0.00016758694127730307,
      "loss": 4.5912,
      "step": 6750
    },
    {
      "epoch": 0.519190831474502,
      "eval_loss": 4.503914833068848,
      "eval_runtime": 18.5932,
      "eval_samples_per_second": 53.783,
      "eval_steps_per_second": 13.446,
      "step": 6750
    },
    {
      "epoch": 0.5230366894854241,
      "grad_norm": 2.5264992713928223,
      "learning_rate": 0.00016732722125548658,
      "loss": 4.3703,
      "step": 6800
    },
    {
      "epoch": 0.5268825474963464,
      "grad_norm": 4.934820175170898,
      "learning_rate": 0.00016706750123367012,
      "loss": 4.5652,
      "step": 6850
    },
    {
      "epoch": 0.5307284055072686,
      "grad_norm": 4.21425724029541,
      "learning_rate": 0.00016680778121185363,
      "loss": 4.4684,
      "step": 6900
    },
    {
      "epoch": 0.5345742635181909,
      "grad_norm": 5.111146926879883,
      "learning_rate": 0.00016654806119003715,
      "loss": 4.5119,
      "step": 6950
    },
    {
      "epoch": 0.5384201215291131,
      "grad_norm": 4.563775062561035,
      "learning_rate": 0.00016628834116822069,
      "loss": 4.489,
      "step": 7000
    },
    {
      "epoch": 0.5384201215291131,
      "eval_loss": 4.429446220397949,
      "eval_runtime": 18.6345,
      "eval_samples_per_second": 53.664,
      "eval_steps_per_second": 13.416,
      "step": 7000
    },
    {
      "epoch": 0.5422659795400354,
      "grad_norm": 6.055607795715332,
      "learning_rate": 0.00016602862114640417,
      "loss": 4.3805,
      "step": 7050
    },
    {
      "epoch": 0.5461118375509576,
      "grad_norm": 3.190605878829956,
      "learning_rate": 0.00016576890112458768,
      "loss": 4.561,
      "step": 7100
    },
    {
      "epoch": 0.5499576955618799,
      "grad_norm": 2.95857834815979,
      "learning_rate": 0.00016550918110277122,
      "loss": 4.5957,
      "step": 7150
    },
    {
      "epoch": 0.5538035535728021,
      "grad_norm": 5.055838108062744,
      "learning_rate": 0.00016524946108095474,
      "loss": 4.3916,
      "step": 7200
    },
    {
      "epoch": 0.5576494115837243,
      "grad_norm": 6.25083589553833,
      "learning_rate": 0.00016498974105913825,
      "loss": 4.6649,
      "step": 7250
    },
    {
      "epoch": 0.5576494115837243,
      "eval_loss": 4.42822265625,
      "eval_runtime": 18.5876,
      "eval_samples_per_second": 53.799,
      "eval_steps_per_second": 13.45,
      "step": 7250
    },
    {
      "epoch": 0.5614952695946466,
      "grad_norm": 3.6657636165618896,
      "learning_rate": 0.0001647300210373218,
      "loss": 4.5059,
      "step": 7300
    },
    {
      "epoch": 0.5653411276055688,
      "grad_norm": 3.6100645065307617,
      "learning_rate": 0.0001644703010155053,
      "loss": 4.4379,
      "step": 7350
    },
    {
      "epoch": 0.5691869856164911,
      "grad_norm": 3.535804271697998,
      "learning_rate": 0.0001642105809936888,
      "loss": 4.3881,
      "step": 7400
    },
    {
      "epoch": 0.5730328436274132,
      "grad_norm": 2.9636013507843018,
      "learning_rate": 0.00016395086097187233,
      "loss": 4.491,
      "step": 7450
    },
    {
      "epoch": 0.5768787016383355,
      "grad_norm": 2.9023678302764893,
      "learning_rate": 0.00016369114095005584,
      "loss": 4.5567,
      "step": 7500
    },
    {
      "epoch": 0.5768787016383355,
      "eval_loss": 4.459234237670898,
      "eval_runtime": 18.5868,
      "eval_samples_per_second": 53.802,
      "eval_steps_per_second": 13.45,
      "step": 7500
    },
    {
      "epoch": 0.5807245596492577,
      "grad_norm": 4.381130218505859,
      "learning_rate": 0.00016343142092823938,
      "loss": 4.521,
      "step": 7550
    },
    {
      "epoch": 0.58457041766018,
      "grad_norm": 2.533957004547119,
      "learning_rate": 0.0001631717009064229,
      "loss": 4.3233,
      "step": 7600
    },
    {
      "epoch": 0.5884162756711022,
      "grad_norm": 3.277646541595459,
      "learning_rate": 0.0001629119808846064,
      "loss": 4.5503,
      "step": 7650
    },
    {
      "epoch": 0.5922621336820244,
      "grad_norm": 3.871952772140503,
      "learning_rate": 0.00016265226086278992,
      "loss": 4.4196,
      "step": 7700
    },
    {
      "epoch": 0.5961079916929467,
      "grad_norm": 3.191589832305908,
      "learning_rate": 0.00016239254084097343,
      "loss": 4.479,
      "step": 7750
    },
    {
      "epoch": 0.5961079916929467,
      "eval_loss": 4.44810152053833,
      "eval_runtime": 18.425,
      "eval_samples_per_second": 54.274,
      "eval_steps_per_second": 13.569,
      "step": 7750
    },
    {
      "epoch": 0.5999538497038689,
      "grad_norm": 3.7022933959960938,
      "learning_rate": 0.00016213282081915695,
      "loss": 4.5402,
      "step": 7800
    },
    {
      "epoch": 0.6037997077147912,
      "grad_norm": 2.883859395980835,
      "learning_rate": 0.00016187310079734049,
      "loss": 4.5457,
      "step": 7850
    },
    {
      "epoch": 0.6076455657257134,
      "grad_norm": 3.5229415893554688,
      "learning_rate": 0.000161613380775524,
      "loss": 4.5185,
      "step": 7900
    },
    {
      "epoch": 0.6114914237366357,
      "grad_norm": 4.529599666595459,
      "learning_rate": 0.0001613536607537075,
      "loss": 4.4489,
      "step": 7950
    },
    {
      "epoch": 0.6153372817475579,
      "grad_norm": 5.1023850440979,
      "learning_rate": 0.00016109394073189102,
      "loss": 4.6043,
      "step": 8000
    },
    {
      "epoch": 0.6153372817475579,
      "eval_loss": 4.478011608123779,
      "eval_runtime": 18.576,
      "eval_samples_per_second": 53.833,
      "eval_steps_per_second": 13.458,
      "step": 8000
    },
    {
      "epoch": 0.6191831397584802,
      "grad_norm": 4.5992255210876465,
      "learning_rate": 0.00016083422071007454,
      "loss": 4.3146,
      "step": 8050
    },
    {
      "epoch": 0.6230289977694023,
      "grad_norm": 5.412031650543213,
      "learning_rate": 0.00016057450068825805,
      "loss": 4.5841,
      "step": 8100
    },
    {
      "epoch": 0.6268748557803245,
      "grad_norm": 4.4531779289245605,
      "learning_rate": 0.0001603147806664416,
      "loss": 4.5084,
      "step": 8150
    },
    {
      "epoch": 0.6307207137912468,
      "grad_norm": 3.913174867630005,
      "learning_rate": 0.0001600550606446251,
      "loss": 4.4183,
      "step": 8200
    },
    {
      "epoch": 0.634566571802169,
      "grad_norm": 3.3952407836914062,
      "learning_rate": 0.00015979534062280862,
      "loss": 4.3496,
      "step": 8250
    },
    {
      "epoch": 0.634566571802169,
      "eval_loss": 4.46174430847168,
      "eval_runtime": 18.5381,
      "eval_samples_per_second": 53.943,
      "eval_steps_per_second": 13.486,
      "step": 8250
    },
    {
      "epoch": 0.6384124298130913,
      "grad_norm": 4.156221389770508,
      "learning_rate": 0.00015953562060099213,
      "loss": 4.361,
      "step": 8300
    },
    {
      "epoch": 0.6422582878240135,
      "grad_norm": 3.2021920680999756,
      "learning_rate": 0.00015927590057917564,
      "loss": 4.473,
      "step": 8350
    },
    {
      "epoch": 0.6461041458349358,
      "grad_norm": 5.048036575317383,
      "learning_rate": 0.00015901618055735918,
      "loss": 4.4142,
      "step": 8400
    },
    {
      "epoch": 0.649950003845858,
      "grad_norm": 3.91768217086792,
      "learning_rate": 0.0001587564605355427,
      "loss": 4.4672,
      "step": 8450
    },
    {
      "epoch": 0.6537958618567803,
      "grad_norm": 9.229452133178711,
      "learning_rate": 0.0001584967405137262,
      "loss": 4.7837,
      "step": 8500
    },
    {
      "epoch": 0.6537958618567803,
      "eval_loss": 4.409055709838867,
      "eval_runtime": 18.5943,
      "eval_samples_per_second": 53.78,
      "eval_steps_per_second": 13.445,
      "step": 8500
    },
    {
      "epoch": 0.6576417198677025,
      "grad_norm": 2.7313661575317383,
      "learning_rate": 0.00015823702049190975,
      "loss": 4.4721,
      "step": 8550
    },
    {
      "epoch": 0.6614875778786247,
      "grad_norm": 4.160475730895996,
      "learning_rate": 0.00015797730047009326,
      "loss": 4.501,
      "step": 8600
    },
    {
      "epoch": 0.665333435889547,
      "grad_norm": 11.54045581817627,
      "learning_rate": 0.00015771758044827675,
      "loss": 4.4433,
      "step": 8650
    },
    {
      "epoch": 0.6691792939004692,
      "grad_norm": 4.087617874145508,
      "learning_rate": 0.00015745786042646029,
      "loss": 4.4981,
      "step": 8700
    },
    {
      "epoch": 0.6730251519113915,
      "grad_norm": 4.155121803283691,
      "learning_rate": 0.0001571981404046438,
      "loss": 4.3874,
      "step": 8750
    },
    {
      "epoch": 0.6730251519113915,
      "eval_loss": 4.418811321258545,
      "eval_runtime": 18.6306,
      "eval_samples_per_second": 53.675,
      "eval_steps_per_second": 13.419,
      "step": 8750
    },
    {
      "epoch": 0.6768710099223136,
      "grad_norm": 4.071916580200195,
      "learning_rate": 0.0001569384203828273,
      "loss": 4.5531,
      "step": 8800
    },
    {
      "epoch": 0.6807168679332359,
      "grad_norm": 3.395460605621338,
      "learning_rate": 0.00015667870036101085,
      "loss": 4.4609,
      "step": 8850
    },
    {
      "epoch": 0.6845627259441581,
      "grad_norm": 3.4933230876922607,
      "learning_rate": 0.00015641898033919436,
      "loss": 4.4536,
      "step": 8900
    },
    {
      "epoch": 0.6884085839550804,
      "grad_norm": 6.921072483062744,
      "learning_rate": 0.00015615926031737788,
      "loss": 4.3478,
      "step": 8950
    },
    {
      "epoch": 0.6922544419660026,
      "grad_norm": 3.920626401901245,
      "learning_rate": 0.0001558995402955614,
      "loss": 4.3761,
      "step": 9000
    },
    {
      "epoch": 0.6922544419660026,
      "eval_loss": 4.415992259979248,
      "eval_runtime": 18.5147,
      "eval_samples_per_second": 54.011,
      "eval_steps_per_second": 13.503,
      "step": 9000
    },
    {
      "epoch": 0.6961002999769248,
      "grad_norm": 7.213745594024658,
      "learning_rate": 0.0001556398202737449,
      "loss": 4.323,
      "step": 9050
    },
    {
      "epoch": 0.6999461579878471,
      "grad_norm": 3.2426984310150146,
      "learning_rate": 0.00015538010025192842,
      "loss": 4.3922,
      "step": 9100
    },
    {
      "epoch": 0.7037920159987693,
      "grad_norm": 3.256950855255127,
      "learning_rate": 0.00015512038023011196,
      "loss": 4.2602,
      "step": 9150
    },
    {
      "epoch": 0.7076378740096916,
      "grad_norm": 6.132264614105225,
      "learning_rate": 0.00015486066020829547,
      "loss": 4.3734,
      "step": 9200
    },
    {
      "epoch": 0.7114837320206138,
      "grad_norm": 3.921595573425293,
      "learning_rate": 0.00015460094018647898,
      "loss": 4.4776,
      "step": 9250
    },
    {
      "epoch": 0.7114837320206138,
      "eval_loss": 4.3921356201171875,
      "eval_runtime": 18.5511,
      "eval_samples_per_second": 53.905,
      "eval_steps_per_second": 13.476,
      "step": 9250
    },
    {
      "epoch": 0.7153295900315361,
      "grad_norm": 5.416064739227295,
      "learning_rate": 0.0001543412201646625,
      "loss": 4.5659,
      "step": 9300
    },
    {
      "epoch": 0.7191754480424583,
      "grad_norm": 4.542217254638672,
      "learning_rate": 0.000154081500142846,
      "loss": 4.2557,
      "step": 9350
    },
    {
      "epoch": 0.7230213060533806,
      "grad_norm": 3.7075681686401367,
      "learning_rate": 0.00015382178012102955,
      "loss": 4.453,
      "step": 9400
    },
    {
      "epoch": 0.7268671640643027,
      "grad_norm": 4.457496166229248,
      "learning_rate": 0.00015356206009921306,
      "loss": 4.3861,
      "step": 9450
    },
    {
      "epoch": 0.7307130220752249,
      "grad_norm": 1.7784981727600098,
      "learning_rate": 0.00015330234007739657,
      "loss": 4.3258,
      "step": 9500
    },
    {
      "epoch": 0.7307130220752249,
      "eval_loss": 4.357193470001221,
      "eval_runtime": 18.6372,
      "eval_samples_per_second": 53.656,
      "eval_steps_per_second": 13.414,
      "step": 9500
    },
    {
      "epoch": 0.7345588800861472,
      "grad_norm": 3.8532371520996094,
      "learning_rate": 0.0001530426200555801,
      "loss": 4.3473,
      "step": 9550
    },
    {
      "epoch": 0.7384047380970694,
      "grad_norm": 4.654659271240234,
      "learning_rate": 0.0001527829000337636,
      "loss": 4.457,
      "step": 9600
    },
    {
      "epoch": 0.7422505961079917,
      "grad_norm": 2.420182228088379,
      "learning_rate": 0.0001525231800119471,
      "loss": 4.4521,
      "step": 9650
    },
    {
      "epoch": 0.7460964541189139,
      "grad_norm": 4.189414978027344,
      "learning_rate": 0.00015226345999013065,
      "loss": 4.2569,
      "step": 9700
    },
    {
      "epoch": 0.7499423121298362,
      "grad_norm": 2.824084997177124,
      "learning_rate": 0.00015200373996831416,
      "loss": 4.3409,
      "step": 9750
    },
    {
      "epoch": 0.7499423121298362,
      "eval_loss": 4.378731727600098,
      "eval_runtime": 18.4857,
      "eval_samples_per_second": 54.096,
      "eval_steps_per_second": 13.524,
      "step": 9750
    },
    {
      "epoch": 0.7537881701407584,
      "grad_norm": 6.379781723022461,
      "learning_rate": 0.00015174401994649768,
      "loss": 4.3041,
      "step": 9800
    },
    {
      "epoch": 0.7576340281516807,
      "grad_norm": 1.7334113121032715,
      "learning_rate": 0.00015148429992468122,
      "loss": 4.3096,
      "step": 9850
    },
    {
      "epoch": 0.7614798861626029,
      "grad_norm": 4.287415027618408,
      "learning_rate": 0.00015122457990286473,
      "loss": 4.4411,
      "step": 9900
    },
    {
      "epoch": 0.7653257441735252,
      "grad_norm": 3.3184821605682373,
      "learning_rate": 0.00015096485988104821,
      "loss": 4.3992,
      "step": 9950
    },
    {
      "epoch": 0.7691716021844474,
      "grad_norm": 4.698968887329102,
      "learning_rate": 0.00015070513985923175,
      "loss": 4.4726,
      "step": 10000
    },
    {
      "epoch": 0.7691716021844474,
      "eval_loss": 4.408615589141846,
      "eval_runtime": 18.5408,
      "eval_samples_per_second": 53.935,
      "eval_steps_per_second": 13.484,
      "step": 10000
    },
    {
      "epoch": 0.7730174601953695,
      "grad_norm": 3.882775068283081,
      "learning_rate": 0.00015044541983741527,
      "loss": 4.5207,
      "step": 10050
    },
    {
      "epoch": 0.7768633182062918,
      "grad_norm": 5.814795017242432,
      "learning_rate": 0.0001501856998155988,
      "loss": 4.2462,
      "step": 10100
    },
    {
      "epoch": 0.780709176217214,
      "grad_norm": 4.733581066131592,
      "learning_rate": 0.00014992597979378232,
      "loss": 4.5563,
      "step": 10150
    },
    {
      "epoch": 0.7845550342281363,
      "grad_norm": 4.805403232574463,
      "learning_rate": 0.00014966625977196583,
      "loss": 4.4353,
      "step": 10200
    },
    {
      "epoch": 0.7884008922390585,
      "grad_norm": 5.814332008361816,
      "learning_rate": 0.00014940653975014935,
      "loss": 4.4004,
      "step": 10250
    },
    {
      "epoch": 0.7884008922390585,
      "eval_loss": 4.41144323348999,
      "eval_runtime": 18.5402,
      "eval_samples_per_second": 53.937,
      "eval_steps_per_second": 13.484,
      "step": 10250
    },
    {
      "epoch": 0.7922467502499808,
      "grad_norm": 5.321393013000488,
      "learning_rate": 0.00014914681972833286,
      "loss": 4.4383,
      "step": 10300
    },
    {
      "epoch": 0.796092608260903,
      "grad_norm": 3.681452751159668,
      "learning_rate": 0.00014888709970651637,
      "loss": 4.5094,
      "step": 10350
    },
    {
      "epoch": 0.7999384662718253,
      "grad_norm": 4.766401767730713,
      "learning_rate": 0.0001486273796846999,
      "loss": 4.3104,
      "step": 10400
    },
    {
      "epoch": 0.8037843242827475,
      "grad_norm": 4.676774024963379,
      "learning_rate": 0.00014836765966288342,
      "loss": 4.516,
      "step": 10450
    },
    {
      "epoch": 0.8076301822936697,
      "grad_norm": 3.623643159866333,
      "learning_rate": 0.00014810793964106694,
      "loss": 4.5146,
      "step": 10500
    },
    {
      "epoch": 0.8076301822936697,
      "eval_loss": 4.398375034332275,
      "eval_runtime": 18.4969,
      "eval_samples_per_second": 54.063,
      "eval_steps_per_second": 13.516,
      "step": 10500
    },
    {
      "epoch": 0.811476040304592,
      "grad_norm": 4.379317760467529,
      "learning_rate": 0.00014784821961925045,
      "loss": 4.6715,
      "step": 10550
    },
    {
      "epoch": 0.8153218983155142,
      "grad_norm": 3.034796714782715,
      "learning_rate": 0.00014758849959743396,
      "loss": 4.4511,
      "step": 10600
    },
    {
      "epoch": 0.8191677563264365,
      "grad_norm": 3.8016927242279053,
      "learning_rate": 0.00014732877957561748,
      "loss": 4.2966,
      "step": 10650
    },
    {
      "epoch": 0.8230136143373586,
      "grad_norm": 4.330080509185791,
      "learning_rate": 0.00014706905955380102,
      "loss": 4.6496,
      "step": 10700
    },
    {
      "epoch": 0.8268594723482809,
      "grad_norm": 8.032389640808105,
      "learning_rate": 0.00014680933953198453,
      "loss": 4.2758,
      "step": 10750
    },
    {
      "epoch": 0.8268594723482809,
      "eval_loss": 4.388455867767334,
      "eval_runtime": 18.481,
      "eval_samples_per_second": 54.11,
      "eval_steps_per_second": 13.527,
      "step": 10750
    },
    {
      "epoch": 0.8307053303592031,
      "grad_norm": 3.1724319458007812,
      "learning_rate": 0.00014654961951016807,
      "loss": 4.405,
      "step": 10800
    },
    {
      "epoch": 0.8345511883701254,
      "grad_norm": 3.529196262359619,
      "learning_rate": 0.00014628989948835155,
      "loss": 4.4433,
      "step": 10850
    },
    {
      "epoch": 0.8383970463810476,
      "grad_norm": 4.758362293243408,
      "learning_rate": 0.00014603017946653507,
      "loss": 4.5747,
      "step": 10900
    },
    {
      "epoch": 0.8422429043919698,
      "grad_norm": 3.524068832397461,
      "learning_rate": 0.0001457704594447186,
      "loss": 4.3469,
      "step": 10950
    },
    {
      "epoch": 0.8460887624028921,
      "grad_norm": 4.452401161193848,
      "learning_rate": 0.00014551073942290212,
      "loss": 4.3767,
      "step": 11000
    },
    {
      "epoch": 0.8460887624028921,
      "eval_loss": 4.352676868438721,
      "eval_runtime": 18.5817,
      "eval_samples_per_second": 53.816,
      "eval_steps_per_second": 13.454,
      "step": 11000
    },
    {
      "epoch": 0.8499346204138143,
      "grad_norm": 6.128251075744629,
      "learning_rate": 0.00014525101940108563,
      "loss": 4.3993,
      "step": 11050
    },
    {
      "epoch": 0.8537804784247366,
      "grad_norm": 3.9961323738098145,
      "learning_rate": 0.00014499129937926917,
      "loss": 4.406,
      "step": 11100
    },
    {
      "epoch": 0.8576263364356588,
      "grad_norm": 3.889711856842041,
      "learning_rate": 0.00014473157935745269,
      "loss": 4.2991,
      "step": 11150
    },
    {
      "epoch": 0.8614721944465811,
      "grad_norm": 3.5852463245391846,
      "learning_rate": 0.00014447185933563617,
      "loss": 4.2967,
      "step": 11200
    },
    {
      "epoch": 0.8653180524575033,
      "grad_norm": 3.343247652053833,
      "learning_rate": 0.0001442121393138197,
      "loss": 4.314,
      "step": 11250
    },
    {
      "epoch": 0.8653180524575033,
      "eval_loss": 4.387504577636719,
      "eval_runtime": 18.554,
      "eval_samples_per_second": 53.897,
      "eval_steps_per_second": 13.474,
      "step": 11250
    },
    {
      "epoch": 0.8691639104684256,
      "grad_norm": 3.5600407123565674,
      "learning_rate": 0.00014395241929200322,
      "loss": 4.2958,
      "step": 11300
    },
    {
      "epoch": 0.8730097684793477,
      "grad_norm": 4.299932956695557,
      "learning_rate": 0.00014369269927018674,
      "loss": 4.2966,
      "step": 11350
    },
    {
      "epoch": 0.8768556264902699,
      "grad_norm": 2.5763466358184814,
      "learning_rate": 0.00014343297924837028,
      "loss": 4.3739,
      "step": 11400
    },
    {
      "epoch": 0.8807014845011922,
      "grad_norm": 3.670653820037842,
      "learning_rate": 0.0001431732592265538,
      "loss": 4.3932,
      "step": 11450
    },
    {
      "epoch": 0.8845473425121144,
      "grad_norm": 5.472078800201416,
      "learning_rate": 0.0001429135392047373,
      "loss": 4.3183,
      "step": 11500
    },
    {
      "epoch": 0.8845473425121144,
      "eval_loss": 4.377117156982422,
      "eval_runtime": 18.5574,
      "eval_samples_per_second": 53.887,
      "eval_steps_per_second": 13.472,
      "step": 11500
    },
    {
      "epoch": 0.8883932005230367,
      "grad_norm": 4.711415767669678,
      "learning_rate": 0.00014265381918292082,
      "loss": 4.5701,
      "step": 11550
    },
    {
      "epoch": 0.8922390585339589,
      "grad_norm": 3.1737523078918457,
      "learning_rate": 0.00014239409916110433,
      "loss": 4.4522,
      "step": 11600
    },
    {
      "epoch": 0.8960849165448812,
      "grad_norm": 4.876018047332764,
      "learning_rate": 0.00014213437913928787,
      "loss": 4.3937,
      "step": 11650
    },
    {
      "epoch": 0.8999307745558034,
      "grad_norm": 7.117967128753662,
      "learning_rate": 0.00014187465911747138,
      "loss": 4.3585,
      "step": 11700
    },
    {
      "epoch": 0.9037766325667257,
      "grad_norm": 2.587160587310791,
      "learning_rate": 0.0001416149390956549,
      "loss": 1.5054,
      "step": 11750
    },
    {
      "epoch": 0.9037766325667257,
      "eval_loss": 1.5468424558639526,
      "eval_runtime": 17.9563,
      "eval_samples_per_second": 55.691,
      "eval_steps_per_second": 13.923,
      "step": 11750
    },
    {
      "epoch": 0.9076224905776479,
      "grad_norm": 1.4935526847839355,
      "learning_rate": 0.0001413552190738384,
      "loss": 1.5719,
      "step": 11800
    },
    {
      "epoch": 0.9114683485885701,
      "grad_norm": 1.4879201650619507,
      "learning_rate": 0.00014109549905202192,
      "loss": 1.4862,
      "step": 11850
    },
    {
      "epoch": 0.9153142065994924,
      "grad_norm": 1.7936193943023682,
      "learning_rate": 0.00014083577903020543,
      "loss": 1.488,
      "step": 11900
    },
    {
      "epoch": 0.9191600646104146,
      "grad_norm": 2.14953875541687,
      "learning_rate": 0.00014057605900838897,
      "loss": 1.5383,
      "step": 11950
    },
    {
      "epoch": 0.9230059226213368,
      "grad_norm": 1.5575013160705566,
      "learning_rate": 0.00014031633898657249,
      "loss": 1.4814,
      "step": 12000
    },
    {
      "epoch": 0.9230059226213368,
      "eval_loss": 1.5273067951202393,
      "eval_runtime": 17.7259,
      "eval_samples_per_second": 56.414,
      "eval_steps_per_second": 14.104,
      "step": 12000
    },
    {
      "epoch": 0.926851780632259,
      "grad_norm": 1.9142653942108154,
      "learning_rate": 0.000140056618964756,
      "loss": 1.5436,
      "step": 12050
    },
    {
      "epoch": 0.9306976386431813,
      "grad_norm": 1.9730989933013916,
      "learning_rate": 0.00013979689894293954,
      "loss": 1.5175,
      "step": 12100
    },
    {
      "epoch": 0.9345434966541035,
      "grad_norm": 0.9956797361373901,
      "learning_rate": 0.00013953717892112302,
      "loss": 1.4491,
      "step": 12150
    },
    {
      "epoch": 0.9383893546650258,
      "grad_norm": 1.0358608961105347,
      "learning_rate": 0.00013927745889930654,
      "loss": 1.4488,
      "step": 12200
    },
    {
      "epoch": 0.942235212675948,
      "grad_norm": 1.5184404850006104,
      "learning_rate": 0.00013901773887749008,
      "loss": 1.5037,
      "step": 12250
    },
    {
      "epoch": 0.942235212675948,
      "eval_loss": 1.5220181941986084,
      "eval_runtime": 17.7647,
      "eval_samples_per_second": 56.291,
      "eval_steps_per_second": 14.073,
      "step": 12250
    },
    {
      "epoch": 0.9460810706868702,
      "grad_norm": 1.3881593942642212,
      "learning_rate": 0.0001387580188556736,
      "loss": 1.4952,
      "step": 12300
    },
    {
      "epoch": 0.9499269286977925,
      "grad_norm": 2.319173574447632,
      "learning_rate": 0.0001384982988338571,
      "loss": 1.4667,
      "step": 12350
    },
    {
      "epoch": 0.9537727867087147,
      "grad_norm": 1.9042879343032837,
      "learning_rate": 0.00013823857881204064,
      "loss": 1.6038,
      "step": 12400
    },
    {
      "epoch": 0.957618644719637,
      "grad_norm": 1.9162698984146118,
      "learning_rate": 0.00013797885879022415,
      "loss": 1.4904,
      "step": 12450
    },
    {
      "epoch": 0.9614645027305592,
      "grad_norm": 2.0601863861083984,
      "learning_rate": 0.00013771913876840767,
      "loss": 1.4758,
      "step": 12500
    },
    {
      "epoch": 0.9614645027305592,
      "eval_loss": 1.5356587171554565,
      "eval_runtime": 17.7404,
      "eval_samples_per_second": 56.369,
      "eval_steps_per_second": 14.092,
      "step": 12500
    },
    {
      "epoch": 0.9653103607414815,
      "grad_norm": 1.888836145401001,
      "learning_rate": 0.00013745941874659118,
      "loss": 1.5681,
      "step": 12550
    },
    {
      "epoch": 0.9691562187524037,
      "grad_norm": 1.4329860210418701,
      "learning_rate": 0.0001371996987247747,
      "loss": 1.5018,
      "step": 12600
    },
    {
      "epoch": 0.973002076763326,
      "grad_norm": 1.969533920288086,
      "learning_rate": 0.00013693997870295823,
      "loss": 1.5494,
      "step": 12650
    },
    {
      "epoch": 0.9768479347742481,
      "grad_norm": 2.1219890117645264,
      "learning_rate": 0.00013668025868114175,
      "loss": 1.5386,
      "step": 12700
    },
    {
      "epoch": 0.9806937927851703,
      "grad_norm": 1.6632941961288452,
      "learning_rate": 0.00013642053865932526,
      "loss": 1.5253,
      "step": 12750
    },
    {
      "epoch": 0.9806937927851703,
      "eval_loss": 1.5147372484207153,
      "eval_runtime": 17.8072,
      "eval_samples_per_second": 56.157,
      "eval_steps_per_second": 14.039,
      "step": 12750
    },
    {
      "epoch": 0.9845396507960926,
      "grad_norm": 1.292913794517517,
      "learning_rate": 0.00013616081863750877,
      "loss": 1.4413,
      "step": 12800
    },
    {
      "epoch": 0.9883855088070148,
      "grad_norm": 1.1377824544906616,
      "learning_rate": 0.00013590109861569228,
      "loss": 1.5705,
      "step": 12850
    },
    {
      "epoch": 0.9922313668179371,
      "grad_norm": 1.656996726989746,
      "learning_rate": 0.0001356413785938758,
      "loss": 1.5311,
      "step": 12900
    },
    {
      "epoch": 0.9960772248288593,
      "grad_norm": 1.6639357805252075,
      "learning_rate": 0.00013538165857205934,
      "loss": 1.5676,
      "step": 12950
    },
    {
      "epoch": 0.9999230828397816,
      "grad_norm": 1.0893466472625732,
      "learning_rate": 0.00013512193855024285,
      "loss": 1.5215,
      "step": 13000
    },
    {
      "epoch": 0.9999230828397816,
      "eval_loss": 1.5119102001190186,
      "eval_runtime": 17.7291,
      "eval_samples_per_second": 56.404,
      "eval_steps_per_second": 14.101,
      "step": 13000
    },
    {
      "epoch": 1.0037689408507038,
      "grad_norm": 1.4789248704910278,
      "learning_rate": 0.00013486221852842636,
      "loss": 1.4551,
      "step": 13050
    },
    {
      "epoch": 1.007614798861626,
      "grad_norm": 1.3270663022994995,
      "learning_rate": 0.00013460249850660988,
      "loss": 1.4776,
      "step": 13100
    },
    {
      "epoch": 1.0114606568725482,
      "grad_norm": 1.3546854257583618,
      "learning_rate": 0.0001343427784847934,
      "loss": 1.5807,
      "step": 13150
    },
    {
      "epoch": 1.0153065148834706,
      "grad_norm": 1.303915023803711,
      "learning_rate": 0.0001340830584629769,
      "loss": 1.4309,
      "step": 13200
    },
    {
      "epoch": 1.0191523728943928,
      "grad_norm": 0.8854748606681824,
      "learning_rate": 0.00013382333844116044,
      "loss": 1.4395,
      "step": 13250
    },
    {
      "epoch": 1.0191523728943928,
      "eval_loss": 1.5238608121871948,
      "eval_runtime": 17.7588,
      "eval_samples_per_second": 56.31,
      "eval_steps_per_second": 14.077,
      "step": 13250
    },
    {
      "epoch": 1.022998230905315,
      "grad_norm": 1.5649653673171997,
      "learning_rate": 0.00013356361841934395,
      "loss": 1.5022,
      "step": 13300
    },
    {
      "epoch": 1.0268440889162371,
      "grad_norm": 1.6031616926193237,
      "learning_rate": 0.0001333038983975275,
      "loss": 1.4315,
      "step": 13350
    },
    {
      "epoch": 1.0306899469271595,
      "grad_norm": 1.3788844347000122,
      "learning_rate": 0.00013304417837571098,
      "loss": 1.578,
      "step": 13400
    },
    {
      "epoch": 1.0345358049380817,
      "grad_norm": 1.4347171783447266,
      "learning_rate": 0.0001327844583538945,
      "loss": 1.3993,
      "step": 13450
    },
    {
      "epoch": 1.038381662949004,
      "grad_norm": 1.9777193069458008,
      "learning_rate": 0.00013252473833207803,
      "loss": 1.4331,
      "step": 13500
    },
    {
      "epoch": 1.038381662949004,
      "eval_loss": 1.5231057405471802,
      "eval_runtime": 17.8114,
      "eval_samples_per_second": 56.144,
      "eval_steps_per_second": 14.036,
      "step": 13500
    },
    {
      "epoch": 1.042227520959926,
      "grad_norm": 2.056574583053589,
      "learning_rate": 0.00013226501831026155,
      "loss": 1.5387,
      "step": 13550
    },
    {
      "epoch": 1.0460733789708483,
      "grad_norm": 1.41805899143219,
      "learning_rate": 0.00013200529828844506,
      "loss": 1.4194,
      "step": 13600
    },
    {
      "epoch": 1.0499192369817707,
      "grad_norm": 1.5727626085281372,
      "learning_rate": 0.0001317455782666286,
      "loss": 1.4763,
      "step": 13650
    },
    {
      "epoch": 1.0537650949926929,
      "grad_norm": 1.8175796270370483,
      "learning_rate": 0.0001314858582448121,
      "loss": 1.5232,
      "step": 13700
    },
    {
      "epoch": 1.057610953003615,
      "grad_norm": 1.459721565246582,
      "learning_rate": 0.0001312261382229956,
      "loss": 1.4926,
      "step": 13750
    },
    {
      "epoch": 1.057610953003615,
      "eval_loss": 1.5073590278625488,
      "eval_runtime": 17.8208,
      "eval_samples_per_second": 56.114,
      "eval_steps_per_second": 14.029,
      "step": 13750
    },
    {
      "epoch": 1.0614568110145373,
      "grad_norm": 1.7236889600753784,
      "learning_rate": 0.00013096641820117914,
      "loss": 1.4485,
      "step": 13800
    },
    {
      "epoch": 1.0653026690254597,
      "grad_norm": 1.1652172803878784,
      "learning_rate": 0.00013070669817936265,
      "loss": 1.4706,
      "step": 13850
    },
    {
      "epoch": 1.0691485270363819,
      "grad_norm": 1.1279985904693604,
      "learning_rate": 0.00013044697815754616,
      "loss": 1.5507,
      "step": 13900
    },
    {
      "epoch": 1.072994385047304,
      "grad_norm": 2.2368061542510986,
      "learning_rate": 0.0001301872581357297,
      "loss": 1.5184,
      "step": 13950
    },
    {
      "epoch": 1.0768402430582262,
      "grad_norm": 1.1515541076660156,
      "learning_rate": 0.00012992753811391322,
      "loss": 1.5184,
      "step": 14000
    },
    {
      "epoch": 1.0768402430582262,
      "eval_loss": 1.5123000144958496,
      "eval_runtime": 17.8325,
      "eval_samples_per_second": 56.077,
      "eval_steps_per_second": 14.019,
      "step": 14000
    },
    {
      "epoch": 1.0806861010691486,
      "grad_norm": 2.1172475814819336,
      "learning_rate": 0.00012966781809209673,
      "loss": 1.4055,
      "step": 14050
    },
    {
      "epoch": 1.0845319590800708,
      "grad_norm": 1.196999430656433,
      "learning_rate": 0.00012940809807028024,
      "loss": 1.464,
      "step": 14100
    },
    {
      "epoch": 1.088377817090993,
      "grad_norm": 1.3582040071487427,
      "learning_rate": 0.00012914837804846375,
      "loss": 1.4502,
      "step": 14150
    },
    {
      "epoch": 1.0922236751019152,
      "grad_norm": 1.6588162183761597,
      "learning_rate": 0.0001288886580266473,
      "loss": 1.5174,
      "step": 14200
    },
    {
      "epoch": 1.0960695331128374,
      "grad_norm": 1.7531650066375732,
      "learning_rate": 0.0001286289380048308,
      "loss": 1.505,
      "step": 14250
    },
    {
      "epoch": 1.0960695331128374,
      "eval_loss": 1.5160688161849976,
      "eval_runtime": 17.764,
      "eval_samples_per_second": 56.294,
      "eval_steps_per_second": 14.073,
      "step": 14250
    },
    {
      "epoch": 1.0999153911237598,
      "grad_norm": 1.868784785270691,
      "learning_rate": 0.00012836921798301432,
      "loss": 1.5544,
      "step": 14300
    },
    {
      "epoch": 1.103761249134682,
      "grad_norm": 1.9493080377578735,
      "learning_rate": 0.00012810949796119783,
      "loss": 1.533,
      "step": 14350
    },
    {
      "epoch": 1.1076071071456042,
      "grad_norm": 0.7309526801109314,
      "learning_rate": 0.00012784977793938135,
      "loss": 1.4672,
      "step": 14400
    },
    {
      "epoch": 1.1114529651565264,
      "grad_norm": 1.3281447887420654,
      "learning_rate": 0.00012759005791756486,
      "loss": 1.3874,
      "step": 14450
    },
    {
      "epoch": 1.1152988231674485,
      "grad_norm": 1.0158611536026,
      "learning_rate": 0.0001273303378957484,
      "loss": 1.4966,
      "step": 14500
    },
    {
      "epoch": 1.1152988231674485,
      "eval_loss": 1.5160739421844482,
      "eval_runtime": 17.6672,
      "eval_samples_per_second": 56.602,
      "eval_steps_per_second": 14.151,
      "step": 14500
    },
    {
      "epoch": 1.119144681178371,
      "grad_norm": 1.6422228813171387,
      "learning_rate": 0.0001270706178739319,
      "loss": 1.5494,
      "step": 14550
    },
    {
      "epoch": 1.1229905391892931,
      "grad_norm": 0.7187716960906982,
      "learning_rate": 0.00012681089785211542,
      "loss": 1.4221,
      "step": 14600
    },
    {
      "epoch": 1.1268363972002153,
      "grad_norm": 1.2605098485946655,
      "learning_rate": 0.00012655117783029896,
      "loss": 1.4299,
      "step": 14650
    },
    {
      "epoch": 1.1306822552111375,
      "grad_norm": 2.598015069961548,
      "learning_rate": 0.00012629145780848245,
      "loss": 1.5757,
      "step": 14700
    },
    {
      "epoch": 1.13452811322206,
      "grad_norm": 1.4004614353179932,
      "learning_rate": 0.00012603173778666596,
      "loss": 1.482,
      "step": 14750
    },
    {
      "epoch": 1.13452811322206,
      "eval_loss": 1.5089725255966187,
      "eval_runtime": 17.9036,
      "eval_samples_per_second": 55.855,
      "eval_steps_per_second": 13.964,
      "step": 14750
    },
    {
      "epoch": 1.1383739712329821,
      "grad_norm": 1.3800735473632812,
      "learning_rate": 0.0001257720177648495,
      "loss": 1.5285,
      "step": 14800
    },
    {
      "epoch": 1.1422198292439043,
      "grad_norm": 1.3741459846496582,
      "learning_rate": 0.00012551229774303301,
      "loss": 1.5242,
      "step": 14850
    },
    {
      "epoch": 1.1460656872548265,
      "grad_norm": 2.232680559158325,
      "learning_rate": 0.00012525257772121653,
      "loss": 1.4483,
      "step": 14900
    },
    {
      "epoch": 1.149911545265749,
      "grad_norm": 1.4408409595489502,
      "learning_rate": 0.00012499285769940007,
      "loss": 1.5414,
      "step": 14950
    },
    {
      "epoch": 1.153757403276671,
      "grad_norm": 1.5221819877624512,
      "learning_rate": 0.00012473313767758355,
      "loss": 1.5246,
      "step": 15000
    },
    {
      "epoch": 1.153757403276671,
      "eval_loss": 1.516871452331543,
      "eval_runtime": 17.9308,
      "eval_samples_per_second": 55.77,
      "eval_steps_per_second": 13.943,
      "step": 15000
    },
    {
      "epoch": 1.1576032612875933,
      "grad_norm": 1.5970553159713745,
      "learning_rate": 0.0001244734176557671,
      "loss": 1.4515,
      "step": 15050
    },
    {
      "epoch": 1.1614491192985155,
      "grad_norm": 1.0201988220214844,
      "learning_rate": 0.0001242136976339506,
      "loss": 1.5431,
      "step": 15100
    },
    {
      "epoch": 1.1652949773094377,
      "grad_norm": 0.9830596446990967,
      "learning_rate": 0.00012395397761213412,
      "loss": 1.4581,
      "step": 15150
    },
    {
      "epoch": 1.16914083532036,
      "grad_norm": 2.3363943099975586,
      "learning_rate": 0.00012369425759031766,
      "loss": 1.5147,
      "step": 15200
    },
    {
      "epoch": 1.1729866933312822,
      "grad_norm": 1.8213731050491333,
      "learning_rate": 0.00012343453756850117,
      "loss": 1.5497,
      "step": 15250
    },
    {
      "epoch": 1.1729866933312822,
      "eval_loss": 1.5053696632385254,
      "eval_runtime": 17.7358,
      "eval_samples_per_second": 56.383,
      "eval_steps_per_second": 14.096,
      "step": 15250
    },
    {
      "epoch": 1.1768325513422044,
      "grad_norm": 1.1468195915222168,
      "learning_rate": 0.00012317481754668468,
      "loss": 1.4292,
      "step": 15300
    },
    {
      "epoch": 1.1806784093531266,
      "grad_norm": 1.2235878705978394,
      "learning_rate": 0.0001229150975248682,
      "loss": 1.4428,
      "step": 15350
    },
    {
      "epoch": 1.1845242673640488,
      "grad_norm": 1.5518691539764404,
      "learning_rate": 0.0001226553775030517,
      "loss": 1.4889,
      "step": 15400
    },
    {
      "epoch": 1.1883701253749712,
      "grad_norm": 1.217755913734436,
      "learning_rate": 0.00012239565748123522,
      "loss": 1.4841,
      "step": 15450
    },
    {
      "epoch": 1.1922159833858934,
      "grad_norm": 1.6093647480010986,
      "learning_rate": 0.00012213593745941876,
      "loss": 1.5461,
      "step": 15500
    },
    {
      "epoch": 1.1922159833858934,
      "eval_loss": 1.5013692378997803,
      "eval_runtime": 18.0052,
      "eval_samples_per_second": 55.539,
      "eval_steps_per_second": 13.885,
      "step": 15500
    },
    {
      "epoch": 1.1960618413968156,
      "grad_norm": 3.367638349533081,
      "learning_rate": 0.00012187621743760226,
      "loss": 1.4261,
      "step": 15550
    },
    {
      "epoch": 1.1999076994077378,
      "grad_norm": 1.6329169273376465,
      "learning_rate": 0.00012161649741578578,
      "loss": 1.4475,
      "step": 15600
    },
    {
      "epoch": 1.2037535574186602,
      "grad_norm": 1.397910475730896,
      "learning_rate": 0.00012135677739396932,
      "loss": 1.4747,
      "step": 15650
    },
    {
      "epoch": 1.2075994154295824,
      "grad_norm": 1.7463736534118652,
      "learning_rate": 0.00012109705737215283,
      "loss": 1.4745,
      "step": 15700
    },
    {
      "epoch": 1.2114452734405046,
      "grad_norm": 1.8097542524337769,
      "learning_rate": 0.00012083733735033633,
      "loss": 1.5276,
      "step": 15750
    },
    {
      "epoch": 1.2114452734405046,
      "eval_loss": 1.4955236911773682,
      "eval_runtime": 17.9054,
      "eval_samples_per_second": 55.849,
      "eval_steps_per_second": 13.962,
      "step": 15750
    },
    {
      "epoch": 1.2152911314514268,
      "grad_norm": 1.745730996131897,
      "learning_rate": 0.00012057761732851987,
      "loss": 1.5282,
      "step": 15800
    },
    {
      "epoch": 1.2191369894623492,
      "grad_norm": 1.3859128952026367,
      "learning_rate": 0.00012031789730670338,
      "loss": 1.4795,
      "step": 15850
    },
    {
      "epoch": 1.2229828474732714,
      "grad_norm": 1.5086127519607544,
      "learning_rate": 0.0001200581772848869,
      "loss": 1.437,
      "step": 15900
    },
    {
      "epoch": 1.2268287054841935,
      "grad_norm": 2.088292121887207,
      "learning_rate": 0.00011979845726307042,
      "loss": 1.5011,
      "step": 15950
    },
    {
      "epoch": 1.2306745634951157,
      "grad_norm": 1.2746011018753052,
      "learning_rate": 0.00011953873724125393,
      "loss": 1.4596,
      "step": 16000
    },
    {
      "epoch": 1.2306745634951157,
      "eval_loss": 1.4984314441680908,
      "eval_runtime": 17.8066,
      "eval_samples_per_second": 56.159,
      "eval_steps_per_second": 14.04,
      "step": 16000
    },
    {
      "epoch": 1.234520421506038,
      "grad_norm": 6.292486667633057,
      "learning_rate": 0.00011927901721943746,
      "loss": 1.3934,
      "step": 16050
    },
    {
      "epoch": 1.2383662795169603,
      "grad_norm": 1.6574532985687256,
      "learning_rate": 0.00011901929719762097,
      "loss": 1.4821,
      "step": 16100
    },
    {
      "epoch": 1.2422121375278825,
      "grad_norm": 1.8651037216186523,
      "learning_rate": 0.00011875957717580448,
      "loss": 1.49,
      "step": 16150
    },
    {
      "epoch": 1.2460579955388047,
      "grad_norm": 1.3768175840377808,
      "learning_rate": 0.00011849985715398801,
      "loss": 1.4579,
      "step": 16200
    },
    {
      "epoch": 1.2499038535497269,
      "grad_norm": 1.1569020748138428,
      "learning_rate": 0.00011824013713217152,
      "loss": 1.4029,
      "step": 16250
    },
    {
      "epoch": 1.2499038535497269,
      "eval_loss": 1.4893407821655273,
      "eval_runtime": 17.9552,
      "eval_samples_per_second": 55.694,
      "eval_steps_per_second": 13.924,
      "step": 16250
    },
    {
      "epoch": 1.253749711560649,
      "grad_norm": 1.8632296323776245,
      "learning_rate": 0.00011798041711035504,
      "loss": 1.4592,
      "step": 16300
    },
    {
      "epoch": 1.2575955695715715,
      "grad_norm": 1.8080470561981201,
      "learning_rate": 0.00011772069708853856,
      "loss": 1.4678,
      "step": 16350
    },
    {
      "epoch": 1.2614414275824937,
      "grad_norm": 1.4193981885910034,
      "learning_rate": 0.00011746097706672208,
      "loss": 1.5031,
      "step": 16400
    },
    {
      "epoch": 1.2652872855934159,
      "grad_norm": 1.5050238370895386,
      "learning_rate": 0.00011720125704490559,
      "loss": 1.4107,
      "step": 16450
    },
    {
      "epoch": 1.2691331436043383,
      "grad_norm": 1.12454092502594,
      "learning_rate": 0.00011694153702308911,
      "loss": 1.4572,
      "step": 16500
    },
    {
      "epoch": 1.2691331436043383,
      "eval_loss": 1.4972718954086304,
      "eval_runtime": 17.8677,
      "eval_samples_per_second": 55.967,
      "eval_steps_per_second": 13.992,
      "step": 16500
    },
    {
      "epoch": 1.2729790016152602,
      "grad_norm": 1.3523976802825928,
      "learning_rate": 0.00011668181700127263,
      "loss": 1.4796,
      "step": 16550
    },
    {
      "epoch": 1.2768248596261826,
      "grad_norm": 1.9770869016647339,
      "learning_rate": 0.00011642209697945614,
      "loss": 1.4756,
      "step": 16600
    },
    {
      "epoch": 1.2806707176371048,
      "grad_norm": 1.7973159551620483,
      "learning_rate": 0.00011616237695763967,
      "loss": 1.4163,
      "step": 16650
    },
    {
      "epoch": 1.284516575648027,
      "grad_norm": 1.3054739236831665,
      "learning_rate": 0.00011590265693582318,
      "loss": 1.4105,
      "step": 16700
    },
    {
      "epoch": 1.2883624336589494,
      "grad_norm": 1.457047939300537,
      "learning_rate": 0.00011564293691400672,
      "loss": 1.4698,
      "step": 16750
    },
    {
      "epoch": 1.2883624336589494,
      "eval_loss": 1.4798808097839355,
      "eval_runtime": 17.8696,
      "eval_samples_per_second": 55.961,
      "eval_steps_per_second": 13.99,
      "step": 16750
    },
    {
      "epoch": 1.2922082916698716,
      "grad_norm": 2.117663860321045,
      "learning_rate": 0.00011538841129262656,
      "loss": 1.4252,
      "step": 16800
    },
    {
      "epoch": 1.2960541496807938,
      "grad_norm": 3.7321341037750244,
      "learning_rate": 0.00011512869127081007,
      "loss": 1.4818,
      "step": 16850
    },
    {
      "epoch": 1.299900007691716,
      "grad_norm": 2.497528553009033,
      "learning_rate": 0.00011486897124899358,
      "loss": 1.4176,
      "step": 16900
    },
    {
      "epoch": 1.3037458657026382,
      "grad_norm": 1.6085398197174072,
      "learning_rate": 0.00011460925122717712,
      "loss": 1.4373,
      "step": 16950
    },
    {
      "epoch": 1.3075917237135606,
      "grad_norm": 1.8937525749206543,
      "learning_rate": 0.00011434953120536062,
      "loss": 1.4823,
      "step": 17000
    },
    {
      "epoch": 1.3075917237135606,
      "eval_loss": 1.4854488372802734,
      "eval_runtime": 17.8761,
      "eval_samples_per_second": 55.941,
      "eval_steps_per_second": 13.985,
      "step": 17000
    },
    {
      "epoch": 1.3114375817244828,
      "grad_norm": 1.6637665033340454,
      "learning_rate": 0.00011408981118354414,
      "loss": 1.5071,
      "step": 17050
    },
    {
      "epoch": 1.315283439735405,
      "grad_norm": 2.0815582275390625,
      "learning_rate": 0.00011383009116172768,
      "loss": 1.4606,
      "step": 17100
    },
    {
      "epoch": 1.3191292977463271,
      "grad_norm": 1.6492595672607422,
      "learning_rate": 0.00011357037113991117,
      "loss": 1.4408,
      "step": 17150
    },
    {
      "epoch": 1.3229751557572493,
      "grad_norm": 0.8617509603500366,
      "learning_rate": 0.00011331065111809469,
      "loss": 1.497,
      "step": 17200
    },
    {
      "epoch": 1.3268210137681717,
      "grad_norm": 1.6395294666290283,
      "learning_rate": 0.00011305093109627823,
      "loss": 1.4774,
      "step": 17250
    },
    {
      "epoch": 1.3268210137681717,
      "eval_loss": 1.4835026264190674,
      "eval_runtime": 18.0136,
      "eval_samples_per_second": 55.514,
      "eval_steps_per_second": 13.878,
      "step": 17250
    },
    {
      "epoch": 1.330666871779094,
      "grad_norm": 2.7765560150146484,
      "learning_rate": 0.00011279121107446174,
      "loss": 1.4864,
      "step": 17300
    },
    {
      "epoch": 1.3345127297900161,
      "grad_norm": 1.2104064226150513,
      "learning_rate": 0.00011253149105264524,
      "loss": 1.4075,
      "step": 17350
    },
    {
      "epoch": 1.3383585878009385,
      "grad_norm": 1.6772801876068115,
      "learning_rate": 0.00011227177103082878,
      "loss": 1.4629,
      "step": 17400
    },
    {
      "epoch": 1.3422044458118605,
      "grad_norm": 2.254371404647827,
      "learning_rate": 0.00011201205100901229,
      "loss": 1.4447,
      "step": 17450
    },
    {
      "epoch": 1.346050303822783,
      "grad_norm": 2.2015669345855713,
      "learning_rate": 0.00011175233098719582,
      "loss": 1.4664,
      "step": 17500
    },
    {
      "epoch": 1.346050303822783,
      "eval_loss": 1.496685266494751,
      "eval_runtime": 17.7963,
      "eval_samples_per_second": 56.191,
      "eval_steps_per_second": 14.048,
      "step": 17500
    },
    {
      "epoch": 1.349896161833705,
      "grad_norm": 1.740045428276062,
      "learning_rate": 0.00011149261096537933,
      "loss": 1.4486,
      "step": 17550
    },
    {
      "epoch": 1.3537420198446273,
      "grad_norm": 1.299919605255127,
      "learning_rate": 0.00011123289094356284,
      "loss": 1.4868,
      "step": 17600
    },
    {
      "epoch": 1.3575878778555497,
      "grad_norm": 1.6288009881973267,
      "learning_rate": 0.00011097317092174637,
      "loss": 1.4595,
      "step": 17650
    },
    {
      "epoch": 1.3614337358664719,
      "grad_norm": 0.8747851252555847,
      "learning_rate": 0.00011071345089992988,
      "loss": 1.4241,
      "step": 17700
    },
    {
      "epoch": 1.365279593877394,
      "grad_norm": 1.9510573148727417,
      "learning_rate": 0.0001104537308781134,
      "loss": 1.496,
      "step": 17750
    },
    {
      "epoch": 1.365279593877394,
      "eval_loss": 1.4812238216400146,
      "eval_runtime": 17.8955,
      "eval_samples_per_second": 55.88,
      "eval_steps_per_second": 13.97,
      "step": 17750
    },
    {
      "epoch": 1.3691254518883162,
      "grad_norm": 1.4853876829147339,
      "learning_rate": 0.00011019401085629692,
      "loss": 1.4645,
      "step": 17800
    },
    {
      "epoch": 1.3729713098992384,
      "grad_norm": 1.5125057697296143,
      "learning_rate": 0.00010993429083448044,
      "loss": 1.4052,
      "step": 17850
    },
    {
      "epoch": 1.3768171679101608,
      "grad_norm": 0.7320863008499146,
      "learning_rate": 0.00010967457081266395,
      "loss": 1.5016,
      "step": 17900
    },
    {
      "epoch": 1.380663025921083,
      "grad_norm": 1.9995285272598267,
      "learning_rate": 0.00010941485079084747,
      "loss": 1.4234,
      "step": 17950
    },
    {
      "epoch": 1.3845088839320052,
      "grad_norm": 0.8304823637008667,
      "learning_rate": 0.00010915513076903099,
      "loss": 1.4068,
      "step": 18000
    },
    {
      "epoch": 1.3845088839320052,
      "eval_loss": 1.4845945835113525,
      "eval_runtime": 17.7402,
      "eval_samples_per_second": 56.369,
      "eval_steps_per_second": 14.092,
      "step": 18000
    },
    {
      "epoch": 1.3883547419429274,
      "grad_norm": 1.8189436197280884,
      "learning_rate": 0.0001088954107472145,
      "loss": 1.4872,
      "step": 18050
    },
    {
      "epoch": 1.3922005999538496,
      "grad_norm": 1.4212762117385864,
      "learning_rate": 0.00010863569072539803,
      "loss": 1.4067,
      "step": 18100
    },
    {
      "epoch": 1.396046457964772,
      "grad_norm": 1.9733264446258545,
      "learning_rate": 0.00010837597070358154,
      "loss": 1.4787,
      "step": 18150
    },
    {
      "epoch": 1.3998923159756942,
      "grad_norm": 1.320064663887024,
      "learning_rate": 0.00010811625068176505,
      "loss": 1.5242,
      "step": 18200
    },
    {
      "epoch": 1.4037381739866164,
      "grad_norm": 1.4177141189575195,
      "learning_rate": 0.00010785653065994858,
      "loss": 1.4719,
      "step": 18250
    },
    {
      "epoch": 1.4037381739866164,
      "eval_loss": 1.4992233514785767,
      "eval_runtime": 17.8574,
      "eval_samples_per_second": 55.999,
      "eval_steps_per_second": 14.0,
      "step": 18250
    },
    {
      "epoch": 1.4075840319975388,
      "grad_norm": 1.8219791650772095,
      "learning_rate": 0.00010759681063813209,
      "loss": 1.4392,
      "step": 18300
    },
    {
      "epoch": 1.4114298900084608,
      "grad_norm": 1.3196603059768677,
      "learning_rate": 0.00010733709061631563,
      "loss": 1.4837,
      "step": 18350
    },
    {
      "epoch": 1.4152757480193832,
      "grad_norm": 1.01405668258667,
      "learning_rate": 0.00010707737059449914,
      "loss": 1.455,
      "step": 18400
    },
    {
      "epoch": 1.4191216060303053,
      "grad_norm": 1.8538917303085327,
      "learning_rate": 0.00010681765057268264,
      "loss": 1.4209,
      "step": 18450
    },
    {
      "epoch": 1.4229674640412275,
      "grad_norm": 0.8785907030105591,
      "learning_rate": 0.00010655793055086618,
      "loss": 1.4716,
      "step": 18500
    },
    {
      "epoch": 1.4229674640412275,
      "eval_loss": 1.485592246055603,
      "eval_runtime": 17.7246,
      "eval_samples_per_second": 56.419,
      "eval_steps_per_second": 14.105,
      "step": 18500
    },
    {
      "epoch": 1.42681332205215,
      "grad_norm": 1.0700381994247437,
      "learning_rate": 0.0001062982105290497,
      "loss": 1.3846,
      "step": 18550
    },
    {
      "epoch": 1.4306591800630721,
      "grad_norm": 1.2848351001739502,
      "learning_rate": 0.0001060384905072332,
      "loss": 1.4874,
      "step": 18600
    },
    {
      "epoch": 1.4345050380739943,
      "grad_norm": 1.4261386394500732,
      "learning_rate": 0.00010577877048541674,
      "loss": 1.5771,
      "step": 18650
    },
    {
      "epoch": 1.4383508960849165,
      "grad_norm": 1.3613426685333252,
      "learning_rate": 0.00010551905046360025,
      "loss": 1.4445,
      "step": 18700
    },
    {
      "epoch": 1.4421967540958387,
      "grad_norm": 0.8625685572624207,
      "learning_rate": 0.00010525933044178375,
      "loss": 1.3845,
      "step": 18750
    },
    {
      "epoch": 1.4421967540958387,
      "eval_loss": 1.4757392406463623,
      "eval_runtime": 17.8903,
      "eval_samples_per_second": 55.896,
      "eval_steps_per_second": 13.974,
      "step": 18750
    },
    {
      "epoch": 1.446042612106761,
      "grad_norm": 1.5575672388076782,
      "learning_rate": 0.00010499961041996729,
      "loss": 1.5387,
      "step": 18800
    },
    {
      "epoch": 1.4498884701176833,
      "grad_norm": 2.2173306941986084,
      "learning_rate": 0.0001047398903981508,
      "loss": 1.459,
      "step": 18850
    },
    {
      "epoch": 1.4537343281286055,
      "grad_norm": 1.904809832572937,
      "learning_rate": 0.00010448017037633431,
      "loss": 1.5229,
      "step": 18900
    },
    {
      "epoch": 1.4575801861395277,
      "grad_norm": 1.8832893371582031,
      "learning_rate": 0.00010422045035451784,
      "loss": 1.4093,
      "step": 18950
    },
    {
      "epoch": 1.4614260441504499,
      "grad_norm": 1.852971076965332,
      "learning_rate": 0.00010396073033270135,
      "loss": 1.389,
      "step": 19000
    },
    {
      "epoch": 1.4614260441504499,
      "eval_loss": 1.4788576364517212,
      "eval_runtime": 17.7751,
      "eval_samples_per_second": 56.258,
      "eval_steps_per_second": 14.065,
      "step": 19000
    },
    {
      "epoch": 1.4652719021613723,
      "grad_norm": 1.2875189781188965,
      "learning_rate": 0.00010370101031088487,
      "loss": 1.4826,
      "step": 19050
    },
    {
      "epoch": 1.4691177601722945,
      "grad_norm": 1.7036223411560059,
      "learning_rate": 0.00010344129028906839,
      "loss": 1.4392,
      "step": 19100
    },
    {
      "epoch": 1.4729636181832166,
      "grad_norm": 1.537514328956604,
      "learning_rate": 0.0001031815702672519,
      "loss": 1.4846,
      "step": 19150
    },
    {
      "epoch": 1.476809476194139,
      "grad_norm": 0.9159242510795593,
      "learning_rate": 0.00010292185024543543,
      "loss": 1.4668,
      "step": 19200
    },
    {
      "epoch": 1.480655334205061,
      "grad_norm": 3.47868013381958,
      "learning_rate": 0.00010266213022361894,
      "loss": 1.493,
      "step": 19250
    },
    {
      "epoch": 1.480655334205061,
      "eval_loss": 1.4858986139297485,
      "eval_runtime": 17.7738,
      "eval_samples_per_second": 56.263,
      "eval_steps_per_second": 14.066,
      "step": 19250
    },
    {
      "epoch": 1.4845011922159834,
      "grad_norm": 1.467437505722046,
      "learning_rate": 0.00010240241020180246,
      "loss": 1.4639,
      "step": 19300
    },
    {
      "epoch": 1.4883470502269056,
      "grad_norm": 1.2710049152374268,
      "learning_rate": 0.00010214269017998598,
      "loss": 1.4244,
      "step": 19350
    },
    {
      "epoch": 1.4921929082378278,
      "grad_norm": 2.0059661865234375,
      "learning_rate": 0.0001018829701581695,
      "loss": 1.4258,
      "step": 19400
    },
    {
      "epoch": 1.4960387662487502,
      "grad_norm": 1.7536308765411377,
      "learning_rate": 0.00010162325013635301,
      "loss": 1.396,
      "step": 19450
    },
    {
      "epoch": 1.4998846242596724,
      "grad_norm": 0.9684279561042786,
      "learning_rate": 0.00010136353011453655,
      "loss": 1.4598,
      "step": 19500
    },
    {
      "epoch": 1.4998846242596724,
      "eval_loss": 1.4841110706329346,
      "eval_runtime": 17.9151,
      "eval_samples_per_second": 55.819,
      "eval_steps_per_second": 13.955,
      "step": 19500
    },
    {
      "epoch": 1.5037304822705946,
      "grad_norm": 0.9244908690452576,
      "learning_rate": 0.00010110381009272005,
      "loss": 1.3834,
      "step": 19550
    },
    {
      "epoch": 1.5075763402815168,
      "grad_norm": 1.6488862037658691,
      "learning_rate": 0.00010084409007090356,
      "loss": 1.5311,
      "step": 19600
    },
    {
      "epoch": 1.511422198292439,
      "grad_norm": 1.9130067825317383,
      "learning_rate": 0.0001005843700490871,
      "loss": 1.5212,
      "step": 19650
    },
    {
      "epoch": 1.5152680563033614,
      "grad_norm": 1.326277256011963,
      "learning_rate": 0.0001003246500272706,
      "loss": 1.4067,
      "step": 19700
    },
    {
      "epoch": 1.5191139143142836,
      "grad_norm": 1.7258195877075195,
      "learning_rate": 0.00010006493000545411,
      "loss": 1.4844,
      "step": 19750
    },
    {
      "epoch": 1.5191139143142836,
      "eval_loss": 1.4816969633102417,
      "eval_runtime": 17.826,
      "eval_samples_per_second": 56.098,
      "eval_steps_per_second": 14.024,
      "step": 19750
    },
    {
      "epoch": 1.5229597723252057,
      "grad_norm": 1.8164838552474976,
      "learning_rate": 9.980520998363765e-05,
      "loss": 1.4233,
      "step": 19800
    },
    {
      "epoch": 1.5268056303361282,
      "grad_norm": 1.5884016752243042,
      "learning_rate": 9.954548996182115e-05,
      "loss": 1.4313,
      "step": 19850
    },
    {
      "epoch": 1.5306514883470501,
      "grad_norm": 1.5381648540496826,
      "learning_rate": 9.928576994000468e-05,
      "loss": 1.4789,
      "step": 19900
    },
    {
      "epoch": 1.5344973463579725,
      "grad_norm": 1.6448626518249512,
      "learning_rate": 9.90260499181882e-05,
      "loss": 1.3756,
      "step": 19950
    },
    {
      "epoch": 1.5383432043688947,
      "grad_norm": 1.6137230396270752,
      "learning_rate": 9.876632989637172e-05,
      "loss": 1.4171,
      "step": 20000
    },
    {
      "epoch": 1.5383432043688947,
      "eval_loss": 1.4770597219467163,
      "eval_runtime": 17.9365,
      "eval_samples_per_second": 55.752,
      "eval_steps_per_second": 13.938,
      "step": 20000
    },
    {
      "epoch": 1.542189062379817,
      "grad_norm": 1.7058050632476807,
      "learning_rate": 9.850660987455523e-05,
      "loss": 1.4646,
      "step": 20050
    },
    {
      "epoch": 1.5460349203907393,
      "grad_norm": 2.2624917030334473,
      "learning_rate": 9.824688985273876e-05,
      "loss": 1.4586,
      "step": 20100
    },
    {
      "epoch": 1.5498807784016613,
      "grad_norm": 2.216883420944214,
      "learning_rate": 9.798716983092227e-05,
      "loss": 1.4376,
      "step": 20150
    },
    {
      "epoch": 1.5537266364125837,
      "grad_norm": 1.9749584197998047,
      "learning_rate": 9.772744980910578e-05,
      "loss": 1.47,
      "step": 20200
    },
    {
      "epoch": 1.5575724944235059,
      "grad_norm": 2.185480833053589,
      "learning_rate": 9.746772978728931e-05,
      "loss": 1.39,
      "step": 20250
    },
    {
      "epoch": 1.5575724944235059,
      "eval_loss": 1.4724150896072388,
      "eval_runtime": 17.8191,
      "eval_samples_per_second": 56.12,
      "eval_steps_per_second": 14.03,
      "step": 20250
    },
    {
      "epoch": 1.561418352434428,
      "grad_norm": 1.5810290575027466,
      "learning_rate": 9.720800976547284e-05,
      "loss": 1.4267,
      "step": 20300
    },
    {
      "epoch": 1.5652642104453505,
      "grad_norm": 2.0329344272613525,
      "learning_rate": 9.694828974365633e-05,
      "loss": 1.5066,
      "step": 20350
    },
    {
      "epoch": 1.5691100684562724,
      "grad_norm": 2.7335126399993896,
      "learning_rate": 9.668856972183986e-05,
      "loss": 1.4749,
      "step": 20400
    },
    {
      "epoch": 1.5729559264671948,
      "grad_norm": 1.0576220750808716,
      "learning_rate": 9.642884970002339e-05,
      "loss": 1.4318,
      "step": 20450
    },
    {
      "epoch": 1.576801784478117,
      "grad_norm": 0.6857870817184448,
      "learning_rate": 9.616912967820689e-05,
      "loss": 1.3567,
      "step": 20500
    },
    {
      "epoch": 1.576801784478117,
      "eval_loss": 1.479669213294983,
      "eval_runtime": 17.8378,
      "eval_samples_per_second": 56.061,
      "eval_steps_per_second": 14.015,
      "step": 20500
    },
    {
      "epoch": 1.5806476424890392,
      "grad_norm": 1.430114507675171,
      "learning_rate": 9.590940965639041e-05,
      "loss": 1.3714,
      "step": 20550
    },
    {
      "epoch": 1.5844935004999616,
      "grad_norm": 1.7613717317581177,
      "learning_rate": 9.564968963457394e-05,
      "loss": 1.4285,
      "step": 20600
    },
    {
      "epoch": 1.5883393585108838,
      "grad_norm": 1.3678529262542725,
      "learning_rate": 9.538996961275745e-05,
      "loss": 1.4463,
      "step": 20650
    },
    {
      "epoch": 1.592185216521806,
      "grad_norm": 3.1135504245758057,
      "learning_rate": 9.513024959094097e-05,
      "loss": 1.4287,
      "step": 20700
    },
    {
      "epoch": 1.5960310745327284,
      "grad_norm": 1.7866570949554443,
      "learning_rate": 9.487052956912449e-05,
      "loss": 1.4515,
      "step": 20750
    },
    {
      "epoch": 1.5960310745327284,
      "eval_loss": 1.4817472696304321,
      "eval_runtime": 17.9267,
      "eval_samples_per_second": 55.783,
      "eval_steps_per_second": 13.946,
      "step": 20750
    },
    {
      "epoch": 1.5998769325436504,
      "grad_norm": 1.2406786680221558,
      "learning_rate": 9.4610809547308e-05,
      "loss": 1.5364,
      "step": 20800
    },
    {
      "epoch": 1.6037227905545728,
      "grad_norm": 1.9183951616287231,
      "learning_rate": 9.435108952549152e-05,
      "loss": 1.5169,
      "step": 20850
    },
    {
      "epoch": 1.607568648565495,
      "grad_norm": 1.852089524269104,
      "learning_rate": 9.409136950367504e-05,
      "loss": 1.4668,
      "step": 20900
    },
    {
      "epoch": 1.6114145065764172,
      "grad_norm": 2.3430335521698,
      "learning_rate": 9.383164948185856e-05,
      "loss": 1.4297,
      "step": 20950
    },
    {
      "epoch": 1.6152603645873396,
      "grad_norm": 1.9646743535995483,
      "learning_rate": 9.357192946004207e-05,
      "loss": 1.5143,
      "step": 21000
    },
    {
      "epoch": 1.6152603645873396,
      "eval_loss": 1.4690666198730469,
      "eval_runtime": 17.7563,
      "eval_samples_per_second": 56.318,
      "eval_steps_per_second": 14.079,
      "step": 21000
    },
    {
      "epoch": 1.6191062225982615,
      "grad_norm": 1.8204776048660278,
      "learning_rate": 9.33122094382256e-05,
      "loss": 1.4423,
      "step": 21050
    },
    {
      "epoch": 1.622952080609184,
      "grad_norm": 1.6150448322296143,
      "learning_rate": 9.305248941640912e-05,
      "loss": 1.4368,
      "step": 21100
    },
    {
      "epoch": 1.6267979386201061,
      "grad_norm": 2.4393765926361084,
      "learning_rate": 9.279276939459264e-05,
      "loss": 1.6115,
      "step": 21150
    },
    {
      "epoch": 1.6306437966310283,
      "grad_norm": 1.217399001121521,
      "learning_rate": 9.253304937277615e-05,
      "loss": 1.5792,
      "step": 21200
    },
    {
      "epoch": 1.6344896546419507,
      "grad_norm": 1.4467514753341675,
      "learning_rate": 9.227332935095967e-05,
      "loss": 1.4838,
      "step": 21250
    },
    {
      "epoch": 1.6344896546419507,
      "eval_loss": 1.4741238355636597,
      "eval_runtime": 17.7656,
      "eval_samples_per_second": 56.289,
      "eval_steps_per_second": 14.072,
      "step": 21250
    },
    {
      "epoch": 1.6383355126528727,
      "grad_norm": 1.4274511337280273,
      "learning_rate": 9.201360932914319e-05,
      "loss": 1.4747,
      "step": 21300
    },
    {
      "epoch": 1.642181370663795,
      "grad_norm": 1.5080102682113647,
      "learning_rate": 9.175908370776304e-05,
      "loss": 1.4262,
      "step": 21350
    },
    {
      "epoch": 1.6460272286747173,
      "grad_norm": 1.2717032432556152,
      "learning_rate": 9.149936368594655e-05,
      "loss": 1.4418,
      "step": 21400
    },
    {
      "epoch": 1.6498730866856395,
      "grad_norm": 1.3594080209732056,
      "learning_rate": 9.123964366413006e-05,
      "loss": 1.5184,
      "step": 21450
    },
    {
      "epoch": 1.6537189446965619,
      "grad_norm": 1.896607518196106,
      "learning_rate": 9.097992364231359e-05,
      "loss": 1.4702,
      "step": 21500
    },
    {
      "epoch": 1.6537189446965619,
      "eval_loss": 1.4777735471725464,
      "eval_runtime": 17.8161,
      "eval_samples_per_second": 56.129,
      "eval_steps_per_second": 14.032,
      "step": 21500
    },
    {
      "epoch": 1.657564802707484,
      "grad_norm": 1.3839844465255737,
      "learning_rate": 9.072020362049712e-05,
      "loss": 1.522,
      "step": 21550
    },
    {
      "epoch": 1.6614106607184063,
      "grad_norm": 1.187853455543518,
      "learning_rate": 9.046048359868063e-05,
      "loss": 1.4469,
      "step": 21600
    },
    {
      "epoch": 1.6652565187293287,
      "grad_norm": 2.8398866653442383,
      "learning_rate": 9.020076357686414e-05,
      "loss": 1.4875,
      "step": 21650
    },
    {
      "epoch": 1.6691023767402506,
      "grad_norm": 1.801963448524475,
      "learning_rate": 8.994104355504767e-05,
      "loss": 1.3892,
      "step": 21700
    },
    {
      "epoch": 1.672948234751173,
      "grad_norm": 1.7727116346359253,
      "learning_rate": 8.968132353323118e-05,
      "loss": 1.3886,
      "step": 21750
    },
    {
      "epoch": 1.672948234751173,
      "eval_loss": 1.4662117958068848,
      "eval_runtime": 17.7809,
      "eval_samples_per_second": 56.24,
      "eval_steps_per_second": 14.06,
      "step": 21750
    },
    {
      "epoch": 1.6767940927620952,
      "grad_norm": 1.7052053213119507,
      "learning_rate": 8.94216035114147e-05,
      "loss": 1.4483,
      "step": 21800
    },
    {
      "epoch": 1.6806399507730174,
      "grad_norm": 1.9812465906143188,
      "learning_rate": 8.916188348959822e-05,
      "loss": 1.4325,
      "step": 21850
    },
    {
      "epoch": 1.6844858087839398,
      "grad_norm": 1.2499721050262451,
      "learning_rate": 8.890216346778173e-05,
      "loss": 1.5246,
      "step": 21900
    },
    {
      "epoch": 1.6883316667948618,
      "grad_norm": 1.2503259181976318,
      "learning_rate": 8.864244344596525e-05,
      "loss": 1.4661,
      "step": 21950
    },
    {
      "epoch": 1.6921775248057842,
      "grad_norm": 2.201223134994507,
      "learning_rate": 8.838272342414877e-05,
      "loss": 1.5103,
      "step": 22000
    },
    {
      "epoch": 1.6921775248057842,
      "eval_loss": 1.470120906829834,
      "eval_runtime": 17.7715,
      "eval_samples_per_second": 56.27,
      "eval_steps_per_second": 14.067,
      "step": 22000
    },
    {
      "epoch": 1.6960233828167064,
      "grad_norm": 0.8225556015968323,
      "learning_rate": 8.81230034023323e-05,
      "loss": 1.4372,
      "step": 22050
    },
    {
      "epoch": 1.6998692408276286,
      "grad_norm": 1.495335578918457,
      "learning_rate": 8.78632833805158e-05,
      "loss": 1.4231,
      "step": 22100
    },
    {
      "epoch": 1.703715098838551,
      "grad_norm": 1.6913652420043945,
      "learning_rate": 8.760356335869933e-05,
      "loss": 1.4859,
      "step": 22150
    },
    {
      "epoch": 1.707560956849473,
      "grad_norm": 1.9825598001480103,
      "learning_rate": 8.734384333688285e-05,
      "loss": 1.4354,
      "step": 22200
    },
    {
      "epoch": 1.7114068148603954,
      "grad_norm": 2.212759017944336,
      "learning_rate": 8.708412331506635e-05,
      "loss": 1.3861,
      "step": 22250
    },
    {
      "epoch": 1.7114068148603954,
      "eval_loss": 1.4596961736679077,
      "eval_runtime": 17.8176,
      "eval_samples_per_second": 56.124,
      "eval_steps_per_second": 14.031,
      "step": 22250
    },
    {
      "epoch": 1.7152526728713176,
      "grad_norm": 1.4325975179672241,
      "learning_rate": 8.682440329324988e-05,
      "loss": 1.4805,
      "step": 22300
    },
    {
      "epoch": 1.7190985308822397,
      "grad_norm": 1.9796292781829834,
      "learning_rate": 8.65646832714334e-05,
      "loss": 1.3836,
      "step": 22350
    },
    {
      "epoch": 1.7229443888931621,
      "grad_norm": 1.6221562623977661,
      "learning_rate": 8.630496324961692e-05,
      "loss": 1.4176,
      "step": 22400
    },
    {
      "epoch": 1.7267902469040843,
      "grad_norm": 1.1431959867477417,
      "learning_rate": 8.604524322780043e-05,
      "loss": 1.4553,
      "step": 22450
    },
    {
      "epoch": 1.7306361049150065,
      "grad_norm": 1.1562083959579468,
      "learning_rate": 8.578552320598396e-05,
      "loss": 1.4489,
      "step": 22500
    },
    {
      "epoch": 1.7306361049150065,
      "eval_loss": 1.464021921157837,
      "eval_runtime": 17.8166,
      "eval_samples_per_second": 56.127,
      "eval_steps_per_second": 14.032,
      "step": 22500
    },
    {
      "epoch": 1.734481962925929,
      "grad_norm": 0.8531803488731384,
      "learning_rate": 8.552580318416747e-05,
      "loss": 1.4298,
      "step": 22550
    },
    {
      "epoch": 1.738327820936851,
      "grad_norm": 1.3987632989883423,
      "learning_rate": 8.526608316235098e-05,
      "loss": 1.4629,
      "step": 22600
    },
    {
      "epoch": 1.7421736789477733,
      "grad_norm": 1.4521870613098145,
      "learning_rate": 8.500636314053451e-05,
      "loss": 1.4005,
      "step": 22650
    },
    {
      "epoch": 1.7460195369586955,
      "grad_norm": 1.0557054281234741,
      "learning_rate": 8.474664311871803e-05,
      "loss": 1.4079,
      "step": 22700
    },
    {
      "epoch": 1.7498653949696177,
      "grad_norm": 1.5067927837371826,
      "learning_rate": 8.448692309690155e-05,
      "loss": 1.4664,
      "step": 22750
    },
    {
      "epoch": 1.7498653949696177,
      "eval_loss": 1.4728831052780151,
      "eval_runtime": 17.7456,
      "eval_samples_per_second": 56.352,
      "eval_steps_per_second": 14.088,
      "step": 22750
    },
    {
      "epoch": 1.75371125298054,
      "grad_norm": 1.3237221240997314,
      "learning_rate": 8.422720307508506e-05,
      "loss": 1.5266,
      "step": 22800
    },
    {
      "epoch": 1.757557110991462,
      "grad_norm": 1.4342700242996216,
      "learning_rate": 8.396748305326859e-05,
      "loss": 1.4929,
      "step": 22850
    },
    {
      "epoch": 1.7614029690023845,
      "grad_norm": 5.2202534675598145,
      "learning_rate": 8.37077630314521e-05,
      "loss": 1.457,
      "step": 22900
    },
    {
      "epoch": 1.7652488270133067,
      "grad_norm": 2.7584545612335205,
      "learning_rate": 8.344804300963561e-05,
      "loss": 1.4523,
      "step": 22950
    },
    {
      "epoch": 1.7690946850242288,
      "grad_norm": 1.8208624124526978,
      "learning_rate": 8.318832298781914e-05,
      "loss": 1.4746,
      "step": 23000
    },
    {
      "epoch": 1.7690946850242288,
      "eval_loss": 1.467396855354309,
      "eval_runtime": 17.663,
      "eval_samples_per_second": 56.616,
      "eval_steps_per_second": 14.154,
      "step": 23000
    },
    {
      "epoch": 1.7729405430351513,
      "grad_norm": 1.6651790142059326,
      "learning_rate": 8.292860296600265e-05,
      "loss": 1.4537,
      "step": 23050
    },
    {
      "epoch": 1.7767864010460732,
      "grad_norm": 1.5910587310791016,
      "learning_rate": 8.266888294418618e-05,
      "loss": 1.5305,
      "step": 23100
    },
    {
      "epoch": 1.7806322590569956,
      "grad_norm": 1.2711199522018433,
      "learning_rate": 8.240916292236969e-05,
      "loss": 1.4877,
      "step": 23150
    },
    {
      "epoch": 1.7844781170679178,
      "grad_norm": 1.7133463621139526,
      "learning_rate": 8.21494429005532e-05,
      "loss": 1.4547,
      "step": 23200
    },
    {
      "epoch": 1.78832397507884,
      "grad_norm": 1.6850301027297974,
      "learning_rate": 8.188972287873673e-05,
      "loss": 1.4228,
      "step": 23250
    },
    {
      "epoch": 1.78832397507884,
      "eval_loss": 1.4735645055770874,
      "eval_runtime": 17.7983,
      "eval_samples_per_second": 56.185,
      "eval_steps_per_second": 14.046,
      "step": 23250
    },
    {
      "epoch": 1.7921698330897624,
      "grad_norm": 1.1047898530960083,
      "learning_rate": 8.163000285692024e-05,
      "loss": 1.4817,
      "step": 23300
    },
    {
      "epoch": 1.7960156911006846,
      "grad_norm": 1.8970743417739868,
      "learning_rate": 8.137028283510376e-05,
      "loss": 1.4884,
      "step": 23350
    },
    {
      "epoch": 1.7998615491116068,
      "grad_norm": 1.4730740785598755,
      "learning_rate": 8.111056281328728e-05,
      "loss": 1.4436,
      "step": 23400
    },
    {
      "epoch": 1.8037074071225292,
      "grad_norm": 1.8288697004318237,
      "learning_rate": 8.08508427914708e-05,
      "loss": 1.4614,
      "step": 23450
    },
    {
      "epoch": 1.8075532651334512,
      "grad_norm": 1.2339516878128052,
      "learning_rate": 8.059112276965432e-05,
      "loss": 1.4276,
      "step": 23500
    },
    {
      "epoch": 1.8075532651334512,
      "eval_loss": 1.4655163288116455,
      "eval_runtime": 17.9781,
      "eval_samples_per_second": 55.623,
      "eval_steps_per_second": 13.906,
      "step": 23500
    },
    {
      "epoch": 1.8113991231443736,
      "grad_norm": 1.8125578165054321,
      "learning_rate": 8.033140274783783e-05,
      "loss": 1.4318,
      "step": 23550
    },
    {
      "epoch": 1.8152449811552958,
      "grad_norm": 1.941846489906311,
      "learning_rate": 8.007168272602136e-05,
      "loss": 1.4569,
      "step": 23600
    },
    {
      "epoch": 1.819090839166218,
      "grad_norm": 2.054161787033081,
      "learning_rate": 7.981196270420487e-05,
      "loss": 1.4426,
      "step": 23650
    },
    {
      "epoch": 1.8229366971771404,
      "grad_norm": 1.6102700233459473,
      "learning_rate": 7.955224268238839e-05,
      "loss": 1.4407,
      "step": 23700
    },
    {
      "epoch": 1.8267825551880623,
      "grad_norm": 1.7302616834640503,
      "learning_rate": 7.929252266057191e-05,
      "loss": 1.4432,
      "step": 23750
    },
    {
      "epoch": 1.8267825551880623,
      "eval_loss": 1.4570631980895996,
      "eval_runtime": 17.8893,
      "eval_samples_per_second": 55.899,
      "eval_steps_per_second": 13.975,
      "step": 23750
    },
    {
      "epoch": 1.8306284131989847,
      "grad_norm": 9.521604537963867,
      "learning_rate": 7.903280263875543e-05,
      "loss": 1.3995,
      "step": 23800
    },
    {
      "epoch": 1.834474271209907,
      "grad_norm": 1.9546847343444824,
      "learning_rate": 7.877308261693894e-05,
      "loss": 1.4957,
      "step": 23850
    },
    {
      "epoch": 1.838320129220829,
      "grad_norm": 1.173722505569458,
      "learning_rate": 7.851336259512246e-05,
      "loss": 1.4157,
      "step": 23900
    },
    {
      "epoch": 1.8421659872317515,
      "grad_norm": 1.5322128534317017,
      "learning_rate": 7.825364257330599e-05,
      "loss": 1.4453,
      "step": 23950
    },
    {
      "epoch": 1.8460118452426735,
      "grad_norm": 1.0376055240631104,
      "learning_rate": 7.799392255148949e-05,
      "loss": 1.4646,
      "step": 24000
    },
    {
      "epoch": 1.8460118452426735,
      "eval_loss": 1.4612687826156616,
      "eval_runtime": 17.984,
      "eval_samples_per_second": 55.605,
      "eval_steps_per_second": 13.901,
      "step": 24000
    },
    {
      "epoch": 1.8498577032535959,
      "grad_norm": 1.5047483444213867,
      "learning_rate": 7.773420252967302e-05,
      "loss": 1.4546,
      "step": 24050
    },
    {
      "epoch": 1.853703561264518,
      "grad_norm": 1.0463405847549438,
      "learning_rate": 7.747448250785654e-05,
      "loss": 1.5014,
      "step": 24100
    },
    {
      "epoch": 1.8575494192754403,
      "grad_norm": 1.8368524312973022,
      "learning_rate": 7.721476248604004e-05,
      "loss": 1.4616,
      "step": 24150
    },
    {
      "epoch": 1.8613952772863627,
      "grad_norm": 1.4084677696228027,
      "learning_rate": 7.695504246422357e-05,
      "loss": 1.4255,
      "step": 24200
    },
    {
      "epoch": 1.8652411352972849,
      "grad_norm": 1.2279951572418213,
      "learning_rate": 7.66953224424071e-05,
      "loss": 1.4254,
      "step": 24250
    },
    {
      "epoch": 1.8652411352972849,
      "eval_loss": 1.460336685180664,
      "eval_runtime": 17.7185,
      "eval_samples_per_second": 56.438,
      "eval_steps_per_second": 14.11,
      "step": 24250
    },
    {
      "epoch": 1.869086993308207,
      "grad_norm": 1.9729641675949097,
      "learning_rate": 7.643560242059061e-05,
      "loss": 1.4656,
      "step": 24300
    },
    {
      "epoch": 1.8729328513191295,
      "grad_norm": 0.9121168255805969,
      "learning_rate": 7.617588239877412e-05,
      "loss": 1.3949,
      "step": 24350
    },
    {
      "epoch": 1.8767787093300514,
      "grad_norm": 1.8953206539154053,
      "learning_rate": 7.591616237695765e-05,
      "loss": 1.4006,
      "step": 24400
    },
    {
      "epoch": 1.8806245673409738,
      "grad_norm": 1.5828944444656372,
      "learning_rate": 7.565644235514116e-05,
      "loss": 1.5085,
      "step": 24450
    },
    {
      "epoch": 1.884470425351896,
      "grad_norm": 2.027841329574585,
      "learning_rate": 7.539672233332467e-05,
      "loss": 1.3978,
      "step": 24500
    },
    {
      "epoch": 1.884470425351896,
      "eval_loss": 1.445096731185913,
      "eval_runtime": 18.4795,
      "eval_samples_per_second": 54.114,
      "eval_steps_per_second": 13.528,
      "step": 24500
    },
    {
      "epoch": 1.8883162833628182,
      "grad_norm": 3.0785481929779053,
      "learning_rate": 7.51370023115082e-05,
      "loss": 1.493,
      "step": 24550
    },
    {
      "epoch": 1.8921621413737406,
      "grad_norm": 1.2793898582458496,
      "learning_rate": 7.487728228969173e-05,
      "loss": 1.3865,
      "step": 24600
    },
    {
      "epoch": 1.8960079993846626,
      "grad_norm": 1.474137544631958,
      "learning_rate": 7.461756226787522e-05,
      "loss": 1.4316,
      "step": 24650
    },
    {
      "epoch": 1.899853857395585,
      "grad_norm": 1.270415186882019,
      "learning_rate": 7.435784224605875e-05,
      "loss": 1.3844,
      "step": 24700
    },
    {
      "epoch": 1.9036997154065072,
      "grad_norm": 1.3681602478027344,
      "learning_rate": 7.409812222424228e-05,
      "loss": 1.4202,
      "step": 24750
    },
    {
      "epoch": 1.9036997154065072,
      "eval_loss": 1.4705748558044434,
      "eval_runtime": 18.5933,
      "eval_samples_per_second": 53.783,
      "eval_steps_per_second": 13.446,
      "step": 24750
    },
    {
      "epoch": 1.9075455734174294,
      "grad_norm": 1.028225302696228,
      "learning_rate": 7.384359660286212e-05,
      "loss": 1.4658,
      "step": 24800
    },
    {
      "epoch": 1.9113914314283518,
      "grad_norm": 1.1178765296936035,
      "learning_rate": 7.358387658104564e-05,
      "loss": 1.4616,
      "step": 24850
    },
    {
      "epoch": 1.9152372894392737,
      "grad_norm": 2.1088242530822754,
      "learning_rate": 7.332415655922915e-05,
      "loss": 1.4916,
      "step": 24900
    },
    {
      "epoch": 1.9190831474501961,
      "grad_norm": 0.7096924781799316,
      "learning_rate": 7.306443653741267e-05,
      "loss": 1.3687,
      "step": 24950
    },
    {
      "epoch": 1.9229290054611183,
      "grad_norm": 1.4435713291168213,
      "learning_rate": 7.28047165155962e-05,
      "loss": 1.5273,
      "step": 25000
    },
    {
      "epoch": 1.9229290054611183,
      "eval_loss": 1.4638206958770752,
      "eval_runtime": 18.4856,
      "eval_samples_per_second": 54.096,
      "eval_steps_per_second": 13.524,
      "step": 25000
    },
    {
      "epoch": 1.9267748634720405,
      "grad_norm": 1.4893878698349,
      "learning_rate": 7.25449964937797e-05,
      "loss": 1.5067,
      "step": 25050
    },
    {
      "epoch": 1.930620721482963,
      "grad_norm": 0.8735935091972351,
      "learning_rate": 7.228527647196322e-05,
      "loss": 1.4671,
      "step": 25100
    },
    {
      "epoch": 1.9344665794938851,
      "grad_norm": 1.6086535453796387,
      "learning_rate": 7.202555645014675e-05,
      "loss": 1.4551,
      "step": 25150
    },
    {
      "epoch": 1.9383124375048073,
      "grad_norm": 0.683675229549408,
      "learning_rate": 7.176583642833027e-05,
      "loss": 1.4673,
      "step": 25200
    },
    {
      "epoch": 1.9421582955157297,
      "grad_norm": 1.9318158626556396,
      "learning_rate": 7.150611640651378e-05,
      "loss": 1.4199,
      "step": 25250
    },
    {
      "epoch": 1.9421582955157297,
      "eval_loss": 1.4574114084243774,
      "eval_runtime": 18.5222,
      "eval_samples_per_second": 53.989,
      "eval_steps_per_second": 13.497,
      "step": 25250
    },
    {
      "epoch": 1.9460041535266517,
      "grad_norm": 1.9871971607208252,
      "learning_rate": 7.12463963846973e-05,
      "loss": 1.5002,
      "step": 25300
    },
    {
      "epoch": 1.949850011537574,
      "grad_norm": 1.4302830696105957,
      "learning_rate": 7.098667636288082e-05,
      "loss": 1.46,
      "step": 25350
    },
    {
      "epoch": 1.9536958695484963,
      "grad_norm": 1.8389050960540771,
      "learning_rate": 7.072695634106434e-05,
      "loss": 1.4025,
      "step": 25400
    },
    {
      "epoch": 1.9575417275594185,
      "grad_norm": 1.7089191675186157,
      "learning_rate": 7.046723631924785e-05,
      "loss": 1.4507,
      "step": 25450
    },
    {
      "epoch": 1.9613875855703409,
      "grad_norm": 1.3698766231536865,
      "learning_rate": 7.020751629743138e-05,
      "loss": 1.4954,
      "step": 25500
    },
    {
      "epoch": 1.9613875855703409,
      "eval_loss": 1.454710841178894,
      "eval_runtime": 18.5708,
      "eval_samples_per_second": 53.848,
      "eval_steps_per_second": 13.462,
      "step": 25500
    },
    {
      "epoch": 1.9652334435812628,
      "grad_norm": 1.808030128479004,
      "learning_rate": 6.994779627561489e-05,
      "loss": 1.4254,
      "step": 25550
    },
    {
      "epoch": 1.9690793015921852,
      "grad_norm": 1.634099006652832,
      "learning_rate": 6.96880762537984e-05,
      "loss": 1.3325,
      "step": 25600
    },
    {
      "epoch": 1.9729251596031074,
      "grad_norm": 0.8857108354568481,
      "learning_rate": 6.942835623198193e-05,
      "loss": 1.3983,
      "step": 25650
    },
    {
      "epoch": 1.9767710176140296,
      "grad_norm": 1.6114498376846313,
      "learning_rate": 6.916863621016545e-05,
      "loss": 1.3348,
      "step": 25700
    },
    {
      "epoch": 1.980616875624952,
      "grad_norm": 1.4415462017059326,
      "learning_rate": 6.890891618834895e-05,
      "loss": 1.4617,
      "step": 25750
    },
    {
      "epoch": 1.980616875624952,
      "eval_loss": 1.4457746744155884,
      "eval_runtime": 17.9567,
      "eval_samples_per_second": 55.689,
      "eval_steps_per_second": 13.922,
      "step": 25750
    },
    {
      "epoch": 1.984462733635874,
      "grad_norm": 1.9614554643630981,
      "learning_rate": 6.864919616653248e-05,
      "loss": 1.3813,
      "step": 25800
    },
    {
      "epoch": 1.9883085916467964,
      "grad_norm": 1.2938437461853027,
      "learning_rate": 6.8389476144716e-05,
      "loss": 1.4047,
      "step": 25850
    },
    {
      "epoch": 1.9921544496577186,
      "grad_norm": 2.1129326820373535,
      "learning_rate": 6.812975612289952e-05,
      "loss": 1.4362,
      "step": 25900
    },
    {
      "epoch": 1.9960003076686408,
      "grad_norm": 0.8634279370307922,
      "learning_rate": 6.787003610108303e-05,
      "loss": 1.4805,
      "step": 25950
    },
    {
      "epoch": 1.9998461656795632,
      "grad_norm": 2.995699405670166,
      "learning_rate": 6.761031607926656e-05,
      "loss": 1.5073,
      "step": 26000
    },
    {
      "epoch": 1.9998461656795632,
      "eval_loss": 1.453719973564148,
      "eval_runtime": 18.0595,
      "eval_samples_per_second": 55.373,
      "eval_steps_per_second": 13.843,
      "step": 26000
    },
    {
      "epoch": 2.003692023690485,
      "grad_norm": 1.621793508529663,
      "learning_rate": 6.73557904578864e-05,
      "loss": 1.369,
      "step": 26050
    },
    {
      "epoch": 2.0075378817014076,
      "grad_norm": 2.223520278930664,
      "learning_rate": 6.709607043606992e-05,
      "loss": 1.4269,
      "step": 26100
    },
    {
      "epoch": 2.01138373971233,
      "grad_norm": 1.4860827922821045,
      "learning_rate": 6.683635041425344e-05,
      "loss": 1.3634,
      "step": 26150
    },
    {
      "epoch": 2.015229597723252,
      "grad_norm": 2.0796148777008057,
      "learning_rate": 6.657663039243696e-05,
      "loss": 1.4233,
      "step": 26200
    },
    {
      "epoch": 2.0190754557341744,
      "grad_norm": 1.6398444175720215,
      "learning_rate": 6.631691037062047e-05,
      "loss": 1.4058,
      "step": 26250
    },
    {
      "epoch": 2.0190754557341744,
      "eval_loss": 1.4733901023864746,
      "eval_runtime": 18.0349,
      "eval_samples_per_second": 55.448,
      "eval_steps_per_second": 13.862,
      "step": 26250
    },
    {
      "epoch": 2.0229213137450963,
      "grad_norm": 1.7550077438354492,
      "learning_rate": 6.605719034880399e-05,
      "loss": 1.4436,
      "step": 26300
    },
    {
      "epoch": 2.0267671717560187,
      "grad_norm": 2.3273561000823975,
      "learning_rate": 6.579747032698751e-05,
      "loss": 1.356,
      "step": 26350
    },
    {
      "epoch": 2.030613029766941,
      "grad_norm": 1.1432509422302246,
      "learning_rate": 6.553775030517103e-05,
      "loss": 1.4116,
      "step": 26400
    },
    {
      "epoch": 2.034458887777863,
      "grad_norm": 1.2345376014709473,
      "learning_rate": 6.527803028335455e-05,
      "loss": 1.4465,
      "step": 26450
    },
    {
      "epoch": 2.0383047457887855,
      "grad_norm": 1.485564112663269,
      "learning_rate": 6.501831026153807e-05,
      "loss": 1.3896,
      "step": 26500
    },
    {
      "epoch": 2.0383047457887855,
      "eval_loss": 1.4492217302322388,
      "eval_runtime": 17.9114,
      "eval_samples_per_second": 55.83,
      "eval_steps_per_second": 13.958,
      "step": 26500
    },
    {
      "epoch": 2.042150603799708,
      "grad_norm": 1.0810669660568237,
      "learning_rate": 6.475859023972158e-05,
      "loss": 1.2936,
      "step": 26550
    },
    {
      "epoch": 2.04599646181063,
      "grad_norm": 1.23382568359375,
      "learning_rate": 6.44988702179051e-05,
      "loss": 1.3922,
      "step": 26600
    },
    {
      "epoch": 2.0498423198215523,
      "grad_norm": 1.8218950033187866,
      "learning_rate": 6.423915019608862e-05,
      "loss": 1.4041,
      "step": 26650
    },
    {
      "epoch": 2.0536881778324743,
      "grad_norm": 0.6482899785041809,
      "learning_rate": 6.397943017427213e-05,
      "loss": 1.3609,
      "step": 26700
    },
    {
      "epoch": 2.0575340358433967,
      "grad_norm": 1.9538156986236572,
      "learning_rate": 6.371971015245566e-05,
      "loss": 1.4558,
      "step": 26750
    },
    {
      "epoch": 2.0575340358433967,
      "eval_loss": 1.454428791999817,
      "eval_runtime": 18.2441,
      "eval_samples_per_second": 54.812,
      "eval_steps_per_second": 13.703,
      "step": 26750
    },
    {
      "epoch": 2.061379893854319,
      "grad_norm": 1.087234377861023,
      "learning_rate": 6.345999013063917e-05,
      "loss": 1.3967,
      "step": 26800
    },
    {
      "epoch": 2.065225751865241,
      "grad_norm": 1.638533592224121,
      "learning_rate": 6.32002701088227e-05,
      "loss": 1.4558,
      "step": 26850
    },
    {
      "epoch": 2.0690716098761635,
      "grad_norm": 1.4552900791168213,
      "learning_rate": 6.294055008700621e-05,
      "loss": 1.4553,
      "step": 26900
    },
    {
      "epoch": 2.0729174678870854,
      "grad_norm": 1.6330054998397827,
      "learning_rate": 6.268083006518974e-05,
      "loss": 1.4977,
      "step": 26950
    },
    {
      "epoch": 2.076763325898008,
      "grad_norm": 1.497938632965088,
      "learning_rate": 6.242111004337325e-05,
      "loss": 1.4449,
      "step": 27000
    },
    {
      "epoch": 2.076763325898008,
      "eval_loss": 1.4610888957977295,
      "eval_runtime": 17.9657,
      "eval_samples_per_second": 55.662,
      "eval_steps_per_second": 13.915,
      "step": 27000
    },
    {
      "epoch": 2.0806091839089302,
      "grad_norm": 0.9695401191711426,
      "learning_rate": 6.216139002155676e-05,
      "loss": 1.4943,
      "step": 27050
    },
    {
      "epoch": 2.084455041919852,
      "grad_norm": 1.243717074394226,
      "learning_rate": 6.190166999974029e-05,
      "loss": 1.4199,
      "step": 27100
    },
    {
      "epoch": 2.0883008999307746,
      "grad_norm": 1.6151024103164673,
      "learning_rate": 6.16419499779238e-05,
      "loss": 1.4248,
      "step": 27150
    },
    {
      "epoch": 2.0921467579416966,
      "grad_norm": 1.7448607683181763,
      "learning_rate": 6.138222995610731e-05,
      "loss": 1.3944,
      "step": 27200
    },
    {
      "epoch": 2.095992615952619,
      "grad_norm": 3.0997345447540283,
      "learning_rate": 6.112250993429084e-05,
      "loss": 1.4174,
      "step": 27250
    },
    {
      "epoch": 2.095992615952619,
      "eval_loss": 1.4659229516983032,
      "eval_runtime": 18.0653,
      "eval_samples_per_second": 55.355,
      "eval_steps_per_second": 13.839,
      "step": 27250
    },
    {
      "epoch": 2.0998384739635414,
      "grad_norm": 1.637845754623413,
      "learning_rate": 6.086278991247436e-05,
      "loss": 1.446,
      "step": 27300
    },
    {
      "epoch": 2.1036843319744634,
      "grad_norm": 1.4263664484024048,
      "learning_rate": 6.060306989065787e-05,
      "loss": 0.9793,
      "step": 27350
    },
    {
      "epoch": 2.1075301899853858,
      "grad_norm": 0.8709418773651123,
      "learning_rate": 6.034334986884139e-05,
      "loss": 0.9715,
      "step": 27400
    },
    {
      "epoch": 2.111376047996308,
      "grad_norm": 0.8483341336250305,
      "learning_rate": 6.008362984702491e-05,
      "loss": 0.9163,
      "step": 27450
    },
    {
      "epoch": 2.11522190600723,
      "grad_norm": 2.070937156677246,
      "learning_rate": 5.9823909825208425e-05,
      "loss": 1.0047,
      "step": 27500
    },
    {
      "epoch": 2.11522190600723,
      "eval_loss": 1.0159448385238647,
      "eval_runtime": 17.7021,
      "eval_samples_per_second": 56.49,
      "eval_steps_per_second": 14.123,
      "step": 27500
    },
    {
      "epoch": 2.1190677640181526,
      "grad_norm": 0.9638277292251587,
      "learning_rate": 5.9564189803391944e-05,
      "loss": 0.9753,
      "step": 27550
    },
    {
      "epoch": 2.1229136220290745,
      "grad_norm": 1.1322181224822998,
      "learning_rate": 5.9304469781575464e-05,
      "loss": 0.948,
      "step": 27600
    },
    {
      "epoch": 2.126759480039997,
      "grad_norm": 1.144047737121582,
      "learning_rate": 5.904474975975898e-05,
      "loss": 1.009,
      "step": 27650
    },
    {
      "epoch": 2.1306053380509193,
      "grad_norm": 2.713625431060791,
      "learning_rate": 5.8785029737942496e-05,
      "loss": 1.0069,
      "step": 27700
    },
    {
      "epoch": 2.1344511960618413,
      "grad_norm": 1.1035822629928589,
      "learning_rate": 5.852530971612602e-05,
      "loss": 1.0013,
      "step": 27750
    },
    {
      "epoch": 2.1344511960618413,
      "eval_loss": 1.0144418478012085,
      "eval_runtime": 17.6881,
      "eval_samples_per_second": 56.535,
      "eval_steps_per_second": 14.134,
      "step": 27750
    },
    {
      "epoch": 2.1382970540727637,
      "grad_norm": 1.2658100128173828,
      "learning_rate": 5.826558969430954e-05,
      "loss": 1.0185,
      "step": 27800
    },
    {
      "epoch": 2.1421429120836857,
      "grad_norm": 0.9421238303184509,
      "learning_rate": 5.800586967249305e-05,
      "loss": 0.9892,
      "step": 27850
    },
    {
      "epoch": 2.145988770094608,
      "grad_norm": 0.9409565925598145,
      "learning_rate": 5.7746149650676575e-05,
      "loss": 0.9985,
      "step": 27900
    },
    {
      "epoch": 2.1498346281055305,
      "grad_norm": 1.445890188217163,
      "learning_rate": 5.7486429628860094e-05,
      "loss": 1.0147,
      "step": 27950
    },
    {
      "epoch": 2.1536804861164525,
      "grad_norm": 1.109020709991455,
      "learning_rate": 5.722670960704361e-05,
      "loss": 1.0093,
      "step": 28000
    },
    {
      "epoch": 2.1536804861164525,
      "eval_loss": 1.008616328239441,
      "eval_runtime": 17.6489,
      "eval_samples_per_second": 56.661,
      "eval_steps_per_second": 14.165,
      "step": 28000
    },
    {
      "epoch": 2.157526344127375,
      "grad_norm": 1.3012685775756836,
      "learning_rate": 5.696698958522713e-05,
      "loss": 1.0703,
      "step": 28050
    },
    {
      "epoch": 2.1613722021382973,
      "grad_norm": 0.8459142446517944,
      "learning_rate": 5.6707269563410646e-05,
      "loss": 1.0507,
      "step": 28100
    },
    {
      "epoch": 2.1652180601492192,
      "grad_norm": 0.6889505982398987,
      "learning_rate": 5.6447549541594166e-05,
      "loss": 1.0056,
      "step": 28150
    },
    {
      "epoch": 2.1690639181601417,
      "grad_norm": 1.2236456871032715,
      "learning_rate": 5.618782951977768e-05,
      "loss": 0.988,
      "step": 28200
    },
    {
      "epoch": 2.1729097761710636,
      "grad_norm": 1.3419203758239746,
      "learning_rate": 5.59281094979612e-05,
      "loss": 0.9467,
      "step": 28250
    },
    {
      "epoch": 2.1729097761710636,
      "eval_loss": 1.0140153169631958,
      "eval_runtime": 17.7299,
      "eval_samples_per_second": 56.402,
      "eval_steps_per_second": 14.1,
      "step": 28250
    },
    {
      "epoch": 2.176755634181986,
      "grad_norm": 0.8394871354103088,
      "learning_rate": 5.5668389476144725e-05,
      "loss": 0.9824,
      "step": 28300
    },
    {
      "epoch": 2.1806014921929084,
      "grad_norm": 0.7906908392906189,
      "learning_rate": 5.540866945432823e-05,
      "loss": 0.9915,
      "step": 28350
    },
    {
      "epoch": 2.1844473502038304,
      "grad_norm": 1.0852785110473633,
      "learning_rate": 5.514894943251175e-05,
      "loss": 0.9473,
      "step": 28400
    },
    {
      "epoch": 2.188293208214753,
      "grad_norm": 0.9187583923339844,
      "learning_rate": 5.488922941069528e-05,
      "loss": 0.9418,
      "step": 28450
    },
    {
      "epoch": 2.192139066225675,
      "grad_norm": 0.7200838923454285,
      "learning_rate": 5.4629509388878797e-05,
      "loss": 0.9776,
      "step": 28500
    },
    {
      "epoch": 2.192139066225675,
      "eval_loss": 1.0052642822265625,
      "eval_runtime": 17.5069,
      "eval_samples_per_second": 57.12,
      "eval_steps_per_second": 14.28,
      "step": 28500
    },
    {
      "epoch": 2.195984924236597,
      "grad_norm": 1.00751793384552,
      "learning_rate": 5.436978936706231e-05,
      "loss": 0.9479,
      "step": 28550
    },
    {
      "epoch": 2.1998307822475196,
      "grad_norm": 0.9239784479141235,
      "learning_rate": 5.411006934524583e-05,
      "loss": 0.9896,
      "step": 28600
    },
    {
      "epoch": 2.2036766402584416,
      "grad_norm": 0.9650816917419434,
      "learning_rate": 5.385034932342935e-05,
      "loss": 1.0081,
      "step": 28650
    },
    {
      "epoch": 2.207522498269364,
      "grad_norm": 1.455723524093628,
      "learning_rate": 5.359062930161286e-05,
      "loss": 0.9717,
      "step": 28700
    },
    {
      "epoch": 2.211368356280286,
      "grad_norm": 0.609380304813385,
      "learning_rate": 5.333090927979638e-05,
      "loss": 0.9774,
      "step": 28750
    },
    {
      "epoch": 2.211368356280286,
      "eval_loss": 1.0018259286880493,
      "eval_runtime": 17.3475,
      "eval_samples_per_second": 57.645,
      "eval_steps_per_second": 14.411,
      "step": 28750
    },
    {
      "epoch": 2.2152142142912084,
      "grad_norm": 0.9981700778007507,
      "learning_rate": 5.30711892579799e-05,
      "loss": 0.9873,
      "step": 28800
    },
    {
      "epoch": 2.2190600723021308,
      "grad_norm": 0.5452422499656677,
      "learning_rate": 5.2811469236163414e-05,
      "loss": 1.0078,
      "step": 28850
    },
    {
      "epoch": 2.2229059303130527,
      "grad_norm": 1.4739840030670166,
      "learning_rate": 5.255174921434693e-05,
      "loss": 0.9739,
      "step": 28900
    },
    {
      "epoch": 2.226751788323975,
      "grad_norm": 0.9250359535217285,
      "learning_rate": 5.229202919253045e-05,
      "loss": 0.9723,
      "step": 28950
    },
    {
      "epoch": 2.230597646334897,
      "grad_norm": 1.0843122005462646,
      "learning_rate": 5.203230917071398e-05,
      "loss": 0.989,
      "step": 29000
    },
    {
      "epoch": 2.230597646334897,
      "eval_loss": 0.9999537467956543,
      "eval_runtime": 17.2417,
      "eval_samples_per_second": 57.999,
      "eval_steps_per_second": 14.5,
      "step": 29000
    },
    {
      "epoch": 2.2344435043458195,
      "grad_norm": 0.5559306740760803,
      "learning_rate": 5.1772589148897485e-05,
      "loss": 0.9749,
      "step": 29050
    },
    {
      "epoch": 2.238289362356742,
      "grad_norm": 0.6720598936080933,
      "learning_rate": 5.151286912708101e-05,
      "loss": 0.9662,
      "step": 29100
    },
    {
      "epoch": 2.242135220367664,
      "grad_norm": 1.117200493812561,
      "learning_rate": 5.125314910526453e-05,
      "loss": 0.9552,
      "step": 29150
    },
    {
      "epoch": 2.2459810783785863,
      "grad_norm": 0.6818645000457764,
      "learning_rate": 5.0993429083448044e-05,
      "loss": 0.9032,
      "step": 29200
    },
    {
      "epoch": 2.2498269363895087,
      "grad_norm": 0.9796412587165833,
      "learning_rate": 5.0733709061631564e-05,
      "loss": 0.9429,
      "step": 29250
    },
    {
      "epoch": 2.2498269363895087,
      "eval_loss": 1.0121312141418457,
      "eval_runtime": 17.3211,
      "eval_samples_per_second": 57.733,
      "eval_steps_per_second": 14.433,
      "step": 29250
    },
    {
      "epoch": 2.2536727944004307,
      "grad_norm": 1.021713137626648,
      "learning_rate": 5.047398903981508e-05,
      "loss": 0.979,
      "step": 29300
    },
    {
      "epoch": 2.257518652411353,
      "grad_norm": 1.1321250200271606,
      "learning_rate": 5.02142690179986e-05,
      "loss": 1.0327,
      "step": 29350
    },
    {
      "epoch": 2.261364510422275,
      "grad_norm": 0.7670277953147888,
      "learning_rate": 4.9954548996182116e-05,
      "loss": 0.9668,
      "step": 29400
    },
    {
      "epoch": 2.2652103684331975,
      "grad_norm": 1.447698712348938,
      "learning_rate": 4.9694828974365635e-05,
      "loss": 0.9288,
      "step": 29450
    },
    {
      "epoch": 2.26905622644412,
      "grad_norm": 1.0438776016235352,
      "learning_rate": 4.9435108952549155e-05,
      "loss": 1.0154,
      "step": 29500
    },
    {
      "epoch": 2.26905622644412,
      "eval_loss": 1.0100510120391846,
      "eval_runtime": 17.4293,
      "eval_samples_per_second": 57.375,
      "eval_steps_per_second": 14.344,
      "step": 29500
    },
    {
      "epoch": 2.272902084455042,
      "grad_norm": 0.6814424991607666,
      "learning_rate": 4.9175388930732675e-05,
      "loss": 0.9563,
      "step": 29550
    },
    {
      "epoch": 2.2767479424659642,
      "grad_norm": 0.85086989402771,
      "learning_rate": 4.891566890891619e-05,
      "loss": 0.9778,
      "step": 29600
    },
    {
      "epoch": 2.280593800476886,
      "grad_norm": 1.1364527940750122,
      "learning_rate": 4.8655948887099714e-05,
      "loss": 0.996,
      "step": 29650
    },
    {
      "epoch": 2.2844396584878086,
      "grad_norm": 0.7810873985290527,
      "learning_rate": 4.839622886528323e-05,
      "loss": 1.0141,
      "step": 29700
    },
    {
      "epoch": 2.288285516498731,
      "grad_norm": 1.02638578414917,
      "learning_rate": 4.8136508843466746e-05,
      "loss": 0.97,
      "step": 29750
    },
    {
      "epoch": 2.288285516498731,
      "eval_loss": 1.0048705339431763,
      "eval_runtime": 17.487,
      "eval_samples_per_second": 57.185,
      "eval_steps_per_second": 14.296,
      "step": 29750
    },
    {
      "epoch": 2.292131374509653,
      "grad_norm": 1.334876298904419,
      "learning_rate": 4.7876788821650266e-05,
      "loss": 0.9533,
      "step": 29800
    },
    {
      "epoch": 2.2959772325205754,
      "grad_norm": 1.3690305948257446,
      "learning_rate": 4.761706879983378e-05,
      "loss": 0.992,
      "step": 29850
    },
    {
      "epoch": 2.299823090531498,
      "grad_norm": 1.525981068611145,
      "learning_rate": 4.7357348778017305e-05,
      "loss": 0.949,
      "step": 29900
    },
    {
      "epoch": 2.3036689485424198,
      "grad_norm": 0.6612236499786377,
      "learning_rate": 4.709762875620082e-05,
      "loss": 1.0112,
      "step": 29950
    },
    {
      "epoch": 2.307514806553342,
      "grad_norm": 0.8452871441841125,
      "learning_rate": 4.683790873438433e-05,
      "loss": 0.9858,
      "step": 30000
    },
    {
      "epoch": 2.307514806553342,
      "eval_loss": 0.9945911169052124,
      "eval_runtime": 17.2501,
      "eval_samples_per_second": 57.971,
      "eval_steps_per_second": 14.493,
      "step": 30000
    },
    {
      "epoch": 2.311360664564264,
      "grad_norm": 1.3450604677200317,
      "learning_rate": 4.657818871256786e-05,
      "loss": 0.973,
      "step": 30050
    },
    {
      "epoch": 2.3152065225751866,
      "grad_norm": 1.1462957859039307,
      "learning_rate": 4.631846869075137e-05,
      "loss": 0.9789,
      "step": 30100
    },
    {
      "epoch": 2.3190523805861085,
      "grad_norm": 1.0691301822662354,
      "learning_rate": 4.605874866893489e-05,
      "loss": 0.9937,
      "step": 30150
    },
    {
      "epoch": 2.322898238597031,
      "grad_norm": 1.527723789215088,
      "learning_rate": 4.579902864711841e-05,
      "loss": 0.9276,
      "step": 30200
    },
    {
      "epoch": 2.3267440966079533,
      "grad_norm": 0.7396986484527588,
      "learning_rate": 4.553930862530192e-05,
      "loss": 0.9648,
      "step": 30250
    },
    {
      "epoch": 2.3267440966079533,
      "eval_loss": 1.0029717683792114,
      "eval_runtime": 17.2662,
      "eval_samples_per_second": 57.917,
      "eval_steps_per_second": 14.479,
      "step": 30250
    },
    {
      "epoch": 2.3305899546188753,
      "grad_norm": 0.9694539308547974,
      "learning_rate": 4.527958860348545e-05,
      "loss": 0.9378,
      "step": 30300
    },
    {
      "epoch": 2.3344358126297977,
      "grad_norm": 1.161028504371643,
      "learning_rate": 4.501986858166896e-05,
      "loss": 0.9807,
      "step": 30350
    },
    {
      "epoch": 2.33828167064072,
      "grad_norm": 1.6975845098495483,
      "learning_rate": 4.476014855985248e-05,
      "loss": 0.9809,
      "step": 30400
    },
    {
      "epoch": 2.342127528651642,
      "grad_norm": 1.3228962421417236,
      "learning_rate": 4.4500428538036e-05,
      "loss": 1.0339,
      "step": 30450
    },
    {
      "epoch": 2.3459733866625645,
      "grad_norm": 1.0168397426605225,
      "learning_rate": 4.424070851621952e-05,
      "loss": 0.9544,
      "step": 30500
    },
    {
      "epoch": 2.3459733866625645,
      "eval_loss": 1.002646565437317,
      "eval_runtime": 17.3363,
      "eval_samples_per_second": 57.682,
      "eval_steps_per_second": 14.421,
      "step": 30500
    },
    {
      "epoch": 2.3498192446734865,
      "grad_norm": 0.9951680302619934,
      "learning_rate": 4.398098849440303e-05,
      "loss": 1.0062,
      "step": 30550
    },
    {
      "epoch": 2.353665102684409,
      "grad_norm": 0.5752933025360107,
      "learning_rate": 4.372126847258655e-05,
      "loss": 0.9649,
      "step": 30600
    },
    {
      "epoch": 2.3575109606953313,
      "grad_norm": 1.0051320791244507,
      "learning_rate": 4.346154845077007e-05,
      "loss": 0.9958,
      "step": 30650
    },
    {
      "epoch": 2.3613568187062532,
      "grad_norm": 0.7760717868804932,
      "learning_rate": 4.320182842895359e-05,
      "loss": 1.0125,
      "step": 30700
    },
    {
      "epoch": 2.3652026767171757,
      "grad_norm": 0.852301836013794,
      "learning_rate": 4.294210840713711e-05,
      "loss": 0.9523,
      "step": 30750
    },
    {
      "epoch": 2.3652026767171757,
      "eval_loss": 1.003655195236206,
      "eval_runtime": 17.3241,
      "eval_samples_per_second": 57.723,
      "eval_steps_per_second": 14.431,
      "step": 30750
    },
    {
      "epoch": 2.3690485347280976,
      "grad_norm": 0.9062100648880005,
      "learning_rate": 4.2682388385320624e-05,
      "loss": 0.9735,
      "step": 30800
    },
    {
      "epoch": 2.37289439273902,
      "grad_norm": 1.309615969657898,
      "learning_rate": 4.2427862763940476e-05,
      "loss": 0.9668,
      "step": 30850
    },
    {
      "epoch": 2.3767402507499424,
      "grad_norm": 1.0907591581344604,
      "learning_rate": 4.2168142742123995e-05,
      "loss": 0.9502,
      "step": 30900
    },
    {
      "epoch": 2.3805861087608644,
      "grad_norm": 1.0946288108825684,
      "learning_rate": 4.190842272030751e-05,
      "loss": 0.9545,
      "step": 30950
    },
    {
      "epoch": 2.384431966771787,
      "grad_norm": 1.225540280342102,
      "learning_rate": 4.164870269849103e-05,
      "loss": 0.9635,
      "step": 31000
    },
    {
      "epoch": 2.384431966771787,
      "eval_loss": 1.0031681060791016,
      "eval_runtime": 17.2062,
      "eval_samples_per_second": 58.119,
      "eval_steps_per_second": 14.53,
      "step": 31000
    },
    {
      "epoch": 2.3882778247827092,
      "grad_norm": 0.8717153668403625,
      "learning_rate": 4.138898267667455e-05,
      "loss": 1.0038,
      "step": 31050
    },
    {
      "epoch": 2.392123682793631,
      "grad_norm": 1.007270097732544,
      "learning_rate": 4.112926265485807e-05,
      "loss": 0.9858,
      "step": 31100
    },
    {
      "epoch": 2.3959695408045536,
      "grad_norm": 1.9409807920455933,
      "learning_rate": 4.0869542633041587e-05,
      "loss": 1.004,
      "step": 31150
    },
    {
      "epoch": 2.3998153988154756,
      "grad_norm": 0.6027572154998779,
      "learning_rate": 4.06098226112251e-05,
      "loss": 0.9958,
      "step": 31200
    },
    {
      "epoch": 2.403661256826398,
      "grad_norm": 0.8274515867233276,
      "learning_rate": 4.035010258940862e-05,
      "loss": 0.9852,
      "step": 31250
    },
    {
      "epoch": 2.403661256826398,
      "eval_loss": 1.0001976490020752,
      "eval_runtime": 17.2233,
      "eval_samples_per_second": 58.061,
      "eval_steps_per_second": 14.515,
      "step": 31250
    },
    {
      "epoch": 2.4075071148373204,
      "grad_norm": 0.7238942384719849,
      "learning_rate": 4.009038256759214e-05,
      "loss": 1.0062,
      "step": 31300
    },
    {
      "epoch": 2.4113529728482423,
      "grad_norm": 0.8912849426269531,
      "learning_rate": 3.983066254577565e-05,
      "loss": 0.9802,
      "step": 31350
    },
    {
      "epoch": 2.4151988308591648,
      "grad_norm": 1.1922829151153564,
      "learning_rate": 3.957094252395918e-05,
      "loss": 1.0317,
      "step": 31400
    },
    {
      "epoch": 2.4190446888700867,
      "grad_norm": 1.3773999214172363,
      "learning_rate": 3.931122250214269e-05,
      "loss": 0.9773,
      "step": 31450
    },
    {
      "epoch": 2.422890546881009,
      "grad_norm": 1.0747745037078857,
      "learning_rate": 3.905150248032621e-05,
      "loss": 1.0208,
      "step": 31500
    },
    {
      "epoch": 2.422890546881009,
      "eval_loss": 1.0009056329727173,
      "eval_runtime": 17.1069,
      "eval_samples_per_second": 58.456,
      "eval_steps_per_second": 14.614,
      "step": 31500
    },
    {
      "epoch": 2.4267364048919315,
      "grad_norm": 1.3409994840621948,
      "learning_rate": 3.879178245850973e-05,
      "loss": 0.9546,
      "step": 31550
    },
    {
      "epoch": 2.4305822629028535,
      "grad_norm": 1.482633352279663,
      "learning_rate": 3.853206243669324e-05,
      "loss": 1.0155,
      "step": 31600
    },
    {
      "epoch": 2.434428120913776,
      "grad_norm": 1.0458152294158936,
      "learning_rate": 3.827234241487677e-05,
      "loss": 0.989,
      "step": 31650
    },
    {
      "epoch": 2.4382739789246983,
      "grad_norm": 0.9805555939674377,
      "learning_rate": 3.801262239306028e-05,
      "loss": 0.9125,
      "step": 31700
    },
    {
      "epoch": 2.4421198369356203,
      "grad_norm": 1.5680670738220215,
      "learning_rate": 3.77529023712438e-05,
      "loss": 0.9713,
      "step": 31750
    },
    {
      "epoch": 2.4421198369356203,
      "eval_loss": 0.9981379508972168,
      "eval_runtime": 17.3284,
      "eval_samples_per_second": 57.709,
      "eval_steps_per_second": 14.427,
      "step": 31750
    },
    {
      "epoch": 2.4459656949465427,
      "grad_norm": 1.0316798686981201,
      "learning_rate": 3.749318234942732e-05,
      "loss": 0.9689,
      "step": 31800
    },
    {
      "epoch": 2.4498115529574647,
      "grad_norm": 0.6630721092224121,
      "learning_rate": 3.7233462327610834e-05,
      "loss": 0.9772,
      "step": 31850
    },
    {
      "epoch": 2.453657410968387,
      "grad_norm": 1.1662702560424805,
      "learning_rate": 3.6973742305794354e-05,
      "loss": 1.0165,
      "step": 31900
    },
    {
      "epoch": 2.457503268979309,
      "grad_norm": 1.0451244115829468,
      "learning_rate": 3.671402228397787e-05,
      "loss": 1.0233,
      "step": 31950
    },
    {
      "epoch": 2.4613491269902315,
      "grad_norm": 0.9077771306037903,
      "learning_rate": 3.645430226216139e-05,
      "loss": 0.9797,
      "step": 32000
    },
    {
      "epoch": 2.4613491269902315,
      "eval_loss": 0.9961766004562378,
      "eval_runtime": 17.2799,
      "eval_samples_per_second": 57.871,
      "eval_steps_per_second": 14.468,
      "step": 32000
    },
    {
      "epoch": 2.465194985001154,
      "grad_norm": 0.999718964099884,
      "learning_rate": 3.619458224034491e-05,
      "loss": 0.9725,
      "step": 32050
    },
    {
      "epoch": 2.469040843012076,
      "grad_norm": 2.740297794342041,
      "learning_rate": 3.5934862218528425e-05,
      "loss": 0.9838,
      "step": 32100
    },
    {
      "epoch": 2.4728867010229982,
      "grad_norm": 1.199425458908081,
      "learning_rate": 3.5675142196711945e-05,
      "loss": 0.9807,
      "step": 32150
    },
    {
      "epoch": 2.4767325590339206,
      "grad_norm": 0.7113758325576782,
      "learning_rate": 3.5415422174895465e-05,
      "loss": 1.0046,
      "step": 32200
    },
    {
      "epoch": 2.4805784170448426,
      "grad_norm": 0.9929390549659729,
      "learning_rate": 3.5155702153078984e-05,
      "loss": 0.9692,
      "step": 32250
    },
    {
      "epoch": 2.4805784170448426,
      "eval_loss": 1.002519965171814,
      "eval_runtime": 17.3042,
      "eval_samples_per_second": 57.789,
      "eval_steps_per_second": 14.447,
      "step": 32250
    },
    {
      "epoch": 2.484424275055765,
      "grad_norm": 0.8170703649520874,
      "learning_rate": 3.48959821312625e-05,
      "loss": 0.982,
      "step": 32300
    },
    {
      "epoch": 2.488270133066687,
      "grad_norm": 0.8909692168235779,
      "learning_rate": 3.463626210944602e-05,
      "loss": 0.9953,
      "step": 32350
    },
    {
      "epoch": 2.4921159910776094,
      "grad_norm": 1.806539535522461,
      "learning_rate": 3.4376542087629536e-05,
      "loss": 1.0069,
      "step": 32400
    },
    {
      "epoch": 2.495961849088532,
      "grad_norm": 1.4509518146514893,
      "learning_rate": 3.4116822065813056e-05,
      "loss": 0.9642,
      "step": 32450
    },
    {
      "epoch": 2.4998077070994538,
      "grad_norm": 1.0312175750732422,
      "learning_rate": 3.3857102043996575e-05,
      "loss": 0.9921,
      "step": 32500
    },
    {
      "epoch": 2.4998077070994538,
      "eval_loss": 0.9993879199028015,
      "eval_runtime": 17.3389,
      "eval_samples_per_second": 57.674,
      "eval_steps_per_second": 14.418,
      "step": 32500
    },
    {
      "epoch": 2.503653565110376,
      "grad_norm": 0.867513120174408,
      "learning_rate": 3.359738202218009e-05,
      "loss": 0.994,
      "step": 32550
    },
    {
      "epoch": 2.507499423121298,
      "grad_norm": 1.7425885200500488,
      "learning_rate": 3.3337662000363615e-05,
      "loss": 1.0068,
      "step": 32600
    },
    {
      "epoch": 2.5113452811322206,
      "grad_norm": 0.9053608775138855,
      "learning_rate": 3.307794197854713e-05,
      "loss": 0.9704,
      "step": 32650
    },
    {
      "epoch": 2.515191139143143,
      "grad_norm": 1.0533051490783691,
      "learning_rate": 3.281822195673064e-05,
      "loss": 0.9506,
      "step": 32700
    },
    {
      "epoch": 2.519036997154065,
      "grad_norm": 1.2495230436325073,
      "learning_rate": 3.255850193491417e-05,
      "loss": 0.9936,
      "step": 32750
    },
    {
      "epoch": 2.519036997154065,
      "eval_loss": 0.9990929961204529,
      "eval_runtime": 17.239,
      "eval_samples_per_second": 58.008,
      "eval_steps_per_second": 14.502,
      "step": 32750
    },
    {
      "epoch": 2.5228828551649873,
      "grad_norm": 0.8339760303497314,
      "learning_rate": 3.229878191309768e-05,
      "loss": 1.0008,
      "step": 32800
    },
    {
      "epoch": 2.5267287131759097,
      "grad_norm": 1.2839399576187134,
      "learning_rate": 3.20390618912812e-05,
      "loss": 1.0109,
      "step": 32850
    },
    {
      "epoch": 2.5305745711868317,
      "grad_norm": 1.2146100997924805,
      "learning_rate": 3.177934186946472e-05,
      "loss": 0.9834,
      "step": 32900
    },
    {
      "epoch": 2.534420429197754,
      "grad_norm": 0.7952923774719238,
      "learning_rate": 3.151962184764823e-05,
      "loss": 1.0013,
      "step": 32950
    },
    {
      "epoch": 2.5382662872086765,
      "grad_norm": 1.67001211643219,
      "learning_rate": 3.125990182583176e-05,
      "loss": 0.9797,
      "step": 33000
    },
    {
      "epoch": 2.5382662872086765,
      "eval_loss": 1.001037359237671,
      "eval_runtime": 17.347,
      "eval_samples_per_second": 57.647,
      "eval_steps_per_second": 14.412,
      "step": 33000
    },
    {
      "epoch": 2.5421121452195985,
      "grad_norm": 0.7536977529525757,
      "learning_rate": 3.100018180401527e-05,
      "loss": 0.9818,
      "step": 33050
    },
    {
      "epoch": 2.5459580032305205,
      "grad_norm": 1.3777302503585815,
      "learning_rate": 3.074046178219879e-05,
      "loss": 0.953,
      "step": 33100
    },
    {
      "epoch": 2.549803861241443,
      "grad_norm": 0.9286239743232727,
      "learning_rate": 3.048074176038231e-05,
      "loss": 1.028,
      "step": 33150
    },
    {
      "epoch": 2.5536497192523653,
      "grad_norm": 0.6741893291473389,
      "learning_rate": 3.022102173856583e-05,
      "loss": 0.9693,
      "step": 33200
    },
    {
      "epoch": 2.5574955772632872,
      "grad_norm": 1.3020586967468262,
      "learning_rate": 2.9961301716749346e-05,
      "loss": 0.9673,
      "step": 33250
    },
    {
      "epoch": 2.5574955772632872,
      "eval_loss": 1.0041394233703613,
      "eval_runtime": 17.32,
      "eval_samples_per_second": 57.737,
      "eval_steps_per_second": 14.434,
      "step": 33250
    },
    {
      "epoch": 2.5613414352742097,
      "grad_norm": 1.4816234111785889,
      "learning_rate": 2.9701581694932862e-05,
      "loss": 0.9964,
      "step": 33300
    },
    {
      "epoch": 2.565187293285132,
      "grad_norm": 0.8813285827636719,
      "learning_rate": 2.9441861673116382e-05,
      "loss": 0.966,
      "step": 33350
    },
    {
      "epoch": 2.569033151296054,
      "grad_norm": 1.0042293071746826,
      "learning_rate": 2.9182141651299898e-05,
      "loss": 0.9254,
      "step": 33400
    },
    {
      "epoch": 2.5728790093069764,
      "grad_norm": 1.5775707960128784,
      "learning_rate": 2.892242162948342e-05,
      "loss": 1.0215,
      "step": 33450
    },
    {
      "epoch": 2.576724867317899,
      "grad_norm": 1.0266311168670654,
      "learning_rate": 2.8662701607666937e-05,
      "loss": 0.9323,
      "step": 33500
    },
    {
      "epoch": 2.576724867317899,
      "eval_loss": 1.0002070665359497,
      "eval_runtime": 17.2317,
      "eval_samples_per_second": 58.033,
      "eval_steps_per_second": 14.508,
      "step": 33500
    },
    {
      "epoch": 2.580570725328821,
      "grad_norm": 0.5430648922920227,
      "learning_rate": 2.8402981585850453e-05,
      "loss": 0.9658,
      "step": 33550
    },
    {
      "epoch": 2.5844165833397432,
      "grad_norm": 0.6678454279899597,
      "learning_rate": 2.8143261564033973e-05,
      "loss": 1.0292,
      "step": 33600
    },
    {
      "epoch": 2.588262441350665,
      "grad_norm": 0.7208724021911621,
      "learning_rate": 2.788354154221749e-05,
      "loss": 0.9505,
      "step": 33650
    },
    {
      "epoch": 2.5921082993615876,
      "grad_norm": 1.2248526811599731,
      "learning_rate": 2.7623821520401012e-05,
      "loss": 0.9728,
      "step": 33700
    },
    {
      "epoch": 2.5959541573725096,
      "grad_norm": 1.0026588439941406,
      "learning_rate": 2.7364101498584525e-05,
      "loss": 0.9783,
      "step": 33750
    },
    {
      "epoch": 2.5959541573725096,
      "eval_loss": 0.9881900548934937,
      "eval_runtime": 17.2742,
      "eval_samples_per_second": 57.89,
      "eval_steps_per_second": 14.472,
      "step": 33750
    },
    {
      "epoch": 2.599800015383432,
      "grad_norm": 0.9579987525939941,
      "learning_rate": 2.710438147676804e-05,
      "loss": 0.9882,
      "step": 33800
    },
    {
      "epoch": 2.6036458733943544,
      "grad_norm": 1.0152076482772827,
      "learning_rate": 2.6844661454951564e-05,
      "loss": 1.041,
      "step": 33850
    },
    {
      "epoch": 2.6074917314052763,
      "grad_norm": 1.1370351314544678,
      "learning_rate": 2.658494143313508e-05,
      "loss": 0.9518,
      "step": 33900
    },
    {
      "epoch": 2.6113375894161988,
      "grad_norm": 0.9851937890052795,
      "learning_rate": 2.63252214113186e-05,
      "loss": 0.9125,
      "step": 33950
    },
    {
      "epoch": 2.615183447427121,
      "grad_norm": 0.8480270504951477,
      "learning_rate": 2.6065501389502116e-05,
      "loss": 0.9736,
      "step": 34000
    },
    {
      "epoch": 2.615183447427121,
      "eval_loss": 0.987713634967804,
      "eval_runtime": 17.2436,
      "eval_samples_per_second": 57.993,
      "eval_steps_per_second": 14.498,
      "step": 34000
    },
    {
      "epoch": 2.619029305438043,
      "grad_norm": 0.5307362079620361,
      "learning_rate": 2.580578136768564e-05,
      "loss": 1.0031,
      "step": 34050
    },
    {
      "epoch": 2.6228751634489655,
      "grad_norm": 1.1112557649612427,
      "learning_rate": 2.5546061345869156e-05,
      "loss": 0.9928,
      "step": 34100
    },
    {
      "epoch": 2.626721021459888,
      "grad_norm": 0.646759569644928,
      "learning_rate": 2.5286341324052672e-05,
      "loss": 1.0173,
      "step": 34150
    },
    {
      "epoch": 2.63056687947081,
      "grad_norm": 1.0114878416061401,
      "learning_rate": 2.502662130223619e-05,
      "loss": 0.9765,
      "step": 34200
    },
    {
      "epoch": 2.6344127374817323,
      "grad_norm": 0.8782021403312683,
      "learning_rate": 2.4766901280419708e-05,
      "loss": 0.9926,
      "step": 34250
    },
    {
      "epoch": 2.6344127374817323,
      "eval_loss": 0.9932020306587219,
      "eval_runtime": 17.1921,
      "eval_samples_per_second": 58.166,
      "eval_steps_per_second": 14.542,
      "step": 34250
    },
    {
      "epoch": 2.6382585954926543,
      "grad_norm": 1.0792268514633179,
      "learning_rate": 2.4507181258603227e-05,
      "loss": 1.0541,
      "step": 34300
    },
    {
      "epoch": 2.6421044535035767,
      "grad_norm": 0.9647793173789978,
      "learning_rate": 2.4247461236786744e-05,
      "loss": 0.9779,
      "step": 34350
    },
    {
      "epoch": 2.6459503115144987,
      "grad_norm": 1.0052498579025269,
      "learning_rate": 2.3987741214970263e-05,
      "loss": 0.9798,
      "step": 34400
    },
    {
      "epoch": 2.649796169525421,
      "grad_norm": 0.8122023344039917,
      "learning_rate": 2.3728021193153783e-05,
      "loss": 1.0205,
      "step": 34450
    },
    {
      "epoch": 2.6536420275363435,
      "grad_norm": 1.452087163925171,
      "learning_rate": 2.3468301171337302e-05,
      "loss": 0.9745,
      "step": 34500
    },
    {
      "epoch": 2.6536420275363435,
      "eval_loss": 0.9959968328475952,
      "eval_runtime": 17.2025,
      "eval_samples_per_second": 58.131,
      "eval_steps_per_second": 14.533,
      "step": 34500
    },
    {
      "epoch": 2.6574878855472654,
      "grad_norm": 1.3329054117202759,
      "learning_rate": 2.3208581149520815e-05,
      "loss": 1.0035,
      "step": 34550
    },
    {
      "epoch": 2.661333743558188,
      "grad_norm": 0.8142715096473694,
      "learning_rate": 2.2948861127704335e-05,
      "loss": 0.9783,
      "step": 34600
    },
    {
      "epoch": 2.6651796015691103,
      "grad_norm": 0.5424798130989075,
      "learning_rate": 2.2689141105887855e-05,
      "loss": 0.9754,
      "step": 34650
    },
    {
      "epoch": 2.6690254595800322,
      "grad_norm": 0.8890462517738342,
      "learning_rate": 2.2429421084071374e-05,
      "loss": 0.9707,
      "step": 34700
    },
    {
      "epoch": 2.6728713175909546,
      "grad_norm": 1.0329838991165161,
      "learning_rate": 2.216970106225489e-05,
      "loss": 0.9845,
      "step": 34750
    },
    {
      "epoch": 2.6728713175909546,
      "eval_loss": 0.9899721741676331,
      "eval_runtime": 17.3008,
      "eval_samples_per_second": 57.801,
      "eval_steps_per_second": 14.45,
      "step": 34750
    },
    {
      "epoch": 2.676717175601877,
      "grad_norm": 1.0697598457336426,
      "learning_rate": 2.190998104043841e-05,
      "loss": 1.0167,
      "step": 34800
    },
    {
      "epoch": 2.680563033612799,
      "grad_norm": 0.8134399056434631,
      "learning_rate": 2.1650261018621926e-05,
      "loss": 0.9001,
      "step": 34850
    },
    {
      "epoch": 2.684408891623721,
      "grad_norm": 1.0405962467193604,
      "learning_rate": 2.1390540996805446e-05,
      "loss": 1.0184,
      "step": 34900
    },
    {
      "epoch": 2.6882547496346434,
      "grad_norm": 1.3229318857192993,
      "learning_rate": 2.1130820974988962e-05,
      "loss": 0.9556,
      "step": 34950
    },
    {
      "epoch": 2.692100607645566,
      "grad_norm": 1.2907413244247437,
      "learning_rate": 2.087110095317248e-05,
      "loss": 0.9355,
      "step": 35000
    },
    {
      "epoch": 2.692100607645566,
      "eval_loss": 0.9918172359466553,
      "eval_runtime": 17.2562,
      "eval_samples_per_second": 57.95,
      "eval_steps_per_second": 14.488,
      "step": 35000
    },
    {
      "epoch": 2.6959464656564878,
      "grad_norm": 1.3044216632843018,
      "learning_rate": 2.0611380931356e-05,
      "loss": 0.9787,
      "step": 35050
    },
    {
      "epoch": 2.69979232366741,
      "grad_norm": 1.4273097515106201,
      "learning_rate": 2.0351660909539518e-05,
      "loss": 0.9775,
      "step": 35100
    },
    {
      "epoch": 2.7036381816783326,
      "grad_norm": 1.6097638607025146,
      "learning_rate": 2.0091940887723034e-05,
      "loss": 0.9926,
      "step": 35150
    },
    {
      "epoch": 2.7074840396892546,
      "grad_norm": 1.4893895387649536,
      "learning_rate": 1.9832220865906553e-05,
      "loss": 1.0435,
      "step": 35200
    },
    {
      "epoch": 2.711329897700177,
      "grad_norm": 1.1620233058929443,
      "learning_rate": 1.9572500844090073e-05,
      "loss": 0.9525,
      "step": 35250
    },
    {
      "epoch": 2.711329897700177,
      "eval_loss": 0.9926208257675171,
      "eval_runtime": 17.3977,
      "eval_samples_per_second": 57.479,
      "eval_steps_per_second": 14.37,
      "step": 35250
    },
    {
      "epoch": 2.7151757557110994,
      "grad_norm": 1.2087517976760864,
      "learning_rate": 1.931278082227359e-05,
      "loss": 1.0005,
      "step": 35300
    },
    {
      "epoch": 2.7190216137220213,
      "grad_norm": 0.8885460495948792,
      "learning_rate": 1.9058255200893437e-05,
      "loss": 0.9911,
      "step": 35350
    },
    {
      "epoch": 2.7228674717329437,
      "grad_norm": 0.9543077945709229,
      "learning_rate": 1.8798535179076957e-05,
      "loss": 1.0437,
      "step": 35400
    },
    {
      "epoch": 2.7267133297438657,
      "grad_norm": 1.2362306118011475,
      "learning_rate": 1.8538815157260476e-05,
      "loss": 0.9766,
      "step": 35450
    },
    {
      "epoch": 2.730559187754788,
      "grad_norm": 1.1946227550506592,
      "learning_rate": 1.8279095135443993e-05,
      "loss": 1.0333,
      "step": 35500
    },
    {
      "epoch": 2.730559187754788,
      "eval_loss": 0.9953948855400085,
      "eval_runtime": 17.1888,
      "eval_samples_per_second": 58.178,
      "eval_steps_per_second": 14.544,
      "step": 35500
    },
    {
      "epoch": 2.73440504576571,
      "grad_norm": 0.66939777135849,
      "learning_rate": 1.801937511362751e-05,
      "loss": 0.8892,
      "step": 35550
    },
    {
      "epoch": 2.7382509037766325,
      "grad_norm": 1.0852998495101929,
      "learning_rate": 1.775965509181103e-05,
      "loss": 0.9502,
      "step": 35600
    },
    {
      "epoch": 2.742096761787555,
      "grad_norm": 0.7603423595428467,
      "learning_rate": 1.7499935069994548e-05,
      "loss": 1.0034,
      "step": 35650
    },
    {
      "epoch": 2.745942619798477,
      "grad_norm": 0.4625702202320099,
      "learning_rate": 1.7240215048178064e-05,
      "loss": 1.0123,
      "step": 35700
    },
    {
      "epoch": 2.7497884778093993,
      "grad_norm": 1.4387953281402588,
      "learning_rate": 1.6980495026361584e-05,
      "loss": 0.9704,
      "step": 35750
    },
    {
      "epoch": 2.7497884778093993,
      "eval_loss": 0.9952225685119629,
      "eval_runtime": 17.4458,
      "eval_samples_per_second": 57.32,
      "eval_steps_per_second": 14.33,
      "step": 35750
    },
    {
      "epoch": 2.7536343358203217,
      "grad_norm": 0.6899126172065735,
      "learning_rate": 1.67207750045451e-05,
      "loss": 0.9627,
      "step": 35800
    },
    {
      "epoch": 2.7574801938312437,
      "grad_norm": 1.0329424142837524,
      "learning_rate": 1.646105498272862e-05,
      "loss": 0.9207,
      "step": 35850
    },
    {
      "epoch": 2.761326051842166,
      "grad_norm": 1.1055504083633423,
      "learning_rate": 1.6201334960912136e-05,
      "loss": 0.9834,
      "step": 35900
    },
    {
      "epoch": 2.7651719098530885,
      "grad_norm": 0.7458188533782959,
      "learning_rate": 1.5941614939095655e-05,
      "loss": 1.003,
      "step": 35950
    },
    {
      "epoch": 2.7690177678640104,
      "grad_norm": 1.112021803855896,
      "learning_rate": 1.5681894917279175e-05,
      "loss": 1.0001,
      "step": 36000
    },
    {
      "epoch": 2.7690177678640104,
      "eval_loss": 0.9910063147544861,
      "eval_runtime": 17.2718,
      "eval_samples_per_second": 57.898,
      "eval_steps_per_second": 14.474,
      "step": 36000
    },
    {
      "epoch": 2.772863625874933,
      "grad_norm": 0.5958703756332397,
      "learning_rate": 1.5422174895462695e-05,
      "loss": 0.9503,
      "step": 36050
    },
    {
      "epoch": 2.776709483885855,
      "grad_norm": 1.1087392568588257,
      "learning_rate": 1.516245487364621e-05,
      "loss": 0.9725,
      "step": 36100
    },
    {
      "epoch": 2.780555341896777,
      "grad_norm": 1.2012007236480713,
      "learning_rate": 1.4902734851829727e-05,
      "loss": 1.0098,
      "step": 36150
    },
    {
      "epoch": 2.784401199907699,
      "grad_norm": 0.9333285093307495,
      "learning_rate": 1.4643014830013247e-05,
      "loss": 0.9764,
      "step": 36200
    },
    {
      "epoch": 2.7882470579186216,
      "grad_norm": 0.9407594203948975,
      "learning_rate": 1.4383294808196765e-05,
      "loss": 0.9404,
      "step": 36250
    },
    {
      "epoch": 2.7882470579186216,
      "eval_loss": 0.9894633889198303,
      "eval_runtime": 17.3239,
      "eval_samples_per_second": 57.724,
      "eval_steps_per_second": 14.431,
      "step": 36250
    },
    {
      "epoch": 2.792092915929544,
      "grad_norm": 0.9794307947158813,
      "learning_rate": 1.4123574786380284e-05,
      "loss": 0.9445,
      "step": 36300
    },
    {
      "epoch": 2.795938773940466,
      "grad_norm": 1.1447358131408691,
      "learning_rate": 1.3863854764563799e-05,
      "loss": 1.0056,
      "step": 36350
    },
    {
      "epoch": 2.7997846319513884,
      "grad_norm": 0.8741857409477234,
      "learning_rate": 1.3604134742747318e-05,
      "loss": 0.9511,
      "step": 36400
    },
    {
      "epoch": 2.803630489962311,
      "grad_norm": 1.0769715309143066,
      "learning_rate": 1.3344414720930836e-05,
      "loss": 0.9789,
      "step": 36450
    },
    {
      "epoch": 2.8074763479732328,
      "grad_norm": 1.1045129299163818,
      "learning_rate": 1.3084694699114356e-05,
      "loss": 0.9958,
      "step": 36500
    },
    {
      "epoch": 2.8074763479732328,
      "eval_loss": 0.9914500117301941,
      "eval_runtime": 17.3793,
      "eval_samples_per_second": 57.54,
      "eval_steps_per_second": 14.385,
      "step": 36500
    },
    {
      "epoch": 2.811322205984155,
      "grad_norm": 0.9856983423233032,
      "learning_rate": 1.2824974677297874e-05,
      "loss": 0.9923,
      "step": 36550
    },
    {
      "epoch": 2.8151680639950776,
      "grad_norm": 1.2112038135528564,
      "learning_rate": 1.2565254655481392e-05,
      "loss": 0.9388,
      "step": 36600
    },
    {
      "epoch": 2.8190139220059995,
      "grad_norm": 1.1824342012405396,
      "learning_rate": 1.230553463366491e-05,
      "loss": 0.984,
      "step": 36650
    },
    {
      "epoch": 2.8228597800169215,
      "grad_norm": 1.3278725147247314,
      "learning_rate": 1.204581461184843e-05,
      "loss": 0.9835,
      "step": 36700
    },
    {
      "epoch": 2.826705638027844,
      "grad_norm": 1.4297553300857544,
      "learning_rate": 1.1786094590031946e-05,
      "loss": 0.9999,
      "step": 36750
    },
    {
      "epoch": 2.826705638027844,
      "eval_loss": 0.9847651720046997,
      "eval_runtime": 17.5097,
      "eval_samples_per_second": 57.111,
      "eval_steps_per_second": 14.278,
      "step": 36750
    },
    {
      "epoch": 2.8305514960387663,
      "grad_norm": 1.1944117546081543,
      "learning_rate": 1.1526374568215465e-05,
      "loss": 0.9758,
      "step": 36800
    },
    {
      "epoch": 2.8343973540496883,
      "grad_norm": 1.1629287004470825,
      "learning_rate": 1.1266654546398983e-05,
      "loss": 0.988,
      "step": 36850
    },
    {
      "epoch": 2.8382432120606107,
      "grad_norm": 1.598382592201233,
      "learning_rate": 1.1006934524582501e-05,
      "loss": 0.9679,
      "step": 36900
    },
    {
      "epoch": 2.842089070071533,
      "grad_norm": 0.9534172415733337,
      "learning_rate": 1.0747214502766019e-05,
      "loss": 0.9921,
      "step": 36950
    },
    {
      "epoch": 2.845934928082455,
      "grad_norm": 1.5716655254364014,
      "learning_rate": 1.0487494480949537e-05,
      "loss": 0.9898,
      "step": 37000
    },
    {
      "epoch": 2.845934928082455,
      "eval_loss": 0.9878412485122681,
      "eval_runtime": 17.2808,
      "eval_samples_per_second": 57.868,
      "eval_steps_per_second": 14.467,
      "step": 37000
    },
    {
      "epoch": 2.8497807860933775,
      "grad_norm": 1.3013421297073364,
      "learning_rate": 1.0227774459133055e-05,
      "loss": 0.9717,
      "step": 37050
    },
    {
      "epoch": 2.8536266441043,
      "grad_norm": 1.263071060180664,
      "learning_rate": 9.968054437316573e-06,
      "loss": 1.0259,
      "step": 37100
    },
    {
      "epoch": 2.857472502115222,
      "grad_norm": 1.1513851881027222,
      "learning_rate": 9.70833441550009e-06,
      "loss": 1.0015,
      "step": 37150
    },
    {
      "epoch": 2.8613183601261443,
      "grad_norm": 0.7431422472000122,
      "learning_rate": 9.448614393683609e-06,
      "loss": 0.999,
      "step": 37200
    },
    {
      "epoch": 2.8651642181370662,
      "grad_norm": 0.6744217872619629,
      "learning_rate": 9.188894371867128e-06,
      "loss": 0.9285,
      "step": 37250
    },
    {
      "epoch": 2.8651642181370662,
      "eval_loss": 0.994976282119751,
      "eval_runtime": 17.3294,
      "eval_samples_per_second": 57.705,
      "eval_steps_per_second": 14.426,
      "step": 37250
    },
    {
      "epoch": 2.8690100761479886,
      "grad_norm": 1.2962367534637451,
      "learning_rate": 8.929174350050646e-06,
      "loss": 0.9433,
      "step": 37300
    },
    {
      "epoch": 2.8728559341589106,
      "grad_norm": 0.9955423474311829,
      "learning_rate": 8.669454328234164e-06,
      "loss": 1.0447,
      "step": 37350
    },
    {
      "epoch": 2.876701792169833,
      "grad_norm": 0.5840064287185669,
      "learning_rate": 8.409734306417682e-06,
      "loss": 0.9504,
      "step": 37400
    },
    {
      "epoch": 2.8805476501807554,
      "grad_norm": 1.0777620077133179,
      "learning_rate": 8.1500142846012e-06,
      "loss": 0.9635,
      "step": 37450
    },
    {
      "epoch": 2.8843935081916774,
      "grad_norm": 0.9312844276428223,
      "learning_rate": 7.890294262784718e-06,
      "loss": 1.005,
      "step": 37500
    },
    {
      "epoch": 2.8843935081916774,
      "eval_loss": 0.9864250421524048,
      "eval_runtime": 17.2238,
      "eval_samples_per_second": 58.059,
      "eval_steps_per_second": 14.515,
      "step": 37500
    },
    {
      "epoch": 2.8882393662026,
      "grad_norm": 0.6800574660301208,
      "learning_rate": 7.630574240968237e-06,
      "loss": 0.983,
      "step": 37550
    },
    {
      "epoch": 2.892085224213522,
      "grad_norm": 0.9417561292648315,
      "learning_rate": 7.3708542191517545e-06,
      "loss": 0.973,
      "step": 37600
    },
    {
      "epoch": 2.895931082224444,
      "grad_norm": 0.6454310417175293,
      "learning_rate": 7.111134197335273e-06,
      "loss": 1.0493,
      "step": 37650
    },
    {
      "epoch": 2.8997769402353666,
      "grad_norm": 1.0287562608718872,
      "learning_rate": 6.851414175518792e-06,
      "loss": 0.9821,
      "step": 37700
    },
    {
      "epoch": 2.903622798246289,
      "grad_norm": 1.6874816417694092,
      "learning_rate": 6.591694153702309e-06,
      "loss": 0.9761,
      "step": 37750
    },
    {
      "epoch": 2.903622798246289,
      "eval_loss": 0.9886119961738586,
      "eval_runtime": 17.2542,
      "eval_samples_per_second": 57.957,
      "eval_steps_per_second": 14.489,
      "step": 37750
    },
    {
      "epoch": 2.907468656257211,
      "grad_norm": 0.7677489519119263,
      "learning_rate": 6.331974131885828e-06,
      "loss": 1.0421,
      "step": 37800
    },
    {
      "epoch": 2.9113145142681334,
      "grad_norm": 0.9784395098686218,
      "learning_rate": 6.072254110069346e-06,
      "loss": 0.9288,
      "step": 37850
    },
    {
      "epoch": 2.9151603722790553,
      "grad_norm": 0.6236763000488281,
      "learning_rate": 5.812534088252864e-06,
      "loss": 0.9777,
      "step": 37900
    },
    {
      "epoch": 2.9190062302899777,
      "grad_norm": 1.0675079822540283,
      "learning_rate": 5.552814066436382e-06,
      "loss": 0.9659,
      "step": 37950
    },
    {
      "epoch": 2.9228520883008997,
      "grad_norm": 0.9310262203216553,
      "learning_rate": 5.2930940446198996e-06,
      "loss": 0.9334,
      "step": 38000
    },
    {
      "epoch": 2.9228520883008997,
      "eval_loss": 0.98431396484375,
      "eval_runtime": 17.3967,
      "eval_samples_per_second": 57.482,
      "eval_steps_per_second": 14.371,
      "step": 38000
    },
    {
      "epoch": 2.926697946311822,
      "grad_norm": 0.9615415930747986,
      "learning_rate": 5.038568423239748e-06,
      "loss": 0.9715,
      "step": 38050
    },
    {
      "epoch": 2.9305438043227445,
      "grad_norm": 0.5959100127220154,
      "learning_rate": 4.778848401423266e-06,
      "loss": 0.936,
      "step": 38100
    },
    {
      "epoch": 2.9343896623336665,
      "grad_norm": 1.0181951522827148,
      "learning_rate": 4.519128379606784e-06,
      "loss": 0.969,
      "step": 38150
    },
    {
      "epoch": 2.938235520344589,
      "grad_norm": 0.8716458082199097,
      "learning_rate": 4.259408357790302e-06,
      "loss": 1.002,
      "step": 38200
    },
    {
      "epoch": 2.9420813783555113,
      "grad_norm": 1.6100435256958008,
      "learning_rate": 3.99968833597382e-06,
      "loss": 0.9425,
      "step": 38250
    },
    {
      "epoch": 2.9420813783555113,
      "eval_loss": 0.9931854009628296,
      "eval_runtime": 17.3406,
      "eval_samples_per_second": 57.668,
      "eval_steps_per_second": 14.417,
      "step": 38250
    },
    {
      "epoch": 2.9459272363664333,
      "grad_norm": 1.1552485227584839,
      "learning_rate": 3.7399683141573383e-06,
      "loss": 1.0424,
      "step": 38300
    },
    {
      "epoch": 2.9497730943773557,
      "grad_norm": 1.441148042678833,
      "learning_rate": 3.480248292340857e-06,
      "loss": 1.0688,
      "step": 38350
    },
    {
      "epoch": 2.953618952388278,
      "grad_norm": 0.8127447366714478,
      "learning_rate": 3.220528270524375e-06,
      "loss": 0.9497,
      "step": 38400
    },
    {
      "epoch": 2.9574648103992,
      "grad_norm": 0.9566346406936646,
      "learning_rate": 2.9608082487078933e-06,
      "loss": 0.9628,
      "step": 38450
    },
    {
      "epoch": 2.961310668410122,
      "grad_norm": 0.7955614924430847,
      "learning_rate": 2.7010882268914113e-06,
      "loss": 1.016,
      "step": 38500
    },
    {
      "epoch": 2.961310668410122,
      "eval_loss": 0.9919273257255554,
      "eval_runtime": 17.3829,
      "eval_samples_per_second": 57.528,
      "eval_steps_per_second": 14.382,
      "step": 38500
    },
    {
      "epoch": 2.9651565264210444,
      "grad_norm": 0.8041125535964966,
      "learning_rate": 2.4413682050749296e-06,
      "loss": 0.9631,
      "step": 38550
    },
    {
      "epoch": 2.969002384431967,
      "grad_norm": 1.0755919218063354,
      "learning_rate": 2.1816481832584475e-06,
      "loss": 0.9689,
      "step": 38600
    },
    {
      "epoch": 2.972848242442889,
      "grad_norm": 0.8630362153053284,
      "learning_rate": 1.9219281614419654e-06,
      "loss": 0.9822,
      "step": 38650
    },
    {
      "epoch": 2.976694100453811,
      "grad_norm": 0.6994553804397583,
      "learning_rate": 1.662208139625484e-06,
      "loss": 0.9706,
      "step": 38700
    },
    {
      "epoch": 2.9805399584647336,
      "grad_norm": 1.1746189594268799,
      "learning_rate": 1.4024881178090021e-06,
      "loss": 0.9695,
      "step": 38750
    },
    {
      "epoch": 2.9805399584647336,
      "eval_loss": 0.9868382215499878,
      "eval_runtime": 17.3783,
      "eval_samples_per_second": 57.543,
      "eval_steps_per_second": 14.386,
      "step": 38750
    },
    {
      "epoch": 2.9843858164756556,
      "grad_norm": 1.0636792182922363,
      "learning_rate": 1.14276809599252e-06,
      "loss": 1.0233,
      "step": 38800
    },
    {
      "epoch": 2.988231674486578,
      "grad_norm": 0.8173992037773132,
      "learning_rate": 8.830480741760383e-07,
      "loss": 0.963,
      "step": 38850
    },
    {
      "epoch": 2.9920775324975004,
      "grad_norm": 1.591539978981018,
      "learning_rate": 6.233280523595564e-07,
      "loss": 0.949,
      "step": 38900
    },
    {
      "epoch": 2.9959233905084224,
      "grad_norm": 0.6467046737670898,
      "learning_rate": 3.636080305430746e-07,
      "loss": 0.9649,
      "step": 38950
    },
    {
      "epoch": 2.999769248519345,
      "grad_norm": 0.9910799860954285,
      "learning_rate": 1.0388800872659275e-07,
      "loss": 1.0502,
      "step": 39000
    },
    {
      "epoch": 2.999769248519345,
      "eval_loss": 0.9894677400588989,
      "eval_runtime": 17.4751,
      "eval_samples_per_second": 57.224,
      "eval_steps_per_second": 14.306,
      "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
}